quality 8 dis

2.PatientSafetyandHealthcareImprovementataGlance.pdf

Home >English homework help >Article writing homework help >quality 8 dis

Patient Safety and Healthcare Improvement at a Glance Edited by Sukhmeet S. Panesar Andrew Carson-Stevens Sarah A. Salvilla Aziz Sheikh

Patient Safety and Healthcare Improvement at a Glance

This title is also available as an e-book. For more details, please see www.wiley.com/buy/9781118361368 or scan this QR code:

FPO

Patient Safety and Healthcare Improvement at a Glance Edited by

Sukhmeet S. Panesar BSc (Hons.), MBBS, AICSM, MPH, MD Honorary Fellow Centre for Population Health Sciences The University of Edinburgh Edinburgh, UK

Andrew Carson-Stevens BSc (Hons.), MB BCh, MPhil Clinical Lecturer in Healthcare Improvement Cochrane Institute of Primary Care and Public Health Cardiff University School of Medicine Cardiff, UK

Sarah A. Salvilla BSc (Hons.), MBBS, MSc Honorary Fellow Centre for Population Health Sciences The University of Edinburgh Edinburgh, UK

Aziz Sheikh BSc, MBBS, MSc, MD, FRCGP, FRCP, FRCPE Professor of Primary Care Research and Development Co-Director of Centre for Population Health Sciences The University of Edinburgh Edinburgh, UK; Visiting Professor of Medicine Harvard Medical School Harkness Fellow in Health Policy and Practice Brigham and Women’s Hospital Harvard Medical School Boston, MA, USA

Registered offi ce: John Wiley & Sons, Ltd, Th e Atrium, Southern Gate, Chichester, West Sussex, PO19 8SQ, UK

Editorial offi ces: 9600 Garsington Road, Oxford, OX4 2DQ, UK Th e Atrium, Southern Gate, Chichester, West Sussex, PO19 8SQ, UK 350 Main Street, Malden, MA 02148-5020, USA

For details of our global editorial offi ces, for customer services and for information about how to apply for permission to reuse the copyright material in this book please see our website at www.wiley.com/wiley-blackwell

Th e right of the authors to be identifi ed as the authors of this work has been asserted in accord- ance with the UK Copyright, Designs and Patents Act 1988.

All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or transmitted, in any form or by any means, electronic, mechanical, photocopying, recording or otherwise, except as permitted by the UK Copyright, Designs and Patents Act 1988, without the prior permission of the publisher.

Designations used by companies to distinguish their products are oft en claimed as trademarks. All brand names and product names used in this book are trade names, service marks, trade- marks or registered trademarks of their respective owners. Th e publisher is not associated with any product or vendor mentioned in this book. It is sold on the understanding that the pub- lisher is not engaged in rendering professional services. If professional advice or other expert assistance is required, the services of a competent professional should be sought.

Th e contents of this work are intended to further general scientifi c research, understanding, and discussion only and are not intended and should not be relied upon as recommending or promoting a specifi c method, diagnosis, or treatment by health science practitioners for any particular patient. Th e publisher and the author make no representations or warranties with respect to the accuracy or completeness of the contents of this work and specifi cally disclaim all warranties, including without limitation any implied warranties of fi tness for a particular pur- pose. In view of ongoing research, equipment modifi cations, changes in governmental regula- tions, and the constant fl ow of information relating to the use of medicines, equipment, and devices, the reader is urged to review and evaluate the information provided in the package insert or instructions for each medicine, equipment, or device for, among other things, any changes in the instructions or indication of usage and for added warnings and precautions. Readers should consult with a specialist where appropriate. Th e fact that an organization or Website is referred to in this work as a citation and/or a potential source of further information does not mean that the author or the publisher endorses the information the organization or Website may provide or recommendations it may make. Further, readers should be aware that Internet Websites listed in this work may have changed or disappeared between when this work was written and when it is read. No warranty may be created or extended by any promotional statements for this work. Neither the publisher nor the author shall be liable for any damages arising herefrom.

Library of Congress Cataloging-in-Publication Data Patient safety and healthcare improvement at a glance / edited by Sukhmeet S. Panesar, Andrew Carson-Stevens, Sarah A. Salvilla, Aziz Sheikh. p. ; cm. Includes bibliographical references and index. ISBN 978-1-118-36136-8 (pbk.) I. Panesar, Sukhmeet S., editor. II. Carson-Stevens, Andrew, editor. III. Salvilla, Sarah A., editor. IV. Sheikh, Aziz, editor. [DNLM: 1. Patient Safety–standards. 2. Medical Errors–prevention & control. 3. Quality of Health Care–standards. 4. Safety Management. WX 185] R729.8 610.28ʹ9–dc23 2014008376

A catalogue record for this book is available from the British Library.

Wiley also publishes its books in a variety of electronic formats. Some content that appears in print may not be available in electronic books.

Cover image: LTH NHS TRUST/SCIENCE PHOTO LIBRARY Cover design by Meaden Creative

Set in Minion Pro 9.5/11.5 by Aptara

1 2014

Contents

Contributors vii Preface xi Acknowledgements xii How to use your revision guide xiii

The essence of patient safety 1 1 Basics of patient safety 2

Kaveh G. Shojania and Sukhmeet S. Panesar

2 Understanding systems 4 Carl Macrae

3 Quality and safety 6 Ranjit Singh and Gurdev Singh

4 Human factors 8 Ken Catchpole

5 Teamwork and communication 12 Sukhmeet S. Panesar, Sarah A. Salvilla, Martin Bromiley and Jane Reid

6 Reporting and learning from errors 14 Tara Lamont

7 Research in patient safety 16 Lilly D. Engineer and Peter Pronovost

Understanding and interpreting risk 19 8 Risk-based patient safety metrics 20

Elizabeth Allen and Sukhmeet S. Panesar

9 Root cause analysis 24 Donna Forsyth and Sundeep Thusu

10 Measuring safety culture 26 Debra de Silva

Risks to patient care 31 11 Medication errors 32

Sarah P. Slight, Tony Avery and David W. Bates

12 Surgical errors 36 Sukhmeet S. Panesar, Bhupinder Mann, Rajan Madhok and Andrew Carson-Stevens

13 Diagnostic errors 40 Ashley N. D. Meyer, Velma L. Payne, Hardeep Singh and Mark L. Graber

14 Maternal and child health errors 44 Susan Leavitt Gullo and Pierre Barker

15 Slips, trips and falls 47 Susan Poulton and Frances Healey

Part 1

Part 2

Part 3

16 Patient safety in paediatrics 50 Peter Lachman and Jane Runnacles

17 Technology in healthcare and e-iatrogenesis 54 Kathrin M. Cresswell

18 Nosocomial infections 58 Imran Qureshi and Sukhmeet S. Panesar

19 Mental health errors 60 Amar Shah and Kevin Cleary

20 Patient safety in primary care 64 Andrew Carson-Stevens and Adrian Edwards

Quality improvement 67 21 Improving the quality of clinical care 68

Andrew Carson-Stevens and Maureen Bisognano

22 Science of improvement 70 Clifford Norman and C. Jane Norman

23 Model for Improvement 74 Clifford Norman and C. Jane Norman

24 Measurement for improvement 78 Mike Davidge

25 Spread and sustainability of improvement 84 Gareth Parry and Andrew Carson-Stevens

26 Quality improvement tools: visualisation 86 Ashley Kay Childers and David M. Neyens

27 Quality improvement: assessing the system 88 Ashley Kay Childers and David M. Neyens

28 Patient stories in improvement 90 Aled Jones and Andrew Carson-Stevens

29 Leading change in healthcare 92 Helen Bevan

30 Public narrative: story of self, us and now 96 Jay D. Bhatt and Andrew Carson-Stevens

31 Planning an improvement project 98 Lakshman Swamy and James Moses

32 Managing an improvement project 100 Valerie P. Pracilio

33 Quality improvement in psychiatry 104 Peter Klinger, Anthony Weiss and Eric Hazen

34 Quality improvement in intensive care 106 Kevin D. Rooney

35 Quality improvement in obstetrics 108 Gloria Esgebona

36 Quality improvement in surgery 110 Shabnam Hafi z

37 Population health and improvement 112 Mohammed Mustafa and Valerie P. Pracilio

Further reading 114 Index 121

Part 4

vii

Contributors

Elizabeth Allen MPH Postgraduate Student Department of Public Health Imperial College London London, UK

Tony Avery MBBS, PhD, FRCGP Professor of Primary Health Care/Joint Head of Division (Primary Care), Faculty of Medicine and Health Sciences The University of Nottingham Nottingham, UK

Pierre Barker MD Senior Vice President Institute for Healthcare Improvement Clinical Professor Maternal and Child Health Department of Public Health University of North Carolina Chapel Hill, NC, USA

David W. Bates MD, MSc Professor of Medicine Harvard Medical School Professor of Health Policy and Management Harvard School of Public Health; Chief of the Division of General Internal Medicine Brigham and Women’s Hospital; Medical Director, Clinical and Quality Analysis Partner’s HealthCare System Boston, MA, USA

Helen Bevan MBA, DBA Chief Transformation Offi cer Horizons Group NHS Improving Quality Coventry, UK

Jay D. Bhatt DO, MPH, MPA, FACP Associate Physician Health System Clinical Adjunct Lecturer Department of Internal Medicine and Geriatrics Northwestern University Chicago, IL, USA

Maureen Bisognano MS President/CEO Institute for Healthcare Improvement Cambridge, MA, USA

Martin Bromiley Airline Transport Pilot’s Licence Chair, Clinical Human Factors Group North Marston, UK

Andrew Carson-Stevens BSc (Hons.), MB BCh, MPhil Clinical Lecturer in Healthcare Improvement Cochrane Institute of Primary Care and Public Health Cardiff University School of Medicine Cardiff, UK

Ken Catchpole BSc (Hons.), PhD Director of Surgical Safety and Human Factors Research Department of Surgery Cedars-Sinai Medical Center Los Angeles, CA, USA

Ashley Kay Childers PhD, CPHQ Research Assistant Professor Department of Industrial Engineering Clemson University Clemson, SC, USA

Kevin Cleary MBChB, FRCPsych Medical Director Director for Quality and Performance and Consultant

Forensic Psychiatrist East London NHS Foundation Trust London, UK

vii

Kathrin M. Cresswell BSc, MSc, PhD Chancellor’s Fellow The School of Health in Social Science The University of Edinburgh Edinburgh, UK

Mike Davidge BSc, BCom Director, NHS Elect London, UK

Adrian Edwards MBBS, MRCP, MRCGP, PhD Institute Director and Professor of Primary Care Cochrane Institute of Primary Care and Public Health Cardiff University School of Medicine Cardiff, UK

Lilly D. Engineer MBBS-MD, DrPH, MHA Associate Director, DrPH Programme in Health Care

Management and Leadership Department of Health Policy and Management Johns Hopkins Bloomberg School of Public Health; Assistant Professor Department of Anesthesiology and Critical Care Medicine Johns Hopkins School of Medicine Baltimore, MD, USA

Gloria Esegbona MBBS, BSc, MSc, MBA, MRCOG Consultant Obstetrician and Gynaecologist & Lecturer Department of Women’s Health Mzati Trust Blantyre, Malawi

Donna Forsyth MSCP, CMIOSH Head of Patient Safety Investigation Department of Patient Safety NHS England London, UK

Mark L. Graber MD, FACP Senior Fellow RTI International Professor Emeritus SUNY Stony Brook School of Medicine Founder and President Society to Improve Diagnosis in

Medicine St. James, NY, USA

Shabnam Hafi z BS, MPH, MD General Surgery Resident Medstar Washington Hospital Center Washington, DC, USA

Eric Hazen MD Instructor in Psychiatry Harvard Medical School; Director, Pediatric Psychiatry Consultation Service Massachusetts General Hospital Boston, MA, USA

Frances Healey RN, PhD Senior Head of Patient Safety Intelligence, Research and

Evaluation Patient Safety Domain NHS England Leeds, UK

Ross W. Hilliard MD Resident, General Internal Medicine The Warren Alpert Medical School of Brown University Providence, RI, USA

Aled Jones PhD, BN (Hons.), RN (Adult), RMN Senior Lecturer School of Healthcare Sciences Cardiff University Cardiff, UK

Peter Klinger MD Instructor in Psychiatry Harvard Medical School Boston, MA, USA

Peter Lachman MD, MMed, MPH, MBBCH, BA, FRCPH, FCP(SA) Deputy Medical Director (Patient Safety) Medical Director Great Ormond Street Hospital Foundation

NHS Trust London, UK

Tara Lamont MSc Scientifi c Advisor NIHR Health Service Delivery and Research (HS&DR)

Programme University of Southampton Southampton, UK

Susan Leavitt Gullo MS, BSN, RN Director Institute for Healthcare Improvement Cambridge, MA, USA

viii

Carl Macrae PhD Senior Research Fellow Centre for Patient Safety and Service Quality Imperial College London London, UK

Rajan Madhok MBBS, MSc, FRCS, FFPH Professor of Public Health Department of Public Health University of Salford Salford, UK

Bhupinder Mann BSc (Hons.), FRCS Consultant Orthopaedic Surgeon Department of Trauma and Orthopaedic Surgery Stoke Mandeville Hospital Aylesbury, UK

Ashley N. D. Meyer PhD Health Science Specialist (Cognitive Psychologist) Veterans Affairs Health Services Research & Development

Center for Innovations in Quality, Effectiveness and Safety Michael E. DeBakey Veterans Affairs Medical Center Houston, TX, USA

James Moses MD, MPH Medical Director of Quality Improvement Department of Quality and Patient Safety Boston University School of Medicine Boston Medical Center Boston, MA, USA

Mohammed Mustafa BSc, MBChB, MRCGP, MSc Clinical Lecturer in Primary Care and Public Health Cochrane Institute of Primary Care and Public Health Cardiff University School of Medicine Cardiff, UK

David M. Neyens PhD, MPH Assistant Professor Department of Industrial Engineering Clemson University Clemson, SC, USA

C. Jane Norman BA, MBA, CQE President Profound Knowledge Products (PKP Inc.) Austin, TX, USA

Clifford L. Norman MA Partner, Associates in Process Improvement (API) Austin, TX, USA

Sukhmeet S. Panesar BSc (Hons.), MBBS, AICSM, MPH, MD Honorary Fellow The Centre for Population Health Sciences The University of Edinburgh Edinburgh, UK

Gareth J. Parry BSc, MSc, PhD Senior Scientist Institute for Healthcare Improvement Cambridge, MA, USA

Velma L. Payne PhD Postdoctoral Fellow (Biomedical Informatics Specialist) Veterans Affairs Health Services Research & Development

Center for Innovations in Quality, Effectiveness and Safety Michael E. DeBakey Veterans Affairs Medical Center Houston, TX, USA

Susan Poulton BM, FRCP Consultant Geriatrician Department of Medicine for Older People, Rehabilitation

and Stroke Portsmouth Hospitals NHS Trust Portsmouth, UK

Valerie P. Pracilio MPH, CPPS Client Services Manager Pascal Metrics Washington, DC, USA

Peter Pronovost MD, PhD, FCCM Sr. Vice President for Patient Safety and Quality Director of the Armstrong Institute for Patient Safety and

Quality Johns Hopkins Medicine; Professor Departments of Anesthesiology/Critical Care Medicine and

Surgery Johns Hopkins University School of Medicine; Professor Department of Health Policy & Management Johns Hopkins Bloomberg School of Public Health; Professor Department of Nursing Johns Hopkins University School of Nursing Baltimore, MD, USA

Imran Qureshi BSc (Hons.), AIEE, MBBS BMJ Cinical Lead for Quality and Safety Specialist Registrar in Medical Microbiology Department of Microbiology St George’s Hospital NHS Trust London, UK

Jane Reid BSc, MSc, PGCEA Professor, Bournemouth University Bournemouth, UK

Kevin D. Rooney MBChB, FRCA, FFICM Professor of Care Improvement Consultant in Anaesthesia and Intensive Care Medicine Institute of Care and Practice Improvement University of the West of Scotland and Royal Alexandra

Hospital Paisley, UK

Jane Runnacles MBBS, BSc (Hons.), MRCPCH, MA Consultant Paediatrician Department of Paediatrics Royal Free London NHS Foundation Trust London, UK

Sarah A. Salvilla BSc (Hons.), MBBS, AICSM, MSc Honorary Fellow The Centre for Population Health Sciences The University of Edinburgh Edinburgh, UK

Amar Shah MBBS, MRCPsych, LLM, PGCMedEd, MBA Associate Medical Director (Quality Improvement) &

Consultant Forensic Psychiatrist East London NHS Foundation Trust London, UK

Kaveh G. Shojania MD Director & Associate Professor of Medicine Centre for Patient Safety University of Toronto Toronto, ON, Canada

Debra de Silva PhD Professor and Head of Evaluation The Evidence Centre London, UK

Gurdev Singh BSc Engg(Alig), MSc Eng, PhD(Birm.) Emeritus Founding Director Patient Safety Research Center State University of New York Buffalo, NY, USA

Hardeep Singh MD, MPH Chief, Health Policy, Quality and Informatics Veterans Affairs Health Services Research & Development

Center for Innovations in Quality, Effectiveness and Safety

Michael E. DeBakey Veterans Affairs Medical Center Houston, TX, USA

Ranjit Singh MA (Cantab.), MB, BChir, MBA Vice Chair for Research Department of Family Medicine State University of New York Buffalo, NY, USA

Sarah P. Slight MPharm, PhD, PGDip Senior Lecturer in Pharmacy Practice School of Medicine, Pharmacy and Health Durham University Durham, UK

Lakshman Swamy BA, MD, MBA Boonshoft School of Medicine Wright State University Fairborn, OH, USA

Sundeep Thusu MEng, MBBS, BDS Clinical Research Fellow Centre for International Child Oral Health Kings College London London, UK

Anthony Weiss BS, MD, MSc Assistant Professor of Psychiatry Harvard Medical School Boston, MA, USA

Preface

Healthcare improvement remains the bedrock of any adaptive, learning and high-quality healthcare system. Th e engagement of frontline clinical staff in advancing this agenda is central to ensuring improvements and safety in care delivery, thereby providing the best possible care for the patient. Since the 1990s, there have been concerted eff orts to empower and equip healthcare professionals, carers, students and patients with the knowledge, skills and tools to execute and achieve safer, high-quality, patient- centred care. Th is book is an attempt to synthesise the key lessons learnt and distil these into practical recommendations.

Infl uential reports have raised awareness of healthcare qual- ity and safety in the professional and public conscience. Seminal amongst these have been To Err Is Human, produced by the US Institute of Medicine (IOM), and An Organisation with a Mem- ory, produced by the UK Government’s Chief Medical Offi cer. Th ese reports highlighted that error was routine during the delivery of healthcare and pointed to steps that should be taken to minimise their occurrence and the adverse consequences resulting from these system failures. Th e IOM advises six aims for quality – safety, eff ectiveness, effi ciency, timeliness, patient- centredness and equity. A focus on patient safety has served as a ‘Trojan horse’ to create urgency for change and highlight the major underlying problems in healthcare, and in doing so it has galvanised the importance of seeking all the aims of qual- ity. More recently, the Institute of Healthcare Improvement (IHI) launched Th e Triple Aim that challenges healthcare organisations

to improve patient experience, improve population health and reduce the per capita cost of healthcare in order to optimise health system performance. Building on this approach, many of our contributors have used the lens of patient safety to highlight concerns about and approaches to enhancing the quality of care provision.

Our hope is that this text – which includes contributions from leading international scholars and clinicians in training – will meet the needs of healthcare students and professionals at all stages of their training: from students and junior doctors who have yet to be introduced to the disciplines of healthcare improvement and patient safety to those who want a quick refresher of core concepts and in areas that would be relevant for healthcare professionals in training. Th is refl ects our core belief that all those serving at the ‘coal face’ of healthcare delivery have the capacity to be the barom- eters of the quality and safety of healthcare provision.

Finally, we are optimistic that all those who read this book will in some way – whether by initiating, leading or contributing to collective eff orts – be inspired to move forward the agenda of safe, high-quality, patient-centred care. It is, aft er all, these enduring values that ensure we are fi tting members of ‘the noble profession’ and that we, like every other generation before us, have fulfi lled the charge of ensuring we take stock of preceding eff orts, enrich them and then hand on these quintessential values.

Sukhmeet S. Panesar, Andrew Carson-Stevens, Sarah A. Salvilla and Aziz Sheikh

xii

Acknowledgements

We wish to record our sincere gratitude to our contributors who have taken time away from their many other commit-ments to share their knowledge and insights. Th ese won- derful colleagues have been a real pleasure to work with, and we wish them all the very best in their future endeavours. We owe par- ticular thanks to colleagues at the Institute for Healthcare Improve- ment, Cambridge, USA* who contributed graciously to this book. Any omissions during the editing phase are ours.

We would also like to take this opportunity to thank our fami- lies for their support throughout the conception, gestation and delivery of this book. Th is work is therefore very much also a fruit of their labours, and we hope that they too will take pride in seeing the ideas contained in this book fl ourish.

*Th e Institute for Healthcare Improvement (IHI) (www.IHI.org) is an independent not-for-profi t organization which hosts the IHI Open School (www.ihi.org/openschool).Th e School exists to advance quality improvement and patient safety competencies in the next generation of health professionals.

xiii

How to use your revision guide

Features contained within your revision guide

Each topic is presented in a double-page spread with

clear, easy-to-follow diagrams supported by succinct

explanatory text.

C hapter 9 R

oot cause analysis

Patient Safety and Healthcare Improvement at a Glance, First Edition. Edited by Sukhmeet S. Panesar, Andrew Carson-Stevens, Sarah A. Salvilla and Aziz Sheikh © 2014 John Wiley & Sons, Ltd. Published 2014 by John Wiley & Sons, Ltd.

What is root cause analysis? Root cause analysis (RCA) is a method of incident investigation.

As such, it is a diagnostic tool rather than a safety solution in itself. RCA allows a systems approach (Chapter 26) to investigation and was selected as the methodology of choice by the National Patient Safety Agency when developing a framework for patient safety investigation in the NHS. The NHS approach aligns well with investigation methods used in healthcare and other high-risk industries across the globe.

Why investigate? The primary aim of patient safety investigation is to learn from inci- dents and to determine what can be done to significantly reduce the likelihood of recurrence; the aim is not to apportion blame.

If, during an investigation, concerns of capability, recklessness or maliciousness arise, the Incident Decision Tree (IDT) should be used to provide guidance on whether and to whom these issues should be referred. Investigation and planned management of these particular concerns should not form part of the patient safety investigation process.

RCA process Investigations can be comprehensive or concise but must always

include the basic elements to help ensure they are thorough, cred- ible and actionable, and represent value for money

Set clear terms of reference and follow them. Secure adequate time and skills, or record and report the impact of constraints

Avoid lots of concise investigations. They can prove false economy 1 Gathering and mapping the information

You have to understand exactly what happened leading up to an incident before you can fully understand why it happened

Investigative interviewing should focus more on listening than on asking questions

Consult the patient and family as part of the investigation; they have a unique perspective and valuable information to share

2 Identifying care and service delivery problems (CDPs and SDPs) – this stage involves identifying all the points at which:

something happened that should not have happened; or something that should have happened did not

3 Analysing problems Using a fishbone diagram (or Ishikawa diagram or cause-

and-effect diagram) as shown in Figure 9.2, place one CDP or

SDP in the head of each fish (not the whole incident), then ana- lyse why that course of action seemed the right thing to do at that time

A few carefully analysed ‘fishbones’ focusing on key CDPs and SDPs will deliver more benefit than many completed quickly

Training in systems thinking and human factors (including error types and biases) will aid impartiality and quality analysis

The root causes are the most significant contributory factors 4. Generating recommendations and solutions

Problems will rarely be resolved for the long term by applying discipline, training and updated procedures alone

Training in improvement science will assist with more effective selection and implementation of solutions

5. Implementing solutions Amalgamate action plans from investigations. This encour-

ages trend analysis and a more cohesive, high-level approach to resolving common issues

Avoid conducting more and more investigations with similar outcomes. Time must be allocated to implementing solutions and monitoring their efficacy

6. Writing the Investigation Report Use an RCA investigation report template to facilitate trend

analysis, audit and shared learning

Effective RCA investigation The components for success in patient safety investigations are the same as those required for successful clinical investigations (Figure 9.1): 1 To avoid the extremes of delayed problem ‘diagnosis’ and resource wastage, triggers or indications for conducting an investi- gation must be correctly identified. 2 To obtain a good-quality, accurate picture of the problem, data gathering must be conducted by those skilled in the process. 3 The findings from the collection of data must be robustly inter- preted and credible conclusions drawn by someone with analyti- cal skills and an understanding of the ‘anatomy, physiology and pathology’ of the issue. 4 To ensure that improvement is achieved and measurable, expert selection, application and monitoring of effective treatment and remedial action are required. 5 If meaningful learning and improvement are expected from incident investigation, there must be organisation-wide support for this process.

Chapter 27 gives an example of a fishbone diagram in use.

9 Root cause analysis Figure 9.1 Steps to an effective root cause analysis (RCA) investigation

Figure 9.2 RCA investigation: fishbone diagram – tool

Good organisational safety culture (infrastructure, resource, support)

Appreciation of the value of,

rationale behind, and indications for

investigation

Appreciation of the value of,

rationale behind, and indications for investigation

Competence in thorough, credible

investigation analysis

Skilled in error wisdom, research,

interpretation and deduction

Skilled in improvement

science

Select and apply effective remedies/ solutions for each

cause ...

and monitor for success

Source: NHS National Patient Safety Agency

Equipment and resource factors

Communication factors

V Wr

Task factors

Individual (staff) factors

y y

r Co e

Patient factors

y or or

y or

Working conditions/ environment factors

A e y

f Wor T

Organisational and strategic factors

rg re

er y

Education and training factors

Team and social factors

+ or

Problem or issue

(CDP/SDP)

ce: NHS National Patient Safety Agency

P art 2 U

nderstanding and interpreting risk

Your textbook is full of illustrations and tables.

P art 4 Q

uality im provem

ent

23 Model for Improvement Figure 23.1 Model for Improvement Figure 23.2 Definition of improvement

Model for Improvement

W to

W ov

Act Plan

DoStudy

Source: Langley G, Moen R, Nolan K et al 2009. Reproduced with permission of John Wiley & Sons Ltd.

Table 23.1 How will we know a change is an improvement? Figure 23.3 The plan–do–study–act cycle

Act Plan

DoStudy omplet ysis

ise s

W to

to ,

ysis

Source: Langley G, Moen R, Nolan K et al 2009. Reproduced with permission of John Wiley & Sons Ltd.

Alter how the work is done ... Improvement is the result of some design or redesign of the system

2 Produce visible, positive differences in results relative to historical norms

And lasting impact

In fe

ct io

Week 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16

12 11 10 9 8 7 6 5 4 3 2 1

Before

Change

After

Source: Langley G, Moen R, Nolan K et al 2009. Reproduced with permission of John Wiley & Sons Ltd.

Decrease CAUTI rate

Decrease inappropriate catheter use

Maintain or increase staff satisfaction

Time between CAUTI rate

atheter days/100 patient days

se compliant to new process on review of catheter

e of the pre and post test

se satisfaction score ysician satisfaction score

Increase nurses’ compliance to new process on review of catheter use

Outcome

ocess

Balancing

Increase nurses’ knowledge on criteria for catheter use

Goals Measures Type of measure

The anytime, anywhere textbook

Wiley E-Text

Your book is also available to purchase as a Wiley E-Text: Powered by VitalSource version – a digital, interactive version of this book which you own as soon as you download it.

our Wiley E-Text allows you to:

Search: Save time by fi nding terms and topics instantly in your book, your notes, even your whole library (once you’ve downloaded more textbooks)

Note and Highlight: Colour code, highlight and make digital notes right in the text so you can fi nd them quickly and easily

Organise: Keep books, notes and class materials organised in folders inside the application

Share: Exchange notes and highlights with friends, classmates and study groups

Upgrade: Your textbook can be transferred when you need to change or upgrade computers

Link: Link directly from the page of your interactive textbook to all of the material contained on the companion website

The Wiley E-Text version will also allow you to copy and paste any photograph or illustration into assignments, presentations and your own notes.

To access your Wiley E-Text: • Visit www.vitalsource.com/software/bookshelf/downloads to download the

Bookshelf application to your computer, laptop, tablet or mobile device.

• Open the Bookshelf application on your computer and register for an account.

• Follow the registration process.

xiv

CourseSmart

CourseSmart gives you instant access (via computer or mobile device) to this Wiley-Blackwell e-book and its extra electronic functionality, at 40% off the recommended retail print price. See all the benefi ts at: www.coursesmart.com/students

Instructors … receive your own digital desk copies! CourseSmart also offers instructors an immediate, effi cient, and environmentally-friendly way to review this book for your course.

For more information visit www.coursesmart.com/instructors.

With CourseSmart, you can create lecture notes quickly with copy and paste, and share pages and notes with your students. Access your CourseSmart digital book from your computer or mobile device instantly for evaluation, class preparation, and as a teaching tool in the classroom.

Simply sign in at http://instructors.coursesmart.com/bookshelf to download your Bookshelf and get started. To request your desk copy, hit ‘Request Online Copy’ on your search results or book product page.

We hope you enjoy using your new book. Good luck with your studies!

Course Learn Smart. Choose Smart.

Smart

Chapter 1 Haemopoiesis / 3

cells are, however,r capable of responding to haemo- poietic growth factors with increased production of one or other cell line when the need arises.

e development of the maturerr cells ,sllecder( granulocytes, monocytes, megakaryocytes and lym- phocytes) is considered further in other sections of

.koobsiht

Bone m arrow s troma e bone marrow forms a suitable environment for

fonoitamrofodnalawener-f-les,lavivrusllecmets erentiated progenitor cells. It is composed of

.)4.1.giF(krowtenralucsavorcimadnaslleclamorts

tseilraeehtsielpmaxenA.aidemdilos-imesni detectable mixed myeloid precursor which gives rise to granulocytes, erythrocytes, monocytes and meg-

gnimrofo-ynoloc(UFCdemretsidnasetycoyraka e bone marrow is also

the primary site of origin of lymphocytes (see erentiate frff om a common

.rosrucerpdiohpmyl e stem cell has the capability for self-ff renewalrr l

-nocsniamerytiralullecworramtahtos)3.1.giF( ere is con-

cation in the system: one stem cell is capable of producing about 106 doolberutam

e precursor

Figure 1.2 Diagrammatic representation of the bone marrow pluripotent stem cell and the cell lines that arise from it. VariousVV progenitor cells can be identified by culture in semi -solid medium by the type of colony they form. It is possible that an erythroid/megakaryocytic progenitor may be formed before the common lymphoid progenitor diverges from the mixed granulocytic/monocyte/eosinophil myeloid progenitor.rr Baso, basophil; BFU, burst-forming unit; CFU, colony -forming unit; E, erythroid; Eo, eosinophil; GEMM, granulocyte, erythroid, monocyte and megakaryocyte; GM, granulocyte, monocyte; Meg, megakaryocyte; NK, natural killer. rr

Pluripotent stem cell

Erythroid progenitors

CFUGEMM Common myeloid progenitor cell

BFUE

CFUE

CFUMeg Megakary- ocyte progenitor

CFUGM Granulocyte monocyte progenitor

CFUEo Eosinophil progenitor

CFUGMEo

CFUbaso

Thymus

CFU-M CFU-G

Common lymphoid progenitor cell

Red cells

Platelets Mono- cytes

Neutro- phils

Eosino- phils

Baso- phils

LymphocyteLL s NK cell

B T NK

The VitalSource Bookshelf can now be used to view your Wiley E-Text on iOS, Android and Kindle Fire!

• For iOS: Visit the app store to download the VitalSource Bookshelf: http://bit.ly/17ib3XS

• For Android: Visit the Google Play Market to download the VitalSource Bookshelf: http://bit.ly/ZMEGvo

• For Kindle Fire, Kindle Fire 2 or Kindle Fire HD: Simply install the VitalSource Bookshelf onto your Fire (see how at http://bit.ly/11BVFn9). You can now sign in with the email address and password you used when you created your VitalSource Bookshelf Account.

Full E-Text support for mobile devices is available at: http://support.vitalsource.com

7 Research in patient safety Figure 7.1 Framework for patient safety research

Patient safety research

Domain 5

Evaluating the association between

organisational characteristics and outcomes

Evaluate organisational characteristics that help or hinder

research efforts or patient safety practices

Domain 4

Identifying and mitigating hazards

Use of retrospective and prospective analyses to identify and mitigate

safety hazards at the microscopic level

Domain 3

Assessing and improving culture

Strategies and interventions to improve safety culture and

communication

Domain 2

Translating evidence into practice

Develop and evaluate interventions that increase the extent to which

patient received evidence- based medicine

Domain 1

Evaluating progress in patient safety

Develop valid and feasible measures to evaluate progress to improve

patient safety

P art 1 T

he essence of patient safety

Figure 7.2 TRIP model

En ga

Ed uctate Execute Evaluate

1. Summarise the evidence

4. Ensure all patients receive the interventions

3. Measure performance

2. Idenfify local barriers to implementation

Introduction After the To Err Is Human report uncovered the magnitude of the problem of patient safety incidents (more than 98,000 deaths per year due to medical errors), another report by RAND in 2003 revealed that hospitalised patients in the United States received only half of the recommended therapies. The framework for patient safety research and improvements (PSRI) described in this chapter evolved out of the need to bridge the gap between the interventions being implemented and scientific assessment of their success and applicability to similar settings.

The framework has five domains (Figure 7.1): 1 Evaluating progress in patient safety 2 Translating evidence into practice 3 Assessing and improving culture 4 Identifying and mitigating hazards 5 Evaluating the association between organisational characteris- tics and outcomes.

Besides these domains, tools such as simulation, health information technology, quantitative data analysis and others are useful in PSRI. Research and improvement must go hand in hand in order to initiate and sustain improvements. Hospitals must address both the techni- cal work and the adaptive work – the former involves known solu- tions and science, and the latter requires changes in values, attitudes

or beliefs to sustain improvement. A research study could be designed to address one or more of the above domains measuring or evaluating either or both of the technical and adaptive aspects, depending on the scope, timeline and access to data for conducting the study. In a col- laborative team project, the centralised research team would do the technical work and the local team would do the adaptive work.

Evaluating progress in patient safety Measures of patient safety involve two balancing acts: 1 Balancing the desire of a global, although more biased, measure of safety versus a more focused, but less biased (robust), measure. A global measure applicable to all patients (e.g. hospital mortality) has extreme bias due to inadequate risk adjustment and accounting for patient preferences. A specific measure (e.g. central line bloodstream infections) is very robust but targets only a subset of patients. Many specific measures are needed to cover the whole patient population. 2 Finding a balance between a measure that is scientifically sound (valid and reliable) and feasible given the limited resources. Use of relatively easy and inexpensive data sources, such as administra- tive data for measures such as deep venous thrombosis, is feasible but correlates poorly with medical chart review data. In order to address these, it is necessary to:

(a) Reduce the quantity but not quality of data. (b) Consider the validity of a measure at two levels: a. Patient safety domain: If it is an outcome, does it represent an impor- tant aspect of quality, and do either variation in practice among organisations or interventions that improve the outcome demon- strate that it is largely preventable? If it is a process measure, does evidence suggest that the intervention will improve outcomes? b. What study design is used to measure the patient safety domain? Are there well-defined research protocol, data collection tools, well-designed databases, clear quality control plans, and detailed analytic plans? Cluster-randomised designs, a stepped wedge-trial design, or a quasi-experimental (time series) design can be used. As most studies tend to be a pre-post design, it is important to adjust for historical bias or changes in performance over time.

Part 1The essence of patient safety

Chapters 1 Basics of patient safety 2 2 Understanding systems 4 3 Quality and safety 6 4 Human factors 8 5 Teamwork and communication 12 6 Reporting and learning from errors 14 7 Research in patient safety 16

P art 1 T

he essence of patient safety

1 Basics of patient safety Table 1.1 Patient safety terms

Harm

Near miss

Adverse event (AE)

Preventable adverse event

Adverse drug event (ADE)

Patient safety incident (PSI)

Critical incident*

Definition

Any physical or psychological injury or damage to the health of a person, either temporary or permanent Harm is usually classified as no harm, low harm, moderate harm, severe harm or death.

Any patients safety incident that had the potential to cause harm but was prevented, resulting in ‘no harm’ (although this is a term of variable definition).

An event involving unintended harm to a patient that resulted from medical care. Traditionally, the term used for an adverse event was ‘iatrogenesis’.

An event involving patient harm as a result of wrong or inappropriate action (‘error of commission’) or failing to do the right thing (‘error of omission’).

Any incident in which the use of a medication (including prescribed drugs, but also dietary supplements) results in harm to a patient. ADEs include adverse drug reactions (i.e. known side effects that occur even when the medication is used as intended), as well as events in which the drug has been used erroneously (prescribed at the wrong dose, administered in the wrong way etc.). ADEs that result from medication errors are often called ‘preventable ADEs’.

Any unintended or unexpected incident that could have harmed or did harm the patient. This includes ‘near misses’. The term ‘patient safety incident’ is preferred to ‘error’, as the latter has a more negative connotation.

A term first coined in the 1950s and made famous by a classic human factors study by Cooper of ‘anaesthetic mishaps’. Cooper and colleagues brought the technique of critical incident analysis to a wide audience in healthcare, and followed the definition of the originator of the technique. They defined critical incidents as occurrences that are ‘significant or pivotal, in either a desirable or an undesirable way’. Cooper and colleagues (and most others since) chose to focus on incidents that had potentially undesirable consequences. This concept is best understood in the context of the type of investigation that follows, which is very much in the style of root cause analysis. Thus, significant or pivotal means that there was significant potential for harm (or actual harm), but also that the event has the potential to reveal important hazards in the organisation. In many ways, it reflects an expression used in quality improvement circles: ‘every defect is a treasure’ . In other words, these incidents, whether near misses or disasters in which significant harm occurred, provide valuable opportunities to learn about individual and organisational factors that can be remedied to prevent similar incidents in the future.

Patient safety term

*Source: Cooper et al 1978. Reproduced with permission of Wolters Kluwer Health.

Figure 1.1 Frequency of errors in medical care (adverse event rate) Figure 1.2 The Swiss cheese model of how defences, barriers and safeguards may be penetrated by an accidentaly trajectory

USA (3.2–5.4%)

UK (11.7%) Denmark

(9%)

Australia (10.6–16.6%)

New Zealand (12.7%)

Latin America

(10%)

Source: Adapted from de Vries et al 2008. Reproduced with permission of BMJ Publishing Group Ltd. Source: Reason J 2000. Reproduced with permission of BMJ Publishing Group Ltd .

Hazards

Losses

3 C

hapter 1 B asics of patient safety

Introduction As healthcare has become more eff ective, it has also become more complex and involves the use of new technologies, medicine and treatments. We are also now treating a greater proportion of older and sicker patients. Th ese factors, coupled with decreased fi nancial resources in most settings, can result in errors.

Two infl uential reports – To Err Is Human (1999) produced by the US Institute of Medicine and An Organisation with a Memory (2000) produced by the UK Government’s Chief Medical Adviser – heralded the start of the global patient safety movement in the late 1990s. Both reports recognised that error was common during the delivery of healthcare: Figure 1.1 gives estimates of harm globally in hospitals. Th e fi gure of 1 in 10 patients being harmed is com- monly quoted in the world of patient safety.

Th e reports drew attention to the poor performance of health- care, as a sector, worldwide on safety compared to most other high- risk industries. Notably, aviation has shown remarkable and sus- tained improvements in levels of risk to air travel passengers over the last four decades. Both reports called for greater focus on, and commitment to, reducing risks in healthcare. In October 2004, the World Health Organization (WHO) launched a patient safety pro- gramme, in response to a World Health Assembly Resolution (2002) urging WHO and member states to pay the closest possible atten- tion to the problem of patient safety. Its establishment underlined the importance of patient safety as a global healthcare issue. In other countries, specifi c bodies dealing with patient safety were set up: the National Patient Safety Agency (NPSA), which is now part of NHS England; the Agency for Healthcare Research and Quality (AHRQ) in the United States; the Canadian Patient Safety Institute (CPSI); and the Australian Commission on Safety and Quality in Health.

Despite these notable eff orts, the current state of patient safety worldwide is still a source of deep concern. As data on the scale and nature of errors and adverse events have been more widely gath- ered, it has become apparent that unsafe actions are a feature of virtually every aspect of healthcare. Furthermore, there is a paucity of research on the frequency of errors and their associated burden of harm in areas such as primary care and mental health. Reports of the deaths of patients regularly feature in media reports in many countries and undermine public confi dence in health services. Moreover, many events recur, with eff orts to prevent them ineff ec- tive. Th ese could be in part due to a punitive culture of individual blame and system failures. Initial, widely quoted estimates of the number of deaths due to medical error may have been exagger- ated. For instance, a study by Hogan et al. (2012) of 1000 deaths at 10 representative UK NHS trusts found that only 5% were judged preventable, with ‘preventable’ being defi ned as having a greater than 50% probability that better care would have prevented death.

Th ere is also growing concern of late amongst patient safety experts that despite all the eff orts made to date, the patient safety momentum might stall as we have been at it for almost a decade and countless initiatives have been thrown at clinicians who may be overwhelmed.

Defi nitions ‘Patient safety’ can be defi ned as reducing the risk of unnecessary harm associated with healthcare to an acceptable minimum. An ‘acceptable minimum’ refers to current knowledge, resources avail- able and the context in which care was delivered, weighed against the risk of non-treatment or other treatment. Simply put, it is the prevention of errors and adverse eff ects to patients associated with healthcare. Further key defi nitions are given in Table 1.1.

Concepts Th e large-scale technological disasters on oil rigs, nuclear power plants and aviation in the 1980s led to more of a systems-thinking approach to developing safer workplaces and safer cultures. Th e same approach applies to healthcare; it is rare that a doctor or nurse is to blame for an error, but the environment and systems they work in play a strong part. James Reason, an eminent psychol- ogist, developed the ‘Swiss cheese’ model (see Figure 1.2) to explain the steps and multiple factors associated with adverse events. Key points to note in this model are: • Defences, barriers and safeguards exist to protect patients from hazards, such as alarms on syringe drivers or anaesthetists remind- ing surgeons to ensure that an adequate pre-operative work-up of the patient has taken place. Th ese defences can be breached, like the holes in slices of Swiss cheese. However, unlike in the cheese, these holes are continually opening, shutting and shift ing their location. Th e presence of holes in any one ‘slice’ does not normally cause a bad outcome. Usually, this only happens when the holes in many layers momentarily line up to permit a trajectory of acci- dent opportunity – bringing hazards into damaging contact with patients. Th e holes occur due to a combination of active failures and latent conditions • Active failures are the unsafe acts committed by people who are in direct contact with the patient or system. Th ey take a var- iety of forms: slips, lapses, fumbles, mistakes and procedural violations • Latent conditions arise from decisions made by designers, build- ers, procedure writers and top-level management. Th ey can trans- late into error-provoking conditions within the local workplace (e.g. understaffi ng requiring the use of locum doctors). Th ey can also create long-lasting holes or weaknesses in the defences (e.g. the intensive care unit being in a diff erent building from the oper- ating theatre)

Another notable individual, Jens Rasmussen, suggested that errors occurred due to defi ciencies in skills (e.g. asking a junior doctor to perform a laparotomy), observation of rules (e.g. not washing hands before performing a procedure) or knowledge (e.g. being unaware that gentamicin levels need to be checked).

Subsequent chapters will build on the concepts discussed here and equip the reader with the knowledge to identify and rectify potential threats to patient safety.

P art 1 T

he essence of patient safety

2 Understanding systems Figure 2.1 Healthcare as a complex socio-technical system and system accidents

The delivery of healthcare relies on a complex and wide array of people, activities and technologies

Each of these present opportunities for error, which impact on other parts of the system

Major accidents result from a combination of minor errors spread around these organisational systems

• Protocols • Policies • Procedures • Checklists

• Information technology • Communication systems • Record systems

• Designers • Data analysts • Support staff

• Infrastructure

• Local workplace

• Ward design

• Buildings and estates • Engineers

• Maintenance • Architects

• Universities

• Funders

• Professional bodies

Poorly designed

ward

Mis- specified

policy

• Regulators

• Regulators • Oversight agencies • Professional bodies • Policymakers • Insurers • Commissioners • Educators and trainers

• Procurement • Purchasers • Designers • Manufacturers

• Mistake during manufacture

• Medical equipment • Consumables • Supplies

• Managers • Administrators • Boards • Executives • Safety officers • Clinical leads, etc.

• Doctors • Nurses • Physios • Healthcare assistants

Equipment hard to use

Error using

equipment

Adverse event

• Radiographers • Paramedics • Allied health professionals, etc.

Healthcare professionals:

• Government

Patients

Introduction Modern healthcare organisations are enormously complex (see Figure 2.1). Even the most routine tasks of healthcare now depend on complex systems that connect a multitude of people, activities and technologies. For example, a typical patient in an intensive care unit requires 178 separate actions to be performed for them each day by a range of people, and a patient in their last year of

life will typically be treated by at least 10 specialist doctors. Th e work of any individual healthcare professional is equally depen- dent on a wide range of other factors. Th ese factors include every- thing from eff ective communication with colleagues to the staffi ng levels and resources within an organisation; and from the quality of equipment available locally to the design of wards and infor- mation technology (IT) systems. All of these factors and many

5 C

hapter 2 U nderstanding system

s more contribute to the safety of patients in any given situation. To manage and improve patient safety, it is essential to understand the nature of the complex systems that deliver healthcare, along with the ways that these systems shape – and are shaped by – the work of individuals. Safety improvement eff orts need to be targeted at improving these systems. As human factors expert James Reason has put it: ‘We cannot change the human condition, but we can change the conditions under which humans work’.

From individuals to systems Th ere has been a dramatic shift over the past two decades in how patient safety and error are understood in healthcare. Traditionally, if things went wrong, any investigation and remedial action would focus on the actions of individuals: typically whoever was closest to the adverse event at the time, such as the surgeon who oper- ated on the wrong site or the nurse who miscalculated the drug dose. Th is view has now been replaced by a focus on the safety and reliability of the broader systems within which individuals work. ‘Systems thinking’ has long been established in other safety-critical industries such as aviation. In these industries, decades of detailed investigations have revealed that major accidents are the result of a combination of minor mishaps, inadequacies and errors that occur throughout organisational systems. Th e actions of the individuals who happen to be ‘closest’ to the adverse event are oft en merely the last link in a very long chain of events. Th is systems thinking underpins the entire fi eld of patient safety, and was largely popu- larised in healthcare through two groundbreaking reports: To Err Is Human and An Organisation with a Memory.

Healthcare as a complex socio-technical system Healthcare is a complex socio-technical system, in which even apparently simple tasks can depend on a wide range of social (e.g. psychological, team and managerial) and technical (e.g. equipment, IT and infrastructure) factors. For example, prescribing a medi- cation depends on things such as IT systems that allow access to patient records, communication systems that allow eff ective trans- fer of information between health professionals, purchasing systems that ensure the pharmacy is properly stocked, education systems that ensure health professionals are appropriately trained and reg- ulatory systems that monitor the safety and eff ectiveness of medi- cines. One defi ning feature of complex socio-technical systems is that there are many components and subsystems that must interact with each other to achieve a certain outcome. Th e eff ective inter- action of each of the systems outlined here is essential to the safe prescribing of medications, and each is also a complex system in itself that requires careful management and design. Another defi n- ing feature of complex systems is that many of these components and subsystems are hidden from people working elsewhere in the system. Th e doctor who writes a prescription is unlikely to know much about the purchasing processes in the pharmacy, and yet all of these systems must function eff ectively together to provide safe care.

System reliability Managing the reliability of systems – the ability of a system to routinely perform its function without failure – is a key factor in improving patient safety. Measures of reliability suggest that many

systems in healthcare organisations operate at around 80% reliabil- ity. Th is is an extraordinarily low level. For comparison, if a car was 80% reliable, it would only work 4 days out of 5. Large commercial jets attain on-time reliability rates of around 99.5%. Low levels of system reliability in healthcare include: • Systems that provide patient information for clinical decision making in surgical outpatient clinics operate at around 85% reliability • Within prescribing systems for hospital inpatients, around one in seven prescriptions contain an error. One in fi ve errors would have had serious consequences if not corrected • Processes for ordering surgical theatre equipment operate at around 80% reliability. Half of these failures result in equipment being entirely unavailable

Much work needs to be done to make healthcare systems as reliable as those in other industries. Th is work is advancing rapidly, but it remains at an early stage. Th e Institute for Healthcare Improvement (IHI) uses a three-step model for applying principles of reliability to healthcare systems: • Prevent failure • Identify and mitigate against failure • Redesign the process based on the critical failures identifi ed

Organisational accidents A particular challenge of complex systems is that they can suff er complex and serious breakdowns – ‘organisational’ or ‘system acci- dents’. Minor errors and mishaps in one area of a system are not confi ned to that area, but can impact activities in other areas, oft en in unexpected and dramatic ways. Small mistakes can be ampli- fi ed and have disproportionate eff ects elsewhere in the system. For example, a simple decision about adding a cleaning agent to a hospital’s water system can have severe knock-on eff ects. In 2008, in the United Kingdom, one haemodialysis patient died and four others required blood transfusion aft er a cleaning agent was added to the hospital’s main water supply. It had not been fully realised or properly communicated that the renal unit’s water fi lters could not remove that particular chemical, which passed straight through into the patients’ bloodstream.

Healthcare organisations increasingly rely on a variety of safety defences and controls to assure system safety. Th ese defences aim to prevent errors cascading and aggregating throughout a system in ways that might cause a major system accident. However, these defences themselves can add complexity to the system, which in turn can introduce new risks. Th e water treatment process described here was itself being undertaken to address the safety risks of water-borne infection within the hospital. Safety defences themselves can sometimes introduce new risks, and safety improvements must be carefully designed and assessed from a sys- tem perspective to reduce this possibility. Healthcare depends on a complex, socio-technical system that can be challenging to fully analyse and understand. Improving patient safety requires a deep understanding of the many system interactions and system fac- tors that produce both good and bad outcomes for patients. Every point of potential failure and error is also an opportunity for safety improvement and system redesign.

3 Quality and safety Figure 3.1 Systemic constructs of quality and safety

Timeliness Effectiveness Safety

Caring

Patent centredness

8 dimensions of quality

Continuity of care

Safety is a fundamental system property

A safe organisation is a cost-effective quality organisation

Systemic threats to safety

• Complexity of the process of care

• Variability from patient to patient

• Inconsistency in the standards of care

• Poor interfacing (e.g. transition between settings)

• Lack of error-preventing barriers

• Lack of initiative to handle the unforseen

• Use of inappropriate time constraints

• Use of hiearchical culture in te system

• Human fallibility – to err is human

Assurance function

Donabedian’s house of quality triad:

structure-process-outcome

E ffi

ca cy

E ff

ec ti

ve n

es s

E ffi

ci en

O p

ti m

al it

A cc

ep ta

b ili

L eg

it im

ac y

E q

u it

Fl yi

ng b

ut tre

ss es

Flying buttresses

1. In

te rd

ep en

de nc

2. O

rg an

is at

io na

l d ep

en de

nc y

3. C

on se

ns ua

lit y

4. C

on gr

ue nc

5. C

re di

bi lit

6. R elevance

7. O w

nership

8. M utuality of interests

9. Facilitation

10. C oerciveness

11. Virtue – personal and public

Donabedian’s seven pillars and eleven buttresses of quality

Protec tive roof of quality

Lack of awareness of and attention to the above may lead to adverse events

Without safety there can be no quality of care

Equity Efficiency

(a) (b)

(c)

P art 1 T

he essence of patient safety

7 C

hapter 3 Q uality and safety

Introduction Th e history of safety and quality started with the ‘fi rst do no harm’ call from Hippocrates and the call for hygiene from Florence Nightingale. Despite various policies and initiatives, the progress towards improvement in this fi eld has been unsatisfactory and demonstrates a need for systems thinking.

What is quality healthcare? Th e quality of any healthcare setting is its total system characteristic. It is about doing the right thing (or things), for the right patient, at the right time, with the best results and at aff ordable costs. Quality has eight dimensions (Figure 3.1.a). A quality, cost-eff ective organ- isation springs only from a safe organisation. Overuse, underuse, misuse and the practice of economy with truth (fraud) are com- mon challenges on the journey to better quality.

What is safety? Th is too is a fundamental system property (Figure 3.1.b). Without safety, there can be no quality of care. It is one of the world’s most pressing healthcare challenges. ‘Safety’ can be defi ned as freedom from avoidable injuries. Its goal is to avoid, prevent and amelio- rate adverse outcomes emanating from the care processes. James Reason’s trajectory of errors is an excellent aid to safety improve- ment because it helps in understanding the causes of failures in the form of situational (e.g. a very unusual work load, or a power sup- ply failure at a critical juncture), latent (e.g. defi ciencies in design, operation, maintenance, organisation and management) and active failures (e.g. human fallibility) that can result in adverse events, in the absence of appropriate technical (e.g. the use of informatics and safe-dosage packaging) and administrative (e.g. standard protocols and non-hierarchical team culture) barriers to this trajectory.

Similarities and differences between quality and safety A quality healthcare setting has dimensions of safety, eff ective- ness, timeliness, effi ciency, equity, patient-centredness, caring and care continuity. Th ese can be seen as being housed in a protective Donabedian ‘house of quality triad’, which encompasses ‘structure, process and outcome’ (Figure 3.1c). A safe healthcare setting, whilst being an indispensable and vital dimension of quality, is designed to face systemic threats of complexity in the process of care, vari- ability from patient to patient, inconsistency in the standards of care, poor interfacing (e.g. transition between settings), lack of error-preventing barriers, lack of initiative to handle the unfore- seen, the use of inappropriate time constraints, hierarchical cul- ture in the system and human fallibility. Awareness of these threats leads to the ability to design and manage systems with barriers to prevent errors reaching patients.

The Donabedian framework Th e triad of structure ↔ process ↔ outcome is a very helpful clas- sifi cation scheme for quality evaluation (Figure 3.1c). Structural quality evaluates healthcare system capacities, how the system is confi gured, and its components and their inter-relationships. Organisational culture and stakeholder satisfaction are also important elements. Process quality assesses interactions between patients and clinicians, as well as how care is delivered. Th e best process measures should be based on evidence relating better pro- cesses to better outcomes (e.g. controlling blood pressure reduces strokes and heart disease). Outcomes quality assesses changes in the health status of the patients and patient satisfaction. Th e best outcomes measures are those that are tied to processes over which the healthcare system has infl uence (e.g. the survival rate of pancreatic cancer is not a reliable measure as there is a lack of meaningful treatments aff ecting survival). Donabedian off ered 11 essential principles (buttresses) to support the design, operation and eff ectiveness of the quality-assuring ‘dome’, supported also by seven pillars of quality (Figure 3.1c).

More recently, the Institute for Healthcare Improvement uses the term ‘triple aim’ to foster improvement at a systems level in three areas: improving the individual experience of care, improv- ing the health of populations and reducing the per capita costs of care for populations.

Approaches to improvement Because healthcare organisations regularly face new challenges, they must be adaptive to improve continually. Quality improve- ment in any setting is a systematic, data-informed activity designed to bring about improvement in healthcare delivery. All improve- ment approaches must meet three basic needs: (i) the creation of a culture of safety (Chapter 10) and a high-reliability organisation, (ii) the acknowledgement and treatment of each setting as a unique microsystem and (iii) the facilitation of workfl ow, processes and task assessment and improvement. Th ere are retrospective and pro- spective methods of assessment for management of improvement. Widely used retrospective methods include error reports (root cause analysis, or RCA) (Chapter 9), internal and external audits, quality and safety indicators (Chapter 8) and trigger tools. Each one of these reveals only the tip of the iceberg of the quality gap and diff erent perspectives of the same reality. Trigger tools reveal a much bigger tip, but are cost and time intensive. Generalisations of the results from retrospective methods can lead to stakeholder dis- satisfaction. Th ese methods tend to be top-down and do not fully meet the needs expressed in this chapter. Th e prospective approach is based on the failure modes and eff ects analysis (FMEA), as against RCA used in retrospective methods. FMEA has been widely used in other high-risk industries and has been advocated by the Institute of Medicine as a means of analysing a system to identify its failure modes and possible consequences of failure (eff ects) and to prioritise areas for improvement.

4 Human factors Figure 4.1 Systems Engineering Initiative for Patient Safety (SEIPS)

Organisation • Throughput versus cost versus quality • Safety culture analysis • Resilience engineering • Accident investigation

Environment • Noise, light and heat • Interruptions and distractions • Workspace layout • Geographical distribution

Technology • Product design • Human machine interface • Procurment and integration • Technology surprises and risks

Tasks • Task design • Error analysis • Error prediction • Direct observation

People • Selection, training and assessment • Teamwork and non-technical skills • Decision making and situational awareness

P art 1 T

he essence of patient safety

9 C

hapter 4 H um

an factors Introduction Th e study of ‘clinical human factors’ is defi ned as ‘enhancing clinical performance through an understanding of the eff ects of teamwork, tasks, equipment, workspace, culture, organisation on human behaviour and abilities, and application of that knowledge in clinical settings’.

Th e study of human factors, or ergonomics, examines the rela- tionship between people and systems in order to build the working world around what people do well, rather than around technol- ogy or processes. By placing humans at the centre of our system of work, we can set about providing the best environment that will allow our best clinicians to perform to the highest level of their ability, or our least able clinicians (and there will always be a ‘least able’) to deliver care eff ectively.

This way of thinking originated in the 1940s and 1950s, when it was realised that the design of the displays and controls in aircraft could influence a crash, and thus sometimes made a difference between life and death. This understanding has been applied successfully to most high-risk industries and many con- sumer products. Good human factors enhance our interactions with the world so naturally that they are sometimes invisible.

The ‘Swiss cheese’ model Perhaps the best known theory in human factors is the ‘Swiss cheese’ model (Chapter 1) of accidents where predisposing ‘latent’ factors can pile up on each other in unique situations to cause acci- dents, injuries and deaths. It is not adequate to blame lone indi- viduals for making errors that lead to such catastrophes. Rather, we need to understand the predisposing factors that were always there, lying dormant in our system and only becoming critical when all these ‘holes in the cheese’ lined up.

Th is view helps us to understand that: • humans are not the ‘cause’ of accidents • humans hold complex and defi cient systems together • humans create safety in complex systems • accidents have their roots in organisations, not individuals • complex, high-risk systems are inherently unsafe • accidents signify problems deep in the system • problems can be visible but may seem innocuous • accidents happen when problems combine

Th us, putting the understanding of people at the centre of the system provides a new way of thinking about how healthcare in the future might be better confi gured and delivered. Th e Systems Engi- neering Initiative for Patient Safety (SEIPS) model developed by Carayon and colleagues illustrates the system parameters that can infl uence human performance, and ultimately patient outcomes (Figure 4.1). Th e SEIPS model involves people, tasks, technology, environment and organisation, as explained in this chapter.

People People are at the centre of any system, and they must exist at some level in the system to: • design, operate and maintain technology or processes • perform the key decision-making tasks • work in teams to support each other • circumvent poor processes • avoid potential errors • trap errors as they happen • mitigate the eff ects of those errors

People prevent catastrophic failures and provide the key resources in the system that no machine or process can do. A greater understanding of what people do within our systems can contribute to training, or indeed help us design systems for better human performance. For example, the eff ect of fatigue on a range of human abilities has been studied for many years. We also know that performance at a task can vary with workload, where being ‘under-loaded’ can be as detrimental to performance as being ‘overloaded’.

Human decision making is also of particular interest in human factors. In contrast to views of the human as a linear ana- lytical information processor, we know that human decisions are not always linear, analytical or logical. Decisions may be based on situational factors (what is happening now?), system-wide decisions (what should I be doing?) and the needs of individual patients.

If we wish to improve how people contribute to the perfor- mance of the system, we can improve our ability to select the right person for the job by improving our training regimes or enhancing our assessment systems.

Non-technical skills and situational awareness Th e introduction of non-technical skills training and assessment for surgical skills development can help to overcome some of the teamwork and communication problems that can predispose to surgical errors. Th ese are grouped into: • Social skills

• Leadership and management • Teamwork and co-operation

• Cognitive skills • Problem solving and decision making • Situational awareness Situational awareness describes how we notice information in

our environment, understand what it means within the context we are working in, and are able to project into the future about where we will be. Th is emphasises the ability of experts to be able to accu- rately predict what may happen to a patient (e.g. deterioration) so the response can be timely and appropriate.

Tasks Tasks defi ne what we need to achieve a goal. For example, to get cash from a hole-in-the-wall, we need to put our card into the machine, type our security code, take our card back and get our money. How experienced we are with this task and how clearly it is set out will infl uence our ability to perform it quickly and accu- rately. In some instances, changing the order of the tasks can make a diff erence. For example, if the cash is returned before the card, there is a much higher chance of forgetting the card, since we have completed the primary task (getting cash) before we have com- pletely fi nished.

To understand tasks, we can: • use hierarchical task analysis to describe tasks that users need to perform to achieve a goal • use human reliability analysis techniques and failure modes and eff ects analysis (FMEA) to predict the likelihood and conse- quences of making an error • perform a direct observation to see what people do • examine the diff erence between what is supposed to happen, what people say happens and what really happens

Once risks and barriers to the completion of processes are understood, it is then possible to redesign tasks to make them faster, more effi cient, safer or less error prone. It is also possible to provide methods to assist with tasks, such as: • sign-posting key processes to make tasks easy to do • checking to ensure that errors are captured • using standard methods to perform tasks consistently • using checklists to aid technical processes • holding briefi ngs and debriefi ngs to aid team processes (Chapter 5)

Technology and tools A tool might be something as simple as a checklist or a pencil and paper, but might also be a complex imaging system or a device such as a surgical robot. All such technologies can assist the people at the centre of the system in performing tasks that will help reach their goals. Th e appropriate application of technology and tools can make work faster, safer and more effi cient. However, technol- ogy that is not designed with the end users in mind can be burden- some and have the potential not to reduce errors, but to relocate them. Indeed, technologies are frequently introduced to increase effi ciency and safety, but have the opposite eff ect. Common sur- prises with new technologies are: • it doesn’t replace the need for humans • it requires them to work in diff erent ways • it requires them to work longer, harder or faster • new skills are required • diff erent errors are possible • diff erent people may be better at using the new technologies • it can de-skill people at the old tasks • there can be an over-reliance on the new technology

For example, the growth in laparoscopy requires additional skills aside from the traditional surgical techniques, which in turn means that diff erent people may be suited to the task. Although outcomes and patient experiences are generally better, it has also increased the chances of major complications such as bile duct injuries and, since open laparotomies are less practiced, they are more challenging when needed.

How the human interacts with the technology – known as the human–machine interface – can have huge implications for the

likelihood of errors. For example, using infusion pumps side-by- side, each with a diff erent design, will automatically generate the potential that the wrong buttons will be pressed and the machine will be set up in the wrong way. Training in the use of technol- ogy is also of signifi cant cost, and appropriate or common designs can reduce these costs and improve safety. Th us, buying equipment that is designed around human abilities can result in substantial benefi ts that will outweigh the initial purchase costs. Th is has been convincingly demonstrated in the defence industries.

Another challenge with technology is in integrating it with the existing work systems, in terms of both the new processes and the new tasks required to eff ectively use the equipment. Maintenance and upkeep of the technology are also key factors that may be over- looked. Recent evidence suggests that the employment of human factors considerations in reconfi guring equipment may yield sig- nifi cant performance and behavioural benefi ts.

Environment Th e study of human factors is also concerned with the eff ect of the working environment on human performance, such as: • Noise and lighting • Temperature and ventilation • Workspace and physical location

Noise can be disruptive to communication or thought patterns, and can even be damaging to the listener. However, music in some circumstances may enhance performance. Lighting is an impor- tant consideration for clinical work, not just in surgery, and excel- lent guidance is available for understanding the optimal lighting conditions for particular tasks. Temperature and ventilation con- siderations also impact human performance and infection rates.

Th e physical organisation of the workspace will also impact human performance. Items that are too high or too low, that do not have a place to be stored or that are in storage that is disor- ganised, inconsistent or poorly labelled will be more diffi cult to access, therefore reducing human ability. Rooms that are too small to house equipment, supplies, teams or other necessary require- ments of the task will increase risks and frustration, and reduce performance. Th ey may also increase infection risk.

Th e location where work is performed is also an important design consideration. Bedside handovers may involve the patient, but are prone to interruptions and may have privacy issues during confi dential discussions. Individual rooms improve patient pri- vacy but may require extra monitoring.

Organisation Th e fi nal component in the SEIPS model relates to the organisa- tion, which must support all aspects of the work environment. Considerations include: • safety culture • the balance between throughput, cost and quality • organisational leadership and management structure • preventing organisational drift • learning from safety incidents and achieving a fair culture

Culture is oft en described as ‘how we do it around here’. Th e assessment of safety culture can be particularly important in understanding the general consensus within an organisation about how people within it understand risk. A variety of tools are avail- able for assessing staff perceptions of the levels of risk and safety at diff erent levels within a hospital.

P art 1 T

he essence of patient safety

11 C

hapter 4 H um

an factors Cost and quality management A key feature of any organisation’s safety culture is the trade-off s that are required between throughput, cost and quality. Staff deliv- ering care are constantly required to decide on the right trade-off s to make, and the leadership in any organisation will help to set the priorities. For example, key performance targets and measures for fi nances and throughputs with unclear quality or safety goals will lead to increased throughput and reduced costs, but at the poten- tial expense of safety and quality. Making clear all three expecta- tions for staff is a clear aspect of managing these confl icting goals, and a realistic approach to setting those expectations is required.

An organisation that takes an active approach to monitoring and adjusting staff expectations in relation to these goals is more likely to be successful in managing these targets. An organisation that fails to recognise when safety is being traded for other aspects of system performance can be prone to ‘organisational drift ’, where safety standards become more and more relaxed until a serious tragedy occurs. In contrast, resilient organisations demonstrate resistance to these breaches in safety.

Learning from incidents

A carotid endarterectomy was underway with a 65-year-old male under local anaesthetic. Everything had gone smoothly and nothing appeared to be unusual. Th e consultant anaesthetist left theatre briefl y to use the telephone. While he was outside, the surgeon asked for the heparin to be given, and the anaesthetic registrar picked up the syringe from the workstation and gave the medication. A short time later, the consultant anaesthetist returned, and asked about the heparin. Th e registrar informed him that it had been given, at which the consultant expressed surprise, and pointed to the syringe that he had prepared with heparin before leaving theatre. Since the error was picked up before the cross-clamp was applied, and the injection had been of saline, no harm came to the patient.

Response to safety events will defi ne the ability of an organisation to learn, improve and avoid safety failures and adverse events in the future. Th is might be in the form of a debriefi ng (for minor events), analysis of incident reports (for a range of minor and more serious events) or a root cause analysis (RCA) (for serious events and harm). An organisation that blames individuals, sim- ply adds more checks or only requires staff to re-train in the wake of a serious incident is unlikely to have learned the appropriate lessons, increasing the chances of a repeat injury or other injuries. Looking deeper into a system can be painful, but an organisation that encourages a fair and open culture, where potential safety concerns are openly discussed and where a full systems analysis is conducted that avoids blame as far as possible (and includes consideration of all the components of the SEIPS model), stands a far better chance of improving safety and quality of care in the future.

Summary Th e study and application of human factors allow a diverse range of ways in which to consider and optimise human performance, safety, quality and effi ciency in any environment. Performance and safety are infl uenced by the: • people at the centre of the system • tasks they are required to perform • technology and tools they have to work with • environment in which they work • organisation in which they work

Th ese components also interact with each other – for example, technology changes tasks and may require diff erent training. Th is makes the study of any one system extremely complex. On the other hand, it provides an excellent opportunity to understand and enhance the delivery of clinical care. Th e application of this knowl- edge is known as ‘human factors’.

5 Teamwork and communication Figure 5.1 Case study – Elaine Bromiley

Figure 5.2 SBAR in action

Elaine Bromiley was a fit and healthy young woman who was admitted to hospital for routine sinus surgery. During the anaesthetic she experienced breathing problems and the anaesthetist was unable to insert a device to secure her airway.

After 10 minutes, it was a situation of ‘can’t intubate, can’t ventilate’, a recognised anaesthetic emergency for which guidelines exist. For a further 15 minutes, three highly experienced consultants made numerous unsuccessful attempts to secure Elaine’s airway and she suffered prolonged periods with dangerously low levels of oxygen in her bloodstream.

Early on, nurses had informed the team that they had brought emergency equipment to the room and booked a bed in intensive care, but neither were utilised. Thirty-five minutes after the start of the anaesthetic, it was decided that Elaine should be allowed to wake up naturally and she was transferred to the recovery unit. When she failed to wake up, she was then transferred to the intensive care unit.

Elaine never regained consciousness, and after 13 days the decision was made to withdraw the life support. Several lapses in human factors were noted:

Note: Video of the this incident can be viewed at http://www.chfg.org/resource/films-guides-articles

The stress of the situation meant that the consultants involved became highly focused on repeated attempts to insert the breathing tube. As a result of this, they lost sight of the bigger picture,

i.e. how long these attempts had been taking. This ‘tunnel vision’ meant they had no sense of time passing or the severity of the situation

In the absence of rehearsed strategies, actions were not in line with the emergency protocol.

In the pressure of the moment it's unlikely the team were able to consider options beyond the obvious single solution (which with hindsight was inappropriate)

There was no clear leader. The consultants in the room were all providing help and support but no one person was seen to be in charge throughout. This led to a breakdown in the

decision-making process and communication between the three consultants

Loss of situational awareness

Perception and cognition

Teamwork and communication

Situation

I am the foundation year 1 doctor

looking after Mrs T, who seems to be

deteriorating

Background

Mrs T. is a 70-year-old woman,

previously fit and well, who was

admitted 5 days ago with a lower respiratory tract

infection

Assessment

Her white cell count and other

inflammatory markers are

elevated. She is tachypnoeic, tachycardic,

hypotensive and pyrexial

Recommendation

I would really appreciate the

help of the medical registrar in managing the

patient. I will perform a septic

screen in the interim

P art 1 T

he essence of patient safety

13 C

hapter 5 Team w

ork and com m

unication Introduction A leading and recognised cause of medical failure is when team- work and communication break down. As care becomes more and more complex, healthcare providers must work in large teams, oft en working across multiple sites or settings. In healthcare, a signifi cant percentage of errors can be attributed to communica- tion failures and ineff ective teamwork. Th e Joint Commission has identifi ed both elements as the primary root cause in more than 70% of sentinel (never) events from 1995 to 2003. Despite the crit- ical roles that teamwork and communication play in the delivery of healthcare, professionals are not necessarily trained in these, and, as a consequence, a wide variation can occur in the qual- ity of professionals’ non-technical skills such as communication, situational awareness, decision making and teamwork. Eff ective communication and teamwork are essential for achieving high- reliability organisations and promoting a culture of openness and delivery of safer care.

Communication Two approaches defi ne communication: the ‘information engi- neering’ approach and the ‘social construction’ approach.

Th e information engineering approach defi nes communication as the ‘linear transmission of messages through a conduit’, that is, eff ective communication is the accurate transmission of informa- tion, resulting in the receiver understanding what is said. Noise (both audible and psychological) is the main barrier to eff ective communication in this model.

Social construction theory examines the way in which people work, their inter-relationships and behaviours in a team context and how this positively or adversely impacts the quality of team communication. Th is theory implies that communication is a social process. So much so, that eff orts to improve the transfer of information are limited, unless the ways teams work, their dynam- ics and relationships are considered in parallel. Team communica- tion is, therefore, not just about transmitting information but also about the social process of receiving that information.

Teamwork Th e key features of a team are: • It consists of two or more individuals • Each individual has a specifi c role or task to perform and inter- acts and/or coordinates with other members to achieve a common goal or outcome • A team makes decisions • It embodies specialised knowledge and skills, oft en functioning with a high workload • It exhibits interdependencies with regard to workfl ow, collective action and goals • It is a part of a larger organisational system Th ere is a tendency in healthcare for teams to be organised hierarchically or geographically. Research highlights hierarchy,

inhibits psychological safety adversely impacting the quality of teamwork and communication whilst teams organised across geography have to compensate barriers such as time diff eren- tials and distance. However organised, the priority is for teams to be co-ordinated and co-operative. Members of a team must engage in both task work and teamwork processes to achieve their common goal. Task work is the component of the indi- vidual member’s performance that is independent of interaction with other members. Teamwork is the interdependent compo- nent of performance that is necessary to eff ectively co-ordinate the performance of multiple team members. Team performance is a multilevel process that develops as members engage in task work and teamwork.

Th e characteristics of an eff ective team include elements of: • Organisational structure – clear purpose, appropriate culture, specifi ed task, distinct roles, suitable leadership, relevant members and adequate resources • Individual contribution – self-knowledge, trust, commitment and fl exibility • Team processes – co-ordination, communication, cohesion, decision making, confl ict management, social relationships and performance feedback

Th e case study in Figure 5.1 shows how teamwork and commu- nication can fail and result in patient death.

Tools to improve teamwork and communication • Briefi ngs: When comprehensively performed, these are crucial and determine how cohesive a team is when working on a task. Th ey are initiated at the start of a task and set the tone for team interaction, ensuring that care providers have a shared mental model of what is going to happen during a process, identify any risk points and plan for contingencies. When done eff ectively, briefi ngs can establish predictability, reduce interruptions, prevent delays and build better working relationships • Debriefi ngs: Th ese are short exchanges that occur at the end of a task to identify what happened, what was learned and how improvements can be made for the next occasion • SBAR: Th is acronym stands for situation, background, assess- ment and recommendation, and it provides a structured approach to conveying information to a colleague. Regular use of SBAR has been shown to reduce the number of patient safety incidents. An example of SBAR in action is shown in Figure 5.2

Conclusion It is worth remembering that analysis of errors and the omissions contributing to patient harm, illustrate that the quality of clinical skills or discrete clinical interventions are rarely the root causes. Evidence suggests in fact, that the major causal factors of avoidable harm are due to the cumulative impacts of poor communication and sub-optimal teamwork.

6 Reporting and learning from errors Figure 6.1 From reporting to learning

More like this? National data • 12 reported deaths • 15 severe harm

National action

Why did this happen? Discuss at M&M meeting

What can we do locally? e.g. Check portable

ultrasounds on wards

What can we do nationally? e.g. work with industry,

re: length of dilators

Local action

Doctor reports incident,

re: chest drain insertion

Local hospital/ facility system

National database

Introduction Doctors and nurses are asked to report when things go wrong, so that others can learn from their experiences. Analysis of the event which harmed or could have harmed a patient can help to uncover important system weaknesses. Action can be taken at local and national levels to strengthen systems and prevent fur- ther errors. National reporting data also help to monitor risks over time and establish patterns of events (e.g. ‘hotspots’ follow- ing changes to delivery of home oxygen services) as well as com- mon learning from rare events which may not be seen at every hospital. One Australian study showed that only 10% of all inci- dents that are reported nationally will be seen more than once every 2 months in an average 250-bed hospital. Th e remaining 90% of incidents would occur less oft en and cover 500 diff erent kinds of events.

Reporting without learning serves no purpose. Reporting sys- tems need to include a sense-making function, which looks for clues in the partial accounts from busy healthcare staff and brings incidents together with other sources of information and clini- cal experience to think about ways of making systems safer. At a local level, forums like mortality and morbidity (M&M) meetings in hospitals oft en focus on the single adverse event, or ‘near-miss’, to draw out common problems and unsafe practices. Th ese are usually held monthly, and teams come together to discuss serious untoward events. At a national level, work can be done to stand- ardise processes or ‘design out’ errors – for instance, changing the name of ‘sound-alike’ drugs (Figure 6.1).

How to report incidents Each hospital and healthcare facility has a local reporting system, collecting paper or e-incident forms through a central risk man- agement function. Since 2004, all NHS organisations in England and Wales have been connected to a national reporting system which automatically uploads locally reported incidents to a central database. Around 100,000 incidents a month are reported in this way, making it the most comprehensive national reporting system in the world.

Most healthcare systems now have incident-monitoring sys- tems: • Spain (ISMP-Spain) • Denmark (Danish Society for Patient Safety) • Sweden (National Board of Health and Welfare) • Netherlands (managed by the Dutch Health Care Inspectorate) • France (REEM, Preventing what is preventable in medicine) • United States: state-wide systems (e.g. Pennsylvania State Reporting System), provider-based systems (e.g. Veterans Associa- tion) and theme-based systems (e.g. medication; Institute for Safe Medication Practice) Some are voluntary systems, depending on professional codes of sharing learning. Other countries have reporting systems which are mandatory for all types of incidents (Denmark, Ireland, Czech Republic and some US states) or for deaths and cases of serious harm (Japan, Netherlands and Sweden). Although England and Wales have a voluntary system, organisations are required to report ‘never events’ – certain serious avoidable errors, such as wrong-site

P art 1 T

he essence of patient safety

15 C

hapter 6 R eporting and learning from

errors surgery. Th ere has been little robust research comparing manda- tory and voluntary systems and their impact on the rates, type and quality of reporting by clinical staff .

In addition to local and national systems, some specialties have bespoke reporting systems. For instance, a tailored anaesthetics e-form has been developed in England and Wales with particu- lar prompts on common errors such as failed intubation and ana- phylaxis. Reported incidents are analysed by anaesthetists to make sense of risks and identify areas for action.

As well as short incidents captured in real time by staff , more serious incidents may be subject to more detailed scrutiny. Avoid- able deaths or serious near misses may be subject to local investi- gations, using techniques such as root cause analysis. Th is intense focus on a few critical incidents, taking evidence from many sources to identify system weaknesses, is more akin to reporting systems from high-risk industries, from oil rigs to chemical plants. Indeed, the fi rst system to investigate error was developed in the 1940s to improve the safety and performance of military pilots. In healthcare, other kinds of reporting include local signifi cant event auditing in general practice, where staff are encouraged to identify incidents from which others can learn.

Barriers to reporting Research has identifi ed a number of barriers to reporting – particularly from doctors, who report in smaller numbers than nurses. Th ese include lack of familiarity with the reporting process, uncertainty as to what should be reported (e.g. certain categories of error such as omitted medicines) and scepticism that positive action will be taken by the organisation. Th ere may also be cultural factors that inhibit reporting – fear of punitive action or discrimi- nation. Qualitative research also confi rms deeply held beliefs that may deter reporting by medical staff – that only bad doctors make mistakes.

What difference has it made? Modern healthcare is complex and relies on multiple interactions between staff , and complicated processes and treatments. Report- ing incidents where patients have been harmed or almost harmed (‘near misses’) is invaluable in pinpointing system vulnerabilities. Incidents reported locally might include equipment shortages on a crash call trolley or confusion from using diff erent kinds of hepa- rin. At a local level, actions might include an audit of resuscitation equipment and rationalising purchasing of heparins within the trust.

At a national level, steps can also be taken to make practice safer. Th is might mean standardising procedures (e.g. introducing

a single crash call number of 2222 in all hospitals) or working with manufacturers to change packaging (e.g. more distinct forms of diamorphine to prevent wrong dose errors). In England and Wales, a national patient safety function has issued guidance on a range of issues, from chest drain insertion to over-sedation by midazolam.

Problems with reporting Th ere are limits to the value of reporting systems. Critics have pointed to low levels of reporting, with inherent bias due to under- reporting from key sectors, such as primary care which consti- tutes less than 5% of all reports (although nine in 10 healthcare contacts are in primary care). Studies comparing data from inci- dent reporting with case note review showed that reporting was relatively weak at identifying some of the more serious incidents (although probably providing richer contextual information). Th is is confi rmed by a US overview which estimated that only 5–20% of adverse events in diff erent settings were reported.

Th ere are also systematic biases due to who reports. Most reports come from nurses, and this is refl ected in the type of inci- dents reported, most commonly patient falls. Many reports are incomplete, leading to data quality issues on coding of harm and suffi cient description to act. Systems which are anonymous, rather than confi dential, limit the ability of organisations to go back to reporters for more information. Despite concerns about underre- porting, the large volume of incidents makes it diffi cult to identify the critical incidents which could be prevented.

Where do we go from here? Research suggests that reporting systems are only eff ective with strong feedback loops and evidence of action following reported harm. In order to act on risks identifi ed from reports, mechanisms are needed where clinical and other staff can interpret the story of the incident, combining this with other information on the clinical risk (from sources such as patient complaints or litigation data), and identify actions that will reduce risk. Th is can happen at a national level – for instance, action by professional bodies and manufacturers on spinal connectors following reported errors of intravenous medicines given by the wrong route. Action can also be taken locally, through refl ection and multi-professional learning in forums such as mortality and morbidity meetings. Data on their own will not lead to safer practice. But data which are interpreted and understood, with an eye for strengthening systems rather than focusing on individual error, can transform practice. Each national safety initiative, from methotrexate prescribing to nasogastric tube replacement, started with a single report of patient harm from a busy clinician.

7 Research in patient safety Figure 7.1 Framework for patient safety research

Patient safety research

Domain 5

Evaluating the association between

organisational characteristics and outcomes

Evaluate organisational characteristics that help or hinder

research efforts or patient safety practices

Domain 4

Identifying and mitigating hazards

Use of retrospective and prospective analyses to identify and mitigate

safety hazards at the microscopic level

Domain 3

Assessing and improving culture

Strategies and interventions to improve safety culture and

communication

Domain 2

Translating evidence into practice

Develop and evaluate interventions that increase the extent to which

patient received evidence- based medicine

Domain 1

Evaluating progress in patient safety

Develop valid and feasible measures to evaluate progress to improve

patient safety

P art 1 T

he essence of patient safety

Figure 7.2 TRIP model

En ga

Ed uctate Execute Evaluate

1. Summarise the evidence

4. Ensure all patients receive the interventions

3. Measure performance

2. Idenfify local barriers to implementation

Introduction Aft er the To Err Is Human report uncovered the magnitude of the problem of patient safety incidents (more than 98,000 deaths per year due to medical errors), another report by RAND in 2003 revealed that hospitalised patients in the United States received only half of the recommended therapies. Th e framework for patient safety research and improvements (PSRI) described in this chapter evolved out of the need to bridge the gap between the interventions being implemented and scientifi c assessment of their success and applicability to similar settings.

Th e framework has fi ve domains (Figure 7.1): 1 Evaluating progress in patient safety 2 Translating evidence into practice 3 Assessing and improving culture 4 Identifying and mitigating hazards 5 Evaluating the association between organisational characteris- tics and outcomes.

Evaluating progress in patient safety Measures of patient safety involve two balancing acts: 1 Balancing the desire of a global, although more biased, measure of safety versus a more focused, but less biased (robust), measure. A global measure applicable to all patients (e.g. hospital mortality) has extreme bias due to inadequate risk adjustment and accounting for patient preferences. A specifi c measure (e.g. central line bloodstream infections) is very robust but targets only a subset of patients. Many specifi c measures are needed to cover the whole patient population. 2 Finding a balance between a measure that is scientifi cally sound (valid and reliable) and feasible given the limited resources. Use of relatively easy and inexpensive data sources, such as administra- tive data for measures such as deep venous thrombosis, is feasible but correlates poorly with medical chart review data. In order to address these, it is necessary to:

(a) Reduce the quantity but not quality of data. (b) Consider the validity of a measure at two levels: a. Patient safety domain: If it is an outcome, does it represent an impor- tant aspect of quality, and do either variation in practice among organisations or interventions that improve the outcome demon- strate that it is largely preventable? If it is a process measure, does evidence suggest that the intervention will improve outcomes? b. What study design is used to measure the patient safety domain? Are there well-defi ned research protocol, data collection tools, well-designed databases, clear quality control plans, and detailed analytic plans? Cluster-randomised designs, a stepped wedge-trial design, or a quasi-experimental (time series) design can be used. As most studies tend to be a pre-post design, it is important to adjust for historical bias or changes in performance over time.

17 C

hapter 7 R esearch in patient safety

Translating evidence into practice (TRIP) Th e TRIP model (Figure 7.2) looks to improve the reliability of care by focusing on systems (how we organise our work) and engaging a multi-disciplinary team to assume ownership of the improvement project. It is based on evidence and performance measurement, and creates a collaborative culture that is essential for sustaining results.

Assessing and improving culture Th e four important aspects of assessing and improving culture are: 1 What is safety culture? ‘Th e way we do things around here’ is a practical, easily understandable defi nition. Low-cost, quick, annual assessments of safety culture have led to a reliance on climate ques- tionnaires, which measure a snapshot of the larger culture through diff erent dimensions such as safety climate or teamwork climate. ‘Safety culture’ generally refers to an organisational culture, whereas ‘safety climate’ is more transient and refers to teams. 2 How do you measure safety culture? A Safety Attitudes Question- naire is the most widely used instrument to evaluate staff members’ attitudes towards patient safety. Unit/ward-level, department-level and institution-level assessment can be done around six scales: safety climate, perceptions of management, teamwork climate, job satisfaction, stress recognition and working conditions. Another widely used and tested instrument is the AHRQ Hospital Survey on Patient Safety Culture (HSOPS). 3 How do you use safety culture results? Th e two goals in assessing safety culture progress are to achieve or maintain a unit-level score of at least 60% agreement, and improve last year’s climate score by 10 points or more (on a 100-point scale). Unit/ward-level results help hospitals recognise units that need resources or leadership support, and help health systems identify hospitals that are strug- gling versus those that are thriving.

Identifying and mitigating hazards Th e following two methods are used to identify and analyse the healthcare system at the level of the unit/ward, department or hospital, to determine the source of potential or known risks to patient safety: 1 Retrospective identifi cation of hazards: Th is involves in-depth analy- ses of sentinel events to identify the causes and contributing factors associated with an adverse event, then planning and implementing strategies to prevent the event from recurring. Th is may be formal (e.g. root cause analysis) or informal (e.g. case review by a quality improve- ment committee). Tools such as the ‘Learning from Defects’ tool help with in-depth analysis. Other tools include incident-reporting sys- tems (e.g. the National Reporting and Learning System that collects reports of patient safety incidents); medication error-reporting sys- tems such as MEdMARX, which collects data on medication errors in the United States; and intensive care unit safety reporting systems. 2 Prospective identifi cation of hazards: Th is involves identifying hazards in the system before patient harm occurs. Unfortunately this is limited by institutional resources and capacity to accomplish the task. Failure mode and eff ects analysis (FMEA) is a tool used by the aeronautical industry but its validity, reliability and eff ective- ness have not been well documented (Chapter 27). Simulation is another tool which holds promise to improve patient safety; for example, simulation of resuscitation during cardiac arrest or mass- causality events helps identify hazards in the process of care.

Evaluating the association between organisational characteristics and outcomes Organisational context must be taken into account in research, as social and structural characteristics strongly infl uence employee behaviour. For example: • How to translate evidence into practice?

• What resources to dedicate towards improvement eff orts? • How to mistake-proof day-to-day operations? Organisational variables that can aff ect patient outcomes are: • Organisational design • Organisational culture • Policies, procedures and requirements • Rewards and incentives • Communication networks (formal and informal within and out- side the organisation) • Patient centredness • Skills, knowledge and dedication of leaders Knowledge about the association between the above characteris- tics and patient safety is important, but diffi cult to assess because valid measures of patient safety are diffi cult to obtain and organi- sational variables lack a standard defi nition. Th is could introduce misclassifi cation and measurement bias (e.g. variation in nurse turnover defi nitions at the unit and hospital level).

Challenges for patient safety research Th e following challenges need to be overcome in order to apply an eff ective research and improvement framework: 1 Build capacity: Th e only way to build momentum in the fi eld of patient safety research is to provide trainees with formal course- work in research methods, mentorship and structured research experience. Advanced trainees are encouraged to get a master’s or doctoral degree in their fi eld of work. Due to the paucity of time, a start could be attending weeklong workshops to grasp the con- cepts. Evaluation skills can be obtained by prolonged coursework. In order to provide common guidelines for training in patient safety research, the World Health Organization has produced a guide to enable teaching across the globe in a standardised format. 2 Create infrastructure: A platform needs to be established for there to be an interchange between clinicians and methodologists. One way to do this is to hold regular multi-disciplinary meetings to bring the disciplines of clinical work and quantitative research together. 3 Evaluate the cost–benefi t ratio of improvement eff orts: Research- ers need to be able to articulate the cost–benefi t ratio of interven- tions, such as hiring new staff or purchasing new equipment, to enable the senior leadership and regulators to make informed deci- sions before mandating a safe practice.

Future direction Key recommendations for future work in patient safety research are: 1 Develop valid measures to evaluate patient safety progress. 2 Develop methods to reliably translate evidence into practice. 3 Study the link between culture, behaviour and outcomes. 4 Evaluate teamwork and leadership behaviours. 5 Use simulations to:

(a) evaluate teamwork and technical work (b) train staff (c) identify and mitigate hazards.

6 Coordinate national eff orts to implement industry-wide changes. 7 Explore ways to eff ectively and effi ciently use resources at all lev- els of a service (unit, ward, department and hospital). 8 Advance the science of measurement and reduction of diagnos- tic errors in medicine. 9 Develop patient safety measures for specifi c product lines (e.g. cardiac surgery) that should be mostly generalisable across settings.

Part 2Understanding and interpreting risk

Chapters 8 Risk-based patient safety metrics 20 9 Root cause analysis 24 10 Measuring safety culture 26

P art 2 U

nderstanding and interpreting risk

8 Risk-based patient safety metrics Figure 8.1 Hospital-standardised mortality ratio (HSMR) for Walsall NHS Trust from 1996 to 2004

Note: Treatment changes were implemented in 2001. CUSUM, cumulative sum chart. Source: Jarman B, Bottle A, Aylin P & Browne M 2005. Reproduced with permission of BMJ Publishing Group Ltd.

C U

S U

M s

ta ti

st ic

0 Feb 1996 (1)

April 1997

(8350)

May 1998

(17,795

June 1999

(27,139)

July 2000

(36,575)

Aug 2001

(46,258)

Sept 2002

(56,396)

Nov 2003

(66,469)

Date of admission to hospital

Hospital standardised mortality ratios,

Walsall Hospitals NHS Trust, 1996–2004.

Treatment changes were implemented

in 2001 (see bmj.com for explanation

of cusum method)

21 C

hapter 8 R isk-based patient safety m

etrics Introduction Patient safety metrics are measures of health service performance in the context of patient safety. Th ey are the evidence base that guides quality improvement strategies to minimise harm and enhance patient experience. Patient safety programmes choose a metric this is based on available data, provides reliable and valid measures and supports desired objectives.

Th e two safety metrics most commonly used in healthcare aim to eliminate medical errors and prevent patient injuries. In indus- try, there is a third metric that focuses on proactively identifying and removing hazards to improve workplace safety. Th is last met- ric is not yet widely adopted in patient safety programmes, but may prove a useful systematic approach to design safer systems that support and enhance service delivery performance. Th e suc- cess of safety improvement initiatives is reliant upon these metric characteristics, as fl awed or inappropriate frameworks can mis- guide strategies and have high opportunity costs.

Error-based patient safety metrics Error-based metrics aim to detect and reduce medical errors, regardless of their impact on patient safety. Th e Institute of Medi- cine (IOM) defi nes error as the failure to complete a planned action as intended or the use of the wrong strategy to achieve a desired objective. Th is metric system attempts to capture both of types of error in the error rate measurement:

Error rate = Identifi ed errors / total opportunities for error

Th e magnitude of this rate is compared to national or regional standards to determine the focus of improvement strategies.

Recent reports have highlighted the unacceptably high fre- quency of preventable adverse events and near misses arising from medical errors. Th eir signifi cant impact on mortality, morbidity and healthcare costs makes this approach an important measure of patient safety and a useful guide for harm-preventing strategies.

• Imprecise error rate: The accuracy and reliability of error rates are limited to the method of detection. Current methods are lim- ited to voluntary reporting, chart review and direct observation. These techniques provide complementary results, but they are not independently valid, reliable means of error identifi cation • Hindsight bias: Retrospective error identifi cation is suscep- tible to misinterpretation of events leading to poor outcomes, because the outcome is known. Inaccurate perceptions of cau- sality can misguide improvement efforts • Reinforcement of blame: A focus on individual rather than systemic failures does not address broader safety concerns that could lead to sustained quality improvements. Errors arise from latent and active failures; however, this metric only cap- tures those from active failures, ignoring wider infl uences • Not always related to harm: Many errors do not impact patient safety; thus, a focus on preventing errors may not lead to directly observable improved health outcomes • Negative connotation: Health providers’ fear of malpractice suits may cause a high rate of underreporting

Limitations

Injury-based patient safety metrics Injury-based metrics aim to eliminate preventable adverse events, including those that are not associated with any identifi able error. Th e injury-centred framework is based on the rate of injury that can be monitored over time:

Injury rate = Identifi ed injuries / total opportunities for injury

Unlike error-based safety metrics, this approach allocates resources and sets safety priorities that will lead directly to observable improvements. However, this method is necessarily retrospective and reactive in nature, and is thus subject to simi- lar limitations as error-based metrics and additional ethical constraints.

• Reactive rather than proactive: Fails to proactively prevent harm by requiring that patients be injured before improvement measures are taken • Reinforcement of blame • Poor discrimination of preventability: Without a reliable tool to discern preventability, resources may be wasted attempting to prevent unavoidable injures • Imprecise injury rate: Identifi cation of injuries depends on routinely collected administrative data, which are often incomplete. It is also diffi cult to identify all opportunities for injury. Thus, this rate is susceptible to frequent random variation • Hindsight bias: Retrospective approach makes it susceptible to oversimplifi ed attribution of cause of patient injury • Negative connotation

Limitations

Hazard- or risk-based patient safety metrics Hazard-based metrics aim to proactively identify, measure and remove hazards to create a safer work environment. Th is approach extends patient safety initiatives beyond individual-centred improvements to address the broader interacting healthcare sys- tem elements, including organisational structure, environment and technology. Th e premise of this method is that through designing a safer system, medical error reduction and injury prevention will follow. Th is metric system also has a more positive connotation than error and injury-based metrics and is likely to encourage greater physician involvement.

• Diffi cult to identify all hazards: All potential risks to patient safety will never be eliminated • Limited use in the patient safety context to date: There is little evidence of its success in the healthcare setting, although it is extensively used in industry • Impact on patient safety is not directly measurable

Limitations

Other applications of metrics Death is the most tractable outcome of care – it is easily measured, is of undisputed importance to everyone and is com- mon in hospital settings. Th is led to the creation of hospital- standardised mortality ratios (HSMRs). Th ese are another useful tool to evaluate hospital performance by comparing risk- adjusted mortality rates with the national average. Hospitals monitor changing patterns in performance over time through a graphical presentation termed the ‘cumulative sum chart’, or CUSUM (Figure 8.1). Th is chart demonstrates how HSMR can be analysed at various time intervals to detect unacceptably high mortality rates and the success of subsequent improvement eff orts. A hospital’s HSMRs can then be compared to those of other hospitals, as shown in Figure 8.2.

P art 2 U

nderstanding and interpreting risk

Figure 8.2 Comparing efficiency and hospital-standardised mortality ratio HSMR for fi ve hospitals

Comparing effi ciency and hospital-standardised mortality ratio (HSMR) for fi ve hospitals

Trusts

Your nearest trust

Your next four nearest trusts

Trusts in England

Boundaries

HSMR below expected

HMSR within expected

HMSR above expected

Low efficiency/high mortality

Low efficiency/low mortality

High efficiency/high mortality

High efficiency/low mortality

Source: Reproduced by permission of Dr Foster Intelligence 2013.

23 C

hapter 8 R isk-based patient safety m

etrics A high death rate can be attributed to chance, regression to

the mean, diff erent procedures, inadequate case-mix adjustment, poor data quality or poor quality of care. All other reasons must be investigated before poor performance is considered the cause of the alarm. Th e advantage of using health outcomes to measure performance is that it is objective, readily available and routinely collected, and matters most to patients. However, this measure is also easily aff ected by confounding factors, includes unavoidable mortality, and is diffi cult to adjust for case-mix factors.

Quality indicators Quality indicators are screening tools used to identify poor qual- ity of care and patient safety concerns. Th ese measures are based on routinely collected hospital administrative data that contain information on diagnoses, procedures, patient characteristics and discharge status. Although the data capture a limited picture of service quality, they provide a common denominator on which to compare hospital performance and guide quality improvement eff orts.

Comparative performance assessments based on risk-adjusted quality indicators highlight unacceptable variation in health out- comes between regions, communities and providers. While quality indicators do not provide defi nitive measures of healthcare quality, they are starting points for further investigation and guide pay-for- performance incentives.

Types of indicators 1 Patient Safety Quality Indicators (PSIs): Th ese indicators screen for preventable adverse events and complications arising from medical interventions. PSIs can be based on a single hospital episode (provider-level indicators) or cases from separate hospi- tal episodes (area-level indicators). However, it is diffi cult to dis- tinguish preventable from unpreventable adverse events. 2 Prevention Quality Indicators (PQIs): Th ese assess the quality of health services in local communities (outpatient care) using inpatient hospital data. PQIs focus on ambulatory care sensitive conditions for which good outpatient care can prevent the need for hospitalisation. However, inpatient data do not consider patient preferences for inpatient and outpatient care, or socio-economic infl uences. Th ere is also little evidence to suggest that eff ective out- patient treatments reduce hospital admissions.

• Fails to capture all complications of interest • Prone to incomplete reporting • Over-emphasises surgical procedures • Hard to distinguish medical complications from comorbidities • Collected for billing purposes, not for research • Coding differences across hospitals • Ambiguous timing of condition onset: codes fail to distin- guish if it occurred before or during hospital stay • Limited clinical characterisation: codes tend to group highly heterogeneous clinical conditions into a single code • Limited case-mix factors

Limitations of administrative data

3 Inpatient Quality Indicators (IQIs): Th ese assess inpatient hospi- tal care using in-hospital mortality rates; utilisation measures of procedures in question for overuse, underuse and misuse; and vol- ume measures of procedures for which evidence suggests a link between the number of procedures performed and health out- comes.

Never events ‘Never events’ are preventable, serious and unambiguously defi ned adverse incidents that should not occur if appropriate national safety measures and guidelines are in place. Th ey refer to alarming medical errors such as wrong-site surgery, retained instrument post operation and wrong route of administration of chemotherapy. Never events are warning signs indicating inad- equate and ineff ective patient safety systems that require further investigation. Th ey are an attempt to chase zero avoidable harm in healthcare. Monitoring never events within the contract between commissioners and providers forms part of the wider safety and quality agenda in the English NHS and also in other countries internationally, such as the United States. Increasing pressure is being put on healthcare organisations to eliminate never events. In fact, the US Centre for Medicare and Medicaid Services (CMS) announced in August 2007 that Medicare would no longer pay for additional costs associated with many preventable errors, includ- ing those considered never events.

In the United States, never events are publicly reported, with the goal of increasing accountability and improving the quality of care. Since the National Quality Forum (NQF) disseminated its original never events list in 2002, 11 states have mandated report- ing of these incidents whenever they occur, and an additional 16 states mandate reporting of serious adverse events (including many of the NQF never events). Healthcare facilities are account- able for correcting systematic problems that contributed to the event, with some states (such as Minnesota) mandating perfor- mance of a root cause analysis and reporting its results. Similar initiatives have been seen in the United Kingdom. In January 2012, the Department of Health published its expanded list of ‘25 Never Events’ and, more recently, it has a updated the Never Events Policy Framework to provide greater clarity and recommended responses to these events. Th e document also contains data on the number and types of never events reported, revealing that 326 never events were reported to strategic health authorities in 2011–2012. Th is shows that there is a long way to go before the incidence of never events is brought down to zero.

Conclusion Th e strengths and weaknesses of the metrics and indicators described in this chapter directly infl uence the success and fail- ure of these initiatives. In addition to informing policy decisions, increasing public availability of performance metrics is guiding and empowering patients to make informed decisions regarding their health. Th is growing transparency in healthcare will play a signifi cant role in driving quality improvement through greater accountability to public expectations and pay-for-performance incentives.

9 Root cause analysis Figure 9.1 Steps to an effective root cause analysis (RCA) investigation

Figure 9.2 RCA investigation: fi shbone diagram – tool

Good organisational safety culture (infrastructure, resource, support)

Appreciation of the value of,

rationale behind, and indications for

investigation

Appreciation of the value of,

rationale behind, and indications for investigation

Competence in thorough, credible

investigation analysis

Skilled in error wisdom, research,

interpretation and deduction

Skilled in improvement

science

Select and apply effective remedies/ solutions for each

cause ...

and monitor for success

Source: NHS National Patient Safety Agency

Equipment and resource factors • Displays • Integrity • Positioning • Usability

Communication factors • Verbal • Written • Non-verbal • Management

Task factors • Guidelines/ procedures protocols • Decision aids • Task design

Individual (staff) factors • Physical issues • Psychological • Social/domestic • Personality • Cognitive factors

Patient factors • Clinical condition • Physical factors • Social factors • Psychological/ mental factors • Interpersonal relationships

Working conditions/ environment factors • Administrative • Design of physical environment • Environment • Staffing • Workload and hours • Time

Organisational and strategic factors • Organisational structure • Priorities • Externally imported risks • Safety culture

Education and training factors • Competence • Supervision • Availability/ accessibility • Appropriateness

Team and social factors • Role congruence • Leadership • Support + cultural factors

Problem or issue

(CDP/SDP)

Source: NHS National Patient Safety Agency

P art 2 U

nderstanding and interpreting risk

25 C

hapter 9 R oot cause analysis

What is root cause analysis? Root cause analysis (RCA) is a method of incident investigation.

As such, it is a diagnostic tool rather than a safety solution in itself. RCA allows a systems approach (Chapter 26) to investigation and was selected as the methodology of choice by the National Patient Safety Agency when developing a framework for patient safety investigation in the NHS. Th e NHS approach aligns well with investigation methods used in healthcare and other high-risk industries across the globe.

Why investigate? Th e primary aim of patient safety investigation is to learn from inci- dents and to determine what can be done to signifi cantly reduce the likelihood of recurrence; the aim is not to apportion blame.

RCA process • Investigations can be comprehensive or concise but must always include the basic elements to help ensure they are thorough, cred- ible and actionable, and represent value for money • Set clear terms of reference and follow them. Secure adequate time and skills, or record and report the impact of constraints • Avoid lots of concise investigations. Th ey can prove false economy 1 Gathering and mapping the information

• You have to understand exactly what happened leading up to an incident before you can fully understand why it happened • Investigative interviewing should focus more on listening than on asking questions • Consult the patient and family as part of the investigation; they have a unique perspective and valuable information to share

2 Identifying care and service delivery problems (CDPs and SDPs) – this stage involves identifying all the points at which:

• something happened that should not have happened; or • something that should have happened did not

3 Analysing problems • Using a fi shbone diagram (or Ishikawa diagram or cause- and-eff ect diagram) as shown in Figure 9.2, place one CDP or

SDP in the head of each fi sh (not the whole incident), then ana- lyse why that course of action seemed the right thing to do at that time • A few carefully analysed ‘fi shbones’ focusing on key CDPs and SDPs will deliver more benefi t than many completed quickly • Training in systems thinking and human factors (including error types and biases) will aid impartiality and quality analysis • Th e root causes are the most signifi cant contributory factors

4. Generating recommendations and solutions • Problems will rarely be resolved for the long term by applying discipline, training and updated procedures alone • Training in improvement science will assist with more eff ective selection and implementation of solutions

5. Implementing solutions • Amalgamate action plans from investigations. Th is encour- ages trend analysis and a more cohesive, high-level approach to resolving common issues • Avoid conducting more and more investigations with similar outcomes. Time must be allocated to implementing solutions and monitoring their effi cacy

6. Writing the Investigation Report • Use an RCA investigation report template to facilitate trend analysis, audit and shared learning

Effective RCA investigation Th e components for success in patient safety investigations are the same as those required for successful clinical investigations (Figure 9.1): 1 To avoid the extremes of delayed problem ‘diagnosis’ and resource wastage, triggers or indications for conducting an investi- gation must be correctly identifi ed. 2 To obtain a good-quality, accurate picture of the problem, data gathering must be conducted by those skilled in the process. 3 Th e fi ndings from the collection of data must be robustly inter- preted and credible conclusions drawn by someone with analyti- cal skills and an understanding of the ‘anatomy, physiology and pathology’ of the issue. 4 To ensure that improvement is achieved and measurable, expert selection, application and monitoring of eff ective treatment and remedial action are required. 5 If meaningful learning and improvement are expected from incident investigation, there must be organisation-wide support for this process.

Chapter 27 gives an example of a fi shbone diagram in use.

10 Measuring safety culture

Figure 10.1 Key components of safety culture

Leadership

Reporting systems

Communication and sharing

Rules and processes

Safety culture

Training

Teamwork

Introduction ‘Safety culture’ refers to the way that patient safety is thought about and implemented within an organisation. It is about the way safety is perceived, valued and prioritised. Safety climate is a subset of this, and refers to what staff think about safety. If an organisation has a ‘strong safety culture’, this means that processes are in place to keep people safe and that staff think that the organisation is doing a good job to promote safety.

Th e term ‘safety culture’ fi rst became popular aft er the Cherno- byl nuclear disaster, when people began suggesting that organisa- tions could reduce safety incidents by developing a ‘positive safety culture’. Safety culture has been measured in industries such as aviation, transport, manufacturing and oil and gas production. In healthcare, safety culture has become a key indicator over the past two decades.

Th e main components of safety culture include good leadership and teamwork, staff training, rules and safety processes, monitor- ing systems to identify and learn from incidents and communica- tion and sharing to promote ongoing learning (see Figure 10.1).

Measuring safety culture A number of tools are available to measure safety culture. Most involve surveys completed by managers and frontline staff . Th e most well-known tools are: • Safety Attitudes Questionnaire • Patient Safety Culture in Healthcare Organisations • Hospital Survey on Patient Safety Culture • Safety Climate Survey • Manchester Patient Safety Framework Many other surveys are also available (see Figure 10.2).

Usually the surveys used to measure safety culture are simple and quick to complete. Th ey are designed to be completed regu- larly over time, such as once a year, to see if there have been any changes. Figure 10.3 provides an example of the format of one short safety culture survey.

Most tools are targeted towards hospitals, but a small number have been tested in other settings such as primary care, nursing homes and emergency services.

Box 10.1 describes a tool developed to measure safety culture in the United Kingdom. Th is tool is designed to be used in work- shops with staff . Results are fed back and teams decide on changes to improve safety. Th e tool is then repeated aft er the changes are made – and the improvement cycle continues.

Some tools, such as the Safety Attitudes Questionnaire, have been more widely tested than others, but there is not enough evi- dence to say that one survey is any better or worse than others (see Table 10.1). In general, tools that are short, able to be repeated over time and able to be adapted for use in many diff erent contexts may be most practical.

Much of the research about tools for measuring safety culture comes from the United States. We must be careful when trans- ferring tools that work well in some countries to other countries where services are arranged diff erently. Similarly, tools that work well in hospitals might not be best for primary care, and surveys used in intensive care units or specialist wards might not necessar- ily be best for other services. When deciding which survey to use to measure safety culture, it is important to look at the tool care- fully, test out whether it works on a small group and think about whether it is really relevant to the context you are working in.

P art 2 U

nderstanding and interpreting risk

27 C

hapter 10 M easuring safety culture

Danish patient safety

culture questionnaire

Checklist for assessing institutional resilience

Trainee supplemental

survey

Safety climate assessment

tool

Safety attitudes

questionnaire

Nursing unit cultural assessment instrument

Veteran affairs Palo Alto/ Stanford patient

safety centre

Hospital culture

questionnaire

Safety climate scale

Hospital survey on patient

safety culture

Teamwork and patient

safety attitudes questionnaire

Culture of safety survey

Error orientation

questionnaire

Patient safety climate in

aesthesia

Stanford safety culture

instrument

Vienna safety culture questionnaire

Safety climate survey

Hospital survey on

patient safety

Patient safety culture questionnaire

Patient safety culture in healthcare organisations

survey

TUKU– safety culture in healthcare

survey

Manchester patient safety assessment framework

Veterans health

administration patient safety

culture

World alliance for

patient safety hand hygiene

The culture of this clinical area makes it easy to learn from the mistakes of others

Medical errors are handled appropriately in this clinical area

The senior leaders in my hospital listen to me and care about my concerns

The physician and nurse leaders in my area listen to me and care about my concerns

Leadership is driving us to be a safety-centred institution

My suggestions about safety would be acted upon if I expressed them to management

Management/leadership does not knowingly compromise safety concerns for productivity

I am encouraged by my colleagues to report any patient safety concerns I may have

I know the proper channels to direct questions regarding patient safety

I receive appropriate feedback about my performance

I would feel safe being treated here as a patient

Briefing personnel before the start of a shift (i.e. to plan for possible contingencies) is an important part of

patient safety

Briefings are common here

I am satisfied with availability of clinical leadership (please respond to all three): Physician

Nursing

Pharmacy

This institution is doing more for patient safety now, than it did one year ago

I believe that most adverse events occur as a result of multiple system failures, and are not attributable to one

individual’s actions

The personnel in this clinical area take responsibility for patient safety

Personnel frequently disregard rules or guidelines that are established for this clinical area

Patient safety is constantly reinforced as the priority in this clinical area

10.

11.

12.

13.

14.

15.

16.

17.

18.

19.

A Disagree strongly

B Disagree slightly

C Neutral

D Agree slighty

E Agree strongly

X Not applicable

A B C D E X

Figure 10.2 Examples of the large number of tools available to measure safety culture

Figure 10.3 Example of the questions and format in one safety culture survey

Key weaknesses Tool and developer Where it has been used Key strengths Amount of

evidence available

Hospital survey on patient safety culture (AHRQ)

Safety attitudes questionnaire (developed from aviation tool)

Manchester patient safety framework (NPSA)

Safety climate survey (University of Texas and IHI)

• Hospitals in North America

Patient safety climate in healthcare organisations (Stanford, funded by AHRQ)

• Hospitals in the US • Studies with large sample sizes have validated the tool

• Has been used mainly by one group of researchers • Tested almost exclusively in US hospitals

• Hospitals in the in UK • Pharmacy in UK • Hospitals in Canada

• Hospitals • ICU • Primary term care in many countries

• Hospitals in the UK, Belgium, China, Netherlands, Turkey, Saudi Arabia, Spain, Lebanon, etc.

• Can compare with other countries and industries

• Well validated and established • Can compare with other countries and industries

• Focuses on a broader way of thinking about safety culture • Can be done in staff workshops

• Short and easy to complete • Has been compared with other surveys

• Tested mainly in North America • Developed some time ago

• Little has been published about usage

• Not used much in UK • Some think it takes too much time to complete

• Focuses only on hospitals • Has some validity issues

= Sufficient evidence

= Limited evidence

The Manchester Patient Safety Framework was developed based on literature reviews and expert input. It aims to help organisations assess their progress in developing their safety culture and see how ‘mature’ they are across 10 dimensions: • Continuous improvement • Priority given to safety • System errors and individual responsibility • Recording incidents • Evaluating incidents • Learning and effecting change • Communication • Personnel management • Staff education • Teamwork

The tool asks staff to rank the level of safety maturity in each of these categories using the following subsets: pathological (‘Why waste our time on safety?’), reactive (‘We do something when we have an incident’), bureaucratic (‘We have systems in place to manage safety’), proactive (‘We are always on alert for risks’) and generative (‘Risk management is an integral part of everything we do’).

The tool is designed to be completed in a staff workshop led by a facilitator from the healthcare organisation.

A good point is that this tool can be used in hospitals as well as primary care, mental health and ambulance services and can be applied at an organisational or team level. For example, in England, 10 community pharmacies used the tool and 67 pharmacists and support staff took part in work- shops to help assess the safety culture of their pharmacy. This helped raise awareness about patient safety and high- lighted differences in perceptions between staff and areas for improvement.

This tool can be used to help teams refl ect on safety cul- ture, reveal any differences in attitudes between different types of staff and help understand what a more mature safety culture might look like. It has been used mainly in the United Kingdom, although some validation has taken place in North America.

Box 10.1 Example of using the Manchester Patient Safety Framework

Table 10.1 Features of the top fi ve most well-known tools for measuring safety culture

Why measure safety culture? Safety culture can be used to measure how well an organisation is doing in terms of patient safety. Other measures of patient safety such as error rates, death rates or record reviews can be diffi cult to measure consistently, be labour intensive regarding the collection of data or take a long time to be aff ected by changes in processes and systems. Measuring safety culture gets around some of these problems and can be an easy way to show changes over time.

Because it is based on people’s views, measuring safety culture alone may not give a good indication of the levels of patient safety in an organisation. However, when used alongside other measures, it can provide a quick and inexpensive way of monitoring change.

Does safety culture improve outcomes? Th e way an organisation or healthcare team thinks about and implements patient safety processes may have a signifi cant impact on the people using services and the staff providing them.

Many studies suggest there is a link between safety culture and patient outcomes, but the relationship is not clear-cut. Some research has found a relationship between safety culture and hos- pital morbidity, adverse events and readmission rates. But other studies have found that safety culture has no impact on patient outcomes. Th ere is more evidence that improving safety culture impacts staff safety behaviours and injury rates amongst staff .

Even when there is a clear link, it is not certain that a good safety culture creates better outcomes. For example, researchers in the United States examined the relationship between patient safety culture and rates of rehospitalisation within 30 days of discharge. Survey data from 36,375 staff from 67 hospitals were compared with risk-standardised hospital readmission rates. Poorer views of safety culture were associated with higher readmission rates for heart attacks and heart failure (Hansen LA, 2011). Other research- ers in the United States examined whether safety culture was linked

P art 2 U

nderstanding and interpreting risk

29 C

hapter 10 M easuring safety culture

with patient outcomes in intensive care units (ICUs). Data from 65,978 patients admitted to 30 ICUs were analysed and 2103 staff were surveyed. For every 10% decrease in safety culture, length of stay increased by 15% (Huang DT, 2010).

Aft er initiatives are put in place to improve patient safety, there is sometimes a simultaneous improvement in both safety culture and patient outcomes. Th erefore, rather than a one-way causal relationship where safety culture infl uences behaviours and clini- cal outcomes, there may be a circular relationship whereby changes in behaviours and outcomes also help to improve safety culture. If things are being done better, then staff might feel better about safety overall (see Figure 10.4).

Improving safety culture An organisation with a positive safety culture encourages and retains learning, promotes open and honest reporting of safety incidents, does not penalise staff for systems errors, is prepared to identify its own shortcomings, rewards innovation and accepts constructive suggestions for continuous improvement. Improv- ing safety culture is therefore about enhancing the entire way that patient safety is thought about and acted upon.

Organisations wanting to improve their safety culture may address key drivers such as: • Ensuring leaders prioritise safety in policy and practice • Visible participation by leaders such as ward walkrounds • Having a standardised system-wide approach to safety • Holding managers and staff accountable for safety • Including safety in annual staff performance reviews • Taking steps to minimise sources of error or harm • Regular staff training about safety issues • Asking staff , patients and families for improvement ideas • Constant assessment of the safety signifi cance of events • Fair treatment of those who report safety incidents • Recognising and rewarding good performance • Sharing information about successes and improvement • Listing safety incidents and successes on notice boards A positive safety culture is established when safety is valued as highly as productivity or other outcomes.

Th e most important thing to remember is that the value of measuring and thinking about safety culture lies in raising the profi le of patient safety and promoting conversations about safety.

Th e exact tools used may be less important than how feedback is collated and used. In other words, safety culture is about getting everyone on the same page in terms of improving safety – and this can have benefi ts for staff , for resource use and for patients and their families.

Improvement initiatives

Improved safety culture

Better patient outcomes

Simplistic way people used to think culture affected outcomes

Complex reciprocal relationship supported by research

Patient outcomes

Safety culture

Safety climate

Im pro

vement initiatives

Sta ff outcomes

Af fe

ct e

ac h

ot he

Figure 10.4 Complex relationship between culture and outcomes

Part 3Risks to patient care

Chapters 11 Medication errors 32 12 Surgical errors 36 13 Diagnostic errors 40 14 Maternal and child health errors 44 15 Slips, trips and falls 47 16 Patient safety in paediatrics 50 17 Technology in healthcare and e-iatrogenesis 54 18 Nosocomial infections 58 19 Mental health errors 60 20 Patient safety in primary care 64

P art 3 R

isks to patient care

11 Medication errors Figure 11.1 What are the different types of medication errors?

(e) Monitoring errors

(a) Prescribing errors

(b) Transcription errors (c) Dispensing errors

(d) Administration errors

A prescribing error occurs when, as a result of a prescribing decision or prescription-writing process, there is an unintentional, significant: reduction in the probability of treatment being timely and effective, or increase in the risk of harm when compared to generally accepted practice. For example, prescribing an incorrect drug or dosage

A transcription error is any deviation from the initial prescription or medication order, such as transcribing incorrect information regarding a patient’s history

Errors that occur at the administration stage, such as failing to administer the right drug at the right time to the right patient

A monitoring error occurs when a prescribed medicine is not monitored in the way which would be considered acceptable in routine clinical practice. For example, failure to monitor for drug side effects

A dispensing error is any unintended deviation from an interpretable written prescription or medication order. Both content and labelling errors are included, such as dispensing an incorrect drug strength, form or quantity

33 C

hapter 11 M edication errors

Introduction A medication error can be defi ned as any error that either resulted in, or had the potential to result in, an adverse drug event. Medica- tion errors can occur at any stage of the medication use process, including the prescribing, transcribing, dispensing, administrat- ing and monitoring of a drug. Not all medication errors have the potential to cause harm to the patient.

Adverse drug events (ADEs), or injuries due to drugs, are surprisingly common worldwide. At least 1.5 million prevent- able ADEs occur in the United States each year. Th e picture is no better in the United Kingdom, with approximately one in seven NHS hospital inpatients experiencing an adverse event, half of which were deemed possibly or defi nitely avoidable. For example, a patient with no history of allergic reactions, who experiences an allergic reaction to an antibiotic, has suff ered an ADE. Th is ADE would not be attributable to error. However, an error would have occurred if a patient with a history of documented allergic reactions experiences an allergic reaction to a prescribed antibi- otic because the medical record was unavailable or not consulted. Given the burden of ADEs, it is incumbent on all healthcare pro- fessionals to understand the types and causes of medication errors, to understand how oft en they cause harm and to help build safer medication systems.

Causes and potential prevention strategies for medication errors Errors that cause harm can seldom be attributed to any one indi- vidual or factor. Many diff erent aspects can contribute to an error, such as poor teamwork and communication, insuffi cient or miss- ing information, illegibility, inadequate training, excessive work- load, staff shortages, interruptions and distractions. Th e US Insti- tute of Medicine (IOM) outlined a number of proven medication safety practices in To Err Is Human: Building a Safer Healthcare System to reduce errors in the medication process. Th ese include: • Avoid reliance on memory and vigilance: Work activities should be designed to minimise reliance on human short-term or long- term memory. It is unreasonable to expect individuals to remain vigilant for long periods of time. Th e wise use of protocols and electronic checklists at the point of decision making, for example, can assist physicians with repetitive tasks • Use of constraints and forcing functions: Th is involves structuring critical tasks so as to guide users to the most appropriate action. Designing defaults for processes and devices like an infusion pump to default to shutoff (safe mode) rather than free fl ow can help pre- vent errors • Simplify key processes: An eff ective means of reducing the like- lihood of error is to simplify and standardise key processes. Examples of simplifi cation include reducing the number of hand- off s (e.g. multiple order and data entry) and the number of times a drug is administered per day

Wayne Jowett, aged 18, was diagnosed with acute lympho- blastic leukaemia in June 1999. In June 2000 his disease was in remission and he entered the maintenance phase of his treat- ment. This consisted of the following drugs: 6-mercaptopurine (orally), methotrexate (orally), prednisolone (orally), vincristine (intravenously) and cytosine (intrathecally).

Wayne was scheduled to receive his chemotherapy drugs on the morning of 4 January 2001. On that day, the sister on the ward discovered that the chemotherapy drugs for Wayne were not prepared and asked Dr Musuka, Wayne’s consultant, to prescribe them, which he did. The pharmacist prepared both cytosine and vincristine in the Sterile Production Unit. He had suggested that both drugs be administered one day apart to prevent an error from occurring. He then received a call from the ward to say that Wayne had arrived and so both drugs were sent together to the ward.

When Wayne arrived, the staff nurse, Ms Vallance, informed Dr Morton, a senior House offi cer, who was covering the ward. Dr Musuka was not informed. According to guidelines, intrathe- cal administration of chemotherapeutic agents should be super- vised by a registrar. Therefore Dr Mulhem, the only registrar on the ward on that afternoon, was called to help Dr Morton. Staff nurse Vallance went to the Day Case unit refrigerator, where she found both chemotherapy items packed in one bag. She took the plastic bag to the treatment room and recalls saying: ‘Here’s Wayne’s chemo’. Ms Vallance then left the room and the two doctors went on with the procedure. The lumbar puncture site was marked and local anaesthetic was infi ltrated in the area. Dr Mulhem had a brief look at the prescription chart, but failed to recognise that vincristine was due to be given the following morning, nor that it should be given intravenously. Dr Morton performed the lumbar puncture. Dr Mulhem then read out aloud the name of the patient and the name and dose of the drug to be given. He did not mention the route of administration. Taking the syringe, Dr Morton asked whether the drug was ‘cytosine’ and Dr Mulhem confi rmed that it was. Dr Morton then injected the contents of the syringe into Wayne’s spine. Dr Mulhem then read out aloud the name and dose of the second drug (vincris- tine) and gave the syringe to Dr Morton. A few moments later, Dr Morton injected vincristine in Wayne’s spine. Vincristine, when given intrathecally, causes central nervous system tox- icity, producing progressive ascending myeloencephalopathy. A few minutes later, the two doctors realised that a serious mis- take had happened and called for senior help.

Wayne Jowett suffered leg paralysis and respiratory fail- ure. He was transferred to the ICU, where he was intubated and ventilated. His ventilator was switched off 4 weeks later. Figure 11.2 outlines the failures in the case of Wayne Jowett using the Swiss cheese model.

Box 11.1 The case of Wayne Jowett

P art 3 R

isks to patient care

Health information technology Electronic tools can help reduce the rate of errors and subsequent patient harm, such as: 1 Computerised physician order entry (CPOE) is the direct entry of medication orders into a computer system by a doctor or other authorised prescriber. At a minimum, a CPOE system can greatly reduce medication errors by ensuring orders are legible, complete and structured (i.e. they include a dose, route and fre- quency). CPOE is likely to be more eff ective at detecting and pre- venting errors when paired with other computerised systems, such as clinical decision support. 2 Clinical decision support (CDS) is computer soft ware designed to assist doctors with decision making. Th ese systems can provide clinical knowledge and guidance as well as patient-specifi c recom- mendations by matching patient characteristics (e.g. age and aller- gies) with rules in the computerised knowledge base. CDS systems can assist with the selection of drugs and dosages (alerts for drug– drug interactions and clinical guidelines), follow-up management (corollary orders and reminders for timely follow-up) and cost reductions (drug formulary guidelines). 3 Bar coding with an electronic medication administration sys- tem can substantially reduce transcription and administration errors, by alerting caregivers to potential errors before they occur. Scanning a patient’s bar-coded wristband, for example, can verify the patient’s identity; scanning the bar code of the item can double check that it is the intended drug. 4 Automated dispensing device (ADD) is a computerised drug storage device that allows medicines to be held and dispensed to a specifi c patient. Th ese devices are more effi cient at reducing medication errors if linked with bar coding and interfaced with hospital information systems. Unit-based cabinets can be placed in patient-care areas to help improve security and accountability for medications. 5 Smart intravenous infusion pumps or infusion devices with decision-support soft ware can prevent errors by ensuring that the right infusion rate and duration are used. Th ese devices can be programmed with standardised concentrations and infusion rates, so that a warning is triggered when settings are outside these estab- lished limits. Th e warning prompts the caregiver to reset the pump or override the alert.

Although health information technology (HIT) holds great promise to reduce medication errors, several studies have high- lighted that unintended consequences associated with HIT

implementation can occur with any implementation; for example, a research report observed an unexpected increase in mortality coincident aft er a commercial CPOE application was implemented. In this single-hospital study, it was reported how critically ill chil- dren had to physically arrive in the hospital and be fully registered into the system before order entry was allowed. Consequent delays in the administration of critical medication were reported, with fewer than half of the patients receiving antibiotics and vasoac- tive infusions within national guideline-recommended timelines. Although CPOE systems are still evolving, they are widely recog- nised as critical for reducing medication errors; ongoing assess- ment of systems integration and human–computer interface eff ects on clinical outcomes is essential.

National and international efforts to reduce medication errors In the wake of the IOM report, the US Food and Drug Administra- tion enhanced their eff orts to reduce preventable harm by dedicat- ing more resources to drug safety. Th e UK Department of Health also issued national guidance on the prevention of medication errors aft er a number of deaths were reported from the intrathecal injection of a Vinca alkaloid. In order to avoid such catastrophic clinical errors, the report recommended the use of non-Luer syringes, allowing intravenous administration of Vinca drugs only. Th ese errors are amongst the most catastrophic; one tragic case is that of Wayne Jowett, as explained in Box 11.1.

In 2006, the World Health Organization launched the High 5s Project, an international collaboration aimed at addressing a num- ber of key patient safety challenges in participating countries. Th is project facilitated the development and implementation of stand- ardised operating protocols (SOPs) to prevent medication errors, one of which targeted the preparation, storage or administration of concentrated injectable medicines, such as concentrated potas- sium chloride solution, sodium heparin and injectable morphine preparations.

Conclusion Medication errors are common and costly. Errors can be prevented by designing safety into the healthcare system. Th e use of HIT, such as CPOE, CDS, bar coding and smart pumps, can undoubt- edly play a key role.

35 C

hapter 11 M edication errors

Figure 11.2 Swiss-cheese model outlining failures in the care of Wayne Jowett

Lack of protocols The protocol at the hospital stated that vincristine and cytosine should not be given on the same day, but on two consecutive days, to prevent any errors in their method of administration. Despite this guideline, and the fact that this was also noted by the pharmacist during the prep- aration of the drugs, these were sent together. The pharmacists dispatching them and the nurse receiving them did not question this, and staff nurse Vallance, who was helping with the lumbar puncture, also failed to notice it.

Drugs Both drugs were packaged in separate boxes, but in one plastic envelope. They were also prescribed on the same prescription chart. In a number of hospitals, intrathecal chemotherapy is prescribed on a separate prescription chart, using a different colour from drugs administered intravenously. This would have alerted Dr Mulhem that only one drug was prescribed and should subsequently be administered. There is now a move to have separate Luer connectors for Vinka alkaloids which prevent wrong route administration.

Lack of knowledge and experience Dr Morton had only been on the ward for 5 weeks and was still learning how things worked. That was his fi rst oncology rotation and he had only administered intrathecal chemotherapy once before, under the supervision of his seniors. Dr Morton had not received any training regarding the intrathecal administration of chemotherapy agents. He assumed that Dr Mulhem was very experienced on chemotherapy and followed his advice without questioning it. However, this was Dr Mulhem’s fi rst job as a registrar. He had no experience with chemotherapy in the past. He had not received any training on chemotherapy.

Supervision: unavailable The question remains as to where senior supervision occurred during this procedure.

Seemingly unavailable

Lack of knowledge and experience

Drugs: problems with design, packaging and delivery

Lack of clear protocols

Training Supervision

Physical barriers

Procedures

The gaps

Tragic death

P art 3 R

isks to patient care

12 Surgical errors Figure 12.1 Surgery as a risky business

Very unsafe Ultrasafe

Microlight aircraft or helicopters Road safety

Surgery

Chemical industry Railways

Nuclear industry

Blood transfusion

Total medical risk

Fatal iatrogenic adverse events

Commercial large-jet aviation

Himalayan mountaineering

High-risk cardiac surgery patients

Source: Adapted from Amalberti (2005)

Figure 12.2 Surgical incidents reported to the National Patient Safety Agency from 2005 to 2008

Between 2005 and 2008, 446,184 surgical incidents were reported to the National Patient Safety Agency, with varying degrees of harm. The specialities with the highest number of incidents were: • Trauma and orthopaedics (33%) • General surgery (30%) • Urology (5%)

‘Admitted on 15th with fractured neck of femur. High volume of trauma admitted this week. Additional list booked on Tues 16th to manage excess trauma. Cancellation of Wednesday

list now left us in a position where this patient is having to wait 4 days for operation’

‘Unplanned arrival of 7 general surgical patients from surgical assessment unit 2 trained staff nurses, 1 healthcare assistant and 1 ward clerk on duty – inadequate ratio. Short on surgical doctors too. Insufficient staff to cope with sudden influx’

* These incidents are verbatim reports of those received by the agency

‘Patient missed her 12:00 and 18:00 drugs as doctor failed to write up chart, despite repeated requests’

‘Patient was booked for exploratory laparotomy. We found 9x9 swab inside patient below the gall bladder. It was not from my initial count for this surgery so must have been retained from a previous surgery’

‘Patient attended pre-op assessment clinic listed for arthroscopy left wrist, some letters plus an X-ray showed left others showed right. Patient says it is his right wrist. Secretary informed’

Severe 0.9%

Moderate 5.4%

Low 22.1%

No harm 71.4%

Death 0.2%

37 C

hapter 12 S urgical errors

Introduction Advances in surgery have been made possible through extra- ordinary technological developments in delivering considerable benefi ts for patients. Th e increasing complexity of procedures, the number of professionals involved in the operating room and the complexity of the healthcare system have made it more diffi cult to deliver reliable and safer care. Th e number of patients who receive precisely the treatment expected could be better; whereas the number of patients who experience consequential errors is high. Other industries like aviation, nuclear and rail have managed to design extremely safe systems. Surgery has a long way to go (Figure 12.1).

Epidemiology of surgery • 234 million surgical procedures are undertaken globally (of note, this is nearly double the annual number of childbirths worldwide) • Eight million surgical operations are carried out every year in the United Kingdom • Although death and complication rates aft er surgery are diffi cult to compare since the range of cases is so diverse, major morbidity complicates 3–16% of all inpatient surgical procedures in devel- oped countries, with permanent disability or death rates of about 0.4–0.8% • Almost seven million patients undergoing surgery have major complications and one million die during or immediately aft er sur- gery every year worldwide • Nearly half of these surgical adverse events have been identifi ed as preventable • Figure 12.2 gives a fl avour of the surgical incidents reported to a large patient safety reporting system

Technical versus non-technical skills Technical skills include many diff erent processes, including the recall of factual information, diagnosis and performing of prac- tical procedures. Non-technical skills (human factors, discussed in chapter 5) are cognitive (e.g. decision making) and interper- sonal (e.g. teamwork) skills. Analyses of adverse events in surgery indicate many underlying causes originating from behavioural or non-technical aspects of performance (e.g. poor communication between team members in the operating theatre) rather than a lack of technical expertise (most surgeons are good at operating).

Non-technical skills can be divided into four categories: • Communication and team working – skills that ensure the trans- fer of information within and between clinicians and allied profes- sional groups to maintain a shared knowledge and understanding of the patient (chapter 5) • Leadership – guiding the team by providing direction, demon- strating high standards of clinical practice and care and being con- siderate about the needs of individual team members • Situation awareness – developing and maintaining a dynamic awareness of the situation in theatre based on assembling data from the environment (patient, team, time, informational displays and equipment), understanding the impact of each environmental infl uence and anticipating what may happen next • Decision making – utilising diagnostic skills for the situation and reaching a judgement in order to choose an appropriate course of action

Non-technical skills are as essential as technical skills when it comes to maintaining high levels of performance over a period of time.

Retained surgical materials Th e economic impact of surgical complications relating to reten- tion of surgical materials cost the UK taxpayer around £2 million per year, according to the NHS Litigation Authority (NHSLA). Th is traumatic experience is not limited to patients, but extends to their families. Th e human costs of short- and long-term follow-ups and the psychological implications are severe.

Th e incidence of retained swabs ranges from 1:9000 to 1:19,000 surgical procedures. Considering that 234 million surgical opera- tions occur globally, such statistics are unacceptable. In most oper- ating rooms, a system should exist to verify swab and instrument counts before and aft er surgery. However, most systems are labour intensive, and are perceived to interfere with surgery and other aspects of care, which may also be prone to error. Around one in eight operations involves at least one counting discrepancy, which increases the likelihood of a retained sponge and/or instrument. Sixty per cent of discrepancies detect a misplaced item, which would be categorised as a near miss. A recent 12-month analysis of all patient safety incidents reported to the UK National Patient Safety Agency’s (NPSA) National Reporting and Learning System – the largest repository of medical errors in the world – revealed 496 incidents relating to retained surgical equipment in 5 years. Th e majority (94%) of those failures were fortunately prevented and corrected before any harm came to the patient.

Th e NPSA has taken a lead in identifying and monitoring reports of a national set of ‘never events’ for England and Wales. Th ese events are serious and deemed to be completely preventable. One of the original eight ‘never events’ is retained instrument post- operation. Further work continues to develop technology to reduce human error, for example the development of bar-coded surgical sponges and surgical instruments, which are scanned by the nurses before being handed to the surgeon and vice versa.

Wrong-site surgery Wrong-site or wrong-patient incidents are rare, but the conse- quences can result in considerable harm to the patient. It is oft en catastrophic for the patient (although perhaps less frequently fatal) and is a popular (and clearly delineated) error topic for reporting by mainstream media. Th e cause appears to be systemic predispo- sition. A recent study revealed 5940 cases of wrong-site surgery (2217 wrong-side surgical procedures and 3723 wrong-treatment or wrong-procedure errors) in 13 years. A particular interest has arisen in spinal- and neurosurgery, where wrong-site surgery occurs in between one in 4550 and one in 780 cases, depending on the procedure. Other factors such as fatigue, time pressure, changes to the operation, unusual patient anatomy and radio- graphic problems have all been identifi ed as common risk factors. A one-year review of the database of errors housed at the NPSA revealed 353 cases of the wrong side marked on the theatre list. Th ere were 150 cases of the wrong side marked on the consent form and 46 cases of wrong-side surgery (including some cases of burr holes and craniotomies being performed on the wrong side). Inclusion of wrong-site surgery as a never event (discussed in this chapter) and the use of checklists could help prevent this type of error.

Checklists Checklists are not a new phenomenon and have been used regu- larly in high-risk industries such as oil mining, nuclear energy and aviation. In January 2007, the World Health Organization (WHO)

began a programme aimed at improving the safety of surgical care globally. Th e initiative, called Safe Surgery Saves Lives, is aimed at identifying minimum standards of surgical care that can be uni- versally applied across countries and settings. A core set of safety checks were identifi ed in the form of a WHO Surgical Safety Checklist (Figure 12.3). It was designed to be used in any surgi- cal setting and operating theatre environment. Each step on the checklist is simple, widely applicable and measurable, and it has already been demonstrated that its use can reduce death and major complications.

The checklist in England and Wales – a national perspective In February 2009, the NPSA issued an alert requiring all hospi- tals in England and Wales to implement the WHO Surgical Safety Checklist by February 2010. One of the main drivers for success was the lead taken by the Patient Safety First campaign in England and the 1,000 Lives Campaign in Wales in rolling out the checklist as part of their broader functions. Th ese initiatives apply the princi- ples of a social movement – target individuals, engage on a personal level to inspire action and provide materials to make a diff erence locally, which can then be scaled up at a national level. Th ey cre- ated an inspirational movement that empowered frontline staff to bring about change. By April 2010, all hospitals had begun imple- menting the checklist using diff erent approaches and with vary- ing degrees of success. However, more than two-thirds reported that the checklist improved teamwork and safety, and almost 40% reported that they were able to capture more near-misses.

Th e task of implementing this simple life-saving tool has chal- lenges: • Negative clinician attitude and poor engagement (‘I am a sur- geon and do not make mistakes. Why should I be questioned by the nurse who is using the checklist?’) • Tendency to view the checklist as a mere tick-box exercise (‘If I tick all the boxes haphazardly, I have done the checklist ’) • Placing the checklist as a low priority (‘We need to think about getting more theatre time and doing more operations’) • Lack of clinical leadership (‘Surely, as surgeons and anaesthe- tists, we know what is best for the patients; does everyone in the team need to be counted as essential to patient care?’) • Poor acceptance of the strong evidence base of the checklist (‘Th e checklist was assessed as part of a multi-centre study that involved some developing countries, so it cannot be that good for the UK’)

• Poor implementation strategies (‘Th e managers and clinicians have diff ering agendas’)

A lean intervention – reducing surgical errors locally Th e surgical emergency unit at the John Radcliff e Hospital is a 38-bed acute surgical ward in a teaching hospital in the United Kingdom. It receives all general surgical emergency admissions to the hospital. An audit of patient safety revealed that there was more than a 10% rate of harm experienced by patients on this ward due to defi ciencies in compliance with recommended practice for important care processes – track and trigger tools (discussed in Chapter 16) were used in 77% of patients, and fl uid balance charts were completed in 11% of patients. It was stipulated that a lean intervention (discussed in Chapters 26 and 27) could improve patient safety on the ward. Seven safety processes were selected, and elements of lean methodology were applied: • Direct verbal communication between medical and nursing teams on daily rounds increased by 37% when visual aids and a sample protocol of individual roles were created • Correct administration of prophylaxis for deep vein thrombosis (DVT) increased by 52% when a preprinted drug chart and self- audit were introduced • Reduction of patients with drug-prescribing errors by 13% was achieved through prompt cards for commonly misprescribed drugs that were given to staff • Use of alcohol gel for hand hygiene on entering ward increased by 8% as a result of better visual cues and more alcohol gel bottles being provided at convenient locations in the ward • Correct use of venous site infection protocol increased by 33% • Adequate monitoring of patients’ vital signs and recording of their risk scores increased by 31% via a basic care checklist and a track and trigger tool (see Chapter 16) was implemented • Adequate completion of fl uid balance charts increased by 1% through no active intervention

Conclusion Training in surgery focuses on technical skills. Whilst essential, this fails to recognise that surgeons cannot perform to the best of their technical ability unless they work in a well-functioning team (non-technical skills). Improved training in non-technical skills and use of tools such as the checklist can ensure that we practice safe surgery.

P art 3 R

isks to patient care

39 C

hapter 12 S urgical errors

In 1935, the U.S. Army Air Corps held a flight competition for airplane manufacturers vying to build its next-generation long-range bomber. In early evaluations, the Boeing plane had trounced other designs. The flight “competition”, was regarded as a mere formality. With the most technically gifted test pilot in the army on board, the plane roared down the tarmac, lifted off smoothly, and climbed sharply to 300 feet. Then it stalled, turned on one wing, and crashed in a fiery explosion. Two of the five crew members died, including the pilot. An investigation revealed that nothing mechanical had gone wrong. The pilot had forgotten to release the new locking mechanism on the elevator and rudder controls. A few months later, army pilots were convinced the plane could fly and invented something that would be used on the few planes that had been purchased… . A checklist, with step-by- step checks for takeoff, flight, landing, and taxiing. With the checklist in hand, the pilots went on to fly the model (B-17) a total of 1.8 million miles without one accident, and helped the US army launch its successful bombing campaign across Nazi Germany.

Before induction of anaesthesia Before patient leaves operating room

Confirm all team members have introduced themselves by name and role

Surgeon, Anaesthesia professional and nurse verbally confirm • Patient • Site • Procedure

Anticipated critical events

Surgeon reviews: what are the critical or unexpected steps, operative duration, anticiptaed blood loss?

Anaesthesia team reviews: are therw any patient-specific concerns?

Nursing team reviews: has sterility (including indicator results) been confirmed? Are the equipment issues or any concerns?

Has antibiotic prophylaxis been given within the last 60 minutes? Yes Not applicable

Is essential imaging displayed? Yes Not applicable

Patient has confirmed • Identity • Site • Procedure • Consent

Site marked /not applicable

Anaesthesia safety check completed

Pulse oximeter on patient and functioning

Does patient have a: Known allergy? No Yes

Difficult airway/aspiration risk? No Yes, and equipment/assistance available

Risk of >500mL blood loss (7mL/kg in children)? No Yes, and adequate intravenous access and fluids planned

Nurse verbally confirms with the team:

The name of the procedure recorded

That instrument, sponge and needle counts are correct (or not applicable)

How the specimen is labelled (including patient name)

Whether ther are any equipment problems to be addressed

Surgeon, anaesthesia professional and nurse review the key concerns for recovery and management of this patient

This checklist is not intended to be comprehensive. Additions and modifications to fit local practice are encouraged

Source: Reproduced with permission of WHO

Before skin incision

• Sign in – prior to anaesthesia ensures adequate preparation is made for any predictable difficulties and that the expected patient is about to receive anaesthesia

• Time out – just prior to incision is the final check that everything is in place for the procedure to take place in the safest environment possible. It is the final check that the right thing is about to be performed on the right patient, that everyone knows what the surgeon and anaesthetist are thinking or expecting, that the right equipment is available, that required imaging is displayed and that appropriate infection and venous thromboembolism prophylaxis is in place

• Sign out – ensures that the correct information will be given to recovery staff, that nothing is missing from the scrub trolley and that key specimens are correctly labelled

The checklist outlines essential standards of surgical care and is designed to be a simple tool to improve surgical safety It consists of three key phases:

Surgical Safety Checklist (First edition)

Figure 12.3 The WHO surgical safety checklist

P art 3 R

isks to patient care

13 Diagnostic errors Figure 13.1 Etiology and outcomes of diagnostic error

Physical exam

Synthesis and hypothesis

History Tests Follow-up Atypical or masked symptoms, deceit, diseases not yet defined or that lack a sensitive test, etc

Diagnostic error

Missed, wrong, or delayed diagnosis

No significant impact to patients, providers,

or systems

Diagnostic process error

Outcomes

Root causes

Patient harm

A 48-year-old male presented to A&E with cough, fever, hypox- emia, blood-tinged sputum, a tongue ulcer and an infi ltrate on chest X-ray. The fi ndings were attributed to community-acquired pneumonia. He was admitted to the intensive care unit (ICU) and started on broad-spectrum intravenous antibiotics. A Nephrol- ogy consultant was called to evaluate an elevated creatinine level and suggested the possibility of Wegener’s granuloma- tosis. A sensitive and specifi c test for Wegener’s was ordered (an anti-neutrophil cytoplasmic antibody, or ANCA, test). Five days later, the patient died of pulmonary haemorrhage and Wegener’s granulomatosis was confi rmed at autopsy. The ANCA test was never sent because it required a special form for ‘send out’ tests that was never completed, and the email the lab had sent to the ordering MD had never been read. Anti- biotics are ineffective in treating Wegener’s granulomatosis, but this often fatal disease responds well to immunosuppression.

Box 13.1 Case example of diagnostic error: wrong diagnosis of pneumonia

Introduction A diagnostic error describes situations in which the correct diag- nosis is missed altogether, signifi cantly delayed or simply wrong. Some of these situations are unavoidable, such as when the symp- toms are masked or atypical, current tests lack sensitivity or the disease is at a very early stage. More typically, however, diagnos- tic errors refl ect a preventable breakdown in one or more steps – cognitive, system-related or both – of the diagnostic process (Figure 13.1).

How do we arrive at a diagnosis? Most diagnoses are made from the history and physical examina- tion alone, and many errors can be avoided by paying special atten- tion to these ‘routine’ aspects of patient evaluation. Other errors arise from problems with diagnostic testing, such as with perfor- mance, interpretation and follow-up of labs, imaging and other types of diagnostic test or procedures.

Th e synthesis or reasoning phase of the diagnostic process is especially important in deriving the correct diagnosis, and is the step most prone to cognitive errors. Clinical reasoning is oft en described by the ‘dual process’ paradigm (shown in Figure 13.2). System I: When we encounter a new patient or a new problem, our brains instantaneously decide if the problem is recognised or not. Does it resemble something we have seen before, or something we learned during training? If the problem is recognised as something familiar, the diagnosis emerges eff ortlessly and quickly, based on our intuition, or what is more formally called ‘System I’. System I is an automatic, subconscious thought process that typically works quite well. In the case example in Box 13.1, the physician immediately recognised that the symptoms of cough, fever, hypoxemia and blood- tinged sputum were suggestive of pneumonia, which became the work- ing diagnosis. Intuitive decisions work well for very common problems like pneumonia, but lead to errors when the real problem is some- thing uncommon, like Wegener’s granulomatosis. True experts use intuitive, System I processing almost exclusively because they are so familiar with the content in their area. By defi nition, experts make the fewest mistakes, part of the reason that we trust intuitive deci- sion making so strongly. Unfortunately, intuitive problem solving is

41 C

hapter 13 D iagnostic errors

error-prone, especially if we are not experts. Intuitive processes are prone to a wide range of infl uences, both cognitive and emotional, that detract from reliable decision making. Some of the more com- mon problems are detailed in Table 13.1.

Table 13.1: Common problems using System 1 (intuitive) problem solving

Jumping to a conclusion – Also called premature closure or satisfi cing, this refers to the tendency to accept the fi rst reasonable solution that comes to mind and failing to consider alternatives. Context errors – In trying to make sense of a new situation, the wrong context is envisioned. An example: Assuming the cause is gastrointestinal in a patient presenting with abdominal pain when the cause could be something else, such as shingles or referred pain.

Seeking confi rmation only – Th is refers to the tendency to only look for evidence to confi rm a diagnosis rather than looking for evidence to refute it, which may be more defi nitive.

Diagnostic inertia – Once established, a diagnostic label persists even if new evidence suggests some other possibility.

System II: If we do not recognise a pattern in the presenting symp- toms, we revert to the tried-and-true method of solving puzzles – we apply deliberate, rational thought. Success in this eff ort will depend on our depth of knowledge, our ability to use evidence- based medicine and our skill in reasoning. Th is mode of problem solving, ‘System II’, is also error-prone, and it requires much more time, eff ort and attention than System I, but under normal circum- stances it works. In the case example, the physician should have realised that the elevated creatinine and tongue ulcer were not com- monly associated with pneumonia. If the physician had switched to the analytical, or System II, mode of reasoning, this might have led to the consideration of other diagnostic possibilities, thus avoiding the diagnostic error that was made.

All students start off using analytical reasoning (System II) exclusively, and as they gain knowledge and experience, more and more problems are solved using intuition (System I). When estab- lishing a diagnosis, it is likely that some elements of System II will be at work; for example, when you refl ect, at least briefl y, on whether you are comfortable with the diagnosis or need to con- sider other possibilities.

How common are diagnostic errors? Diagnostic errors are uncomfortably common. It is estimated that 40,000–80,000 preventable deaths occur from misdiagnoses each year in the United States. Furthermore, experts assert that 10–15% of all diagnoses are incorrect (with a small fraction leading to harm) and approximately 5% of people are misdiagnosed every year in the outpatient setting. Autopsy studies routinely fi nd major diagnostic discrepancies in 20–30% of cases. Using chart reviews of discharged inpatients, the landmark Harvard Medical Practice Study identifi ed adverse events in 7% of inpatients, and diagnostic mishaps were among the most common causes. Diagnostic errors generally make up the largest fraction of malpractice claims, result in the highest payments, and are the most diffi cult to defend in the court of law.

In addition to causing overt harm, diagnostic errors can lead to delayed, unnecessary or inappropriate testing and treatment. Besides the stress and concern of the involved patients, these cas- cades increase healthcare costs through problems with underuse, overuse and misuse of diagnostic testing strategies. Making the correct diagnosis, although complicated, is obviously critical to both obtaining optimal patient outcomes and using healthcare resources effi ciently.

What factors contribute to diagnostic error? Th e root causes of diagnostic errors include two categories that doctors or health systems can infl uence, system-related and cogni- tive factors, and a third category, ‘no-fault’ factors that are largely outside their control. Th ese factors might operate synergistically, and typically multiple, contributory factors can be identifi ed in a single case. System-related and cognitive contributions are identi- fi able in most diagnostic error cases, and ‘no-fault’ errors are less common (less than 10%) (Figure 13.3). System-related contributory factors: refer to all the character- istics of the healthcare system and the operating environment that impact the safety and quality of care provided. Are there too many distractions? Is the workload balanced enough to provide each patient the attention they need? Are expertise, tests and equipment available when needed? Is care coordinated across sites and con- sultants? Are handoff s performed regularly and adequately? Is the staff well trained and engaged in their assigned work? Is the care process patient centred? Are trainees adequately supervised?

In the Wegener’s granulomatosis case example, several system- related factors can be identifi ed: • Th e lab required a complex administrative process for a test that should be performed expeditiously, and assumed that the staff they sent the message to read their emails regularly • Th ere was no process in place to ensure that critical tests were per- formed and reported on a timely basis. A process to ‘close the loop’ might have prevented this error by ensuring that the email was read and acted upon Cognitive-related contributory factors: largely fall within three categories: issues related to faulty knowledge, faulty data gathering and/or faulty synthesis. Faulty knowledge includes the physician lacking knowledge of a particular disease or the diagnostic skills to diagnose the current condition. Faulty data gathering involves incomplete gathering of the patient’s history and examination, failure to order the appropriate tests or performing or interpret- ing diagnostic tests incorrectly. Faulty synthesis is by far the most common type of cognitive error, and refl ects a failure in ‘putting it all together’. Th e synthesis phase of diagnosis is especially prone to

Patient presentation

Physician recognises pattern?

Yes No

System I

Automatic, subconscious, heuristic-based

System II

Conscious, deliberate, rational consideration

Patient diagnosed

Figure 13.2 The dual-process model of clinical reasoning

errors due to the many factors that detract from optimal cognitive performance. Th e intuitive diagnoses we derive using System I can be degraded by a wide range of problems, the most common of which are framing bias, context errors and premature closure.

System II errors commonly refl ect inadequate knowledge about all the competing diagnostic possibilities, but can also go astray because of many other factors. Both System I and System II can be negatively impacted by stress, distractions, fatigue and aff ective issues, such as dealing with a relative or friend, or a patient who is disliked for some reason.

In the Wegener’s granulomatosis case, several cognitive root causes can be suspected: • Th e physicians who initially evaluated the patient prematurely accepted the diagnosis of pneumonia because it seemed to explain the major fi ndings, without considering other, more esoteric, possibilities in light of additional fi ndings • Th e ICU physicians who accepted the patient were infl uenced by the framing eff ects of the original ER diagnosis; they did not rethink the case on their own • All of the treating physicians may have been biased by the pre- senting context. That is, the complaints of cough and the find- ings of infiltrate suggested a primary pulmonary etiology. In this case, the correct diagnosis of a primarily vascular process, Wegener’s granulomatosis, may not have been considered because additional clues (tongue ulcer and creatinine) were not given adequate priority No-fault contributory factors: refer to root cause factors beyond the control of the clinician or the local standard healthcare system. Th is can include patient-related factors, such as when a patient fails to follow through with recommended diagnostic tests or follow-up

appointments, or patients who are purposely misleading. Dis- eases that present atypically or at a very early stage, or fi ndings that are non-specifi c, would also fall into this category. Another major category is situations where appropriate diagnostic tests are not yet available to assist in diagnosing a particular disease, but may evolve in the future. Similarly, some diseases can be diagnosed by exceedingly expensive or restricted tests that are not generally available in everyday practice.

How do we avoid diagnostic errors? Diagnostic errors can be avoided by using a number of techniques. To begin with, this includes carrying out a thorough history and physical examination, and gathering all of the relevant background data. Be especially wary if the patient cannot communicate, or if someone else (the patient, another provider or the A&E doctor) suggested the diagnosis. Watch out for context and framing errors. It is worthwhile to spend a moment mentally reviewing some of the ‘high-risk’ situations (see Table 13.2) that predispose to diag- nostic error.

Table 13.2 High-risk situations for diagnostic error

Is this patient unable to communicate their own story?

Are there ‘must-not-miss’ diagnoses that need consideration?

Did I just accept the fi rst diagnosis that came to mind?

Was the diagnosis suggested to me by the patient, a nurse or another MD?

Are there data about this patient I have not obtained and reviewed? Old records? Family? Primary care provider?

Are there any pieces that do not fi t?

Were the X-rays read by a radiologist?

Was this patient handed off to me from a previous shift ?

Was this patient seen in A&E or a clinic recently for the same problem?

Was I interrupted, distracted or cognitively overloaded while evaluating this patient?

Is this a patient I do not like for some reason? Or like too much (e.g. a friend or relative)?

P art 3 R

isks to patient care

46%

19%

28%

No-fault factors

System factors

Cognitive factors

System and cognitive factors

Figure 13.3 Contributors to diagnostic errors.

43 C

hapter 13 D iagnostic errors

Cognitive-related interventions When asked why they missed the correct diagnosis, the most common answer is ‘I just did not think of it’. Th e best antidote to avoiding this problem is to invoke analytical (System II) reasoning. Th is involves taking a diagnostic ‘time out’ and pausing to refl ect on how the diagnosis was established and whether it is trustwor- thy; thinking ‘What else could this be?’; and trying to construct a diff erential diagnosis for every case seen, even if the patient already carries an assigned diagnosis.

Students can work towards increasing their knowledge and expertise, which in turn will improve clinical reasoning skills, by getting additional training and/or taking rotations in domains they are less familiar with.

Students must also never hesitate to ask for help. One can begin by soliciting second opinions from peers or other members of the team. Requesting consultation from a specialist is always war- ranted for cases one is unsure about. Having someone else think through the case independently is a powerful way to avoid diag- nostic errors.

System-related interventions Changes to the healthcare environment may also contribute to enhancing the reliability of the diagnostic process and preventing diagnostic error. Although such changes typically require actions and resources from more senior healthcare administrators, it is the responsibility of all frontline clinicians to bring opportunities for improvement to their attention.

Such changes may include improving clinical workfl ow, mak- ing sure training and supervision are appropriate, having specialist consultants available if needed and providing or enhancing medi- cal hardware and soft ware to support optimal diagnosis. Using a robust electronic medical record system with built-in decision support soft ware can help to improve diagnoses in many ways. For example, the electronic record can improve communication and make it easier to order, fi nd and follow test results. Electronic alerts can help avoid situations, like the Wegener’s granulomatosis case example in Box 13.1, where essential test results have not yet returned.

Web-based decision support tools can be helpful in expanding the list of diagnostic considerations. A free online checklist to assist with diff erential diagnoses of common complaints is available at

http://pie.med.utoronto.ca/DC/index.htm. Alternatively, simple tools like the VITAMIN CC&D mnemonic can bring other diag- nostic possibilities to mind.

Changes in hospital policies and standard practices may also enhance the diagnostic process. For example, routinely following up with patients recently discharged from the hospital, or from an A&E visit, is a potential intervention that may reduce diagnostic errors and improve patient satisfaction at the same time.

Clinicians should also bring examples of diagnostic error to attention, so that they can be studied with the goal of helping avoid similar errors in the future. Th ese exercises can lead to projects to redesign systems of care, and bring attention to cognitive issues as well. Finally, clinicians should always try to learn from their cases by requesting autopsies.

Involving patients Patients should be partners in establishing the correct diagnosis, especially because they have the most at stake (Chapter 28). Clini- cians should ensure that their patient knows how and when to get back to them if the symptoms change, or if they do not respond to treatment as the clinician thinks they should. Clinicians should always keep an open mind, be alert to fi ndings that do not fi t with the working diagnosis and remember that every diagnosis is just a probability, not a certainty.

VITAMIN C C & D

V ascular I nfection & intoxications T rauma & toxins A uto-immune M etabolic I diopathic & iatrogenic N eoplastic C ongenital C onversion (psychiatric) D egenerative

P art 3 R

isks to patient care

14 Maternal and child health errors Figure 14.1 The ‘3-delays’ model

Delay 1 Delay 3

Demand: Delays in

deciding to seek care

Supply: Delays in receiving

routine and emergency high quality

care

Linkage: Delays in identifying

and reaching

appropriate care setting

Delay 2

Figure 14.2 Linking communities to health facilities: roles and responsibilities of the task forces

Health centre

Outreach nurses

Task force

Villages

• Reliable antenatal care • Basic obstetric care • Rapid and effective referral to emergency obstetric care

• Identification of pregnant women (village) • Identification of high risk pregnant women (health centre) • Communication of needs of pregnant women, feedback of customer and provider satisfaction

• Surveillance for new pregnancies • Close follow up of all pregnancy care • Effective transport solutions for high risk pregnant women

Figure 14.3 The Institute for Healthcare Improvement’s (IHI) Improving Perinatal Care driver diagram – a system perspective of patient safety

• Align unit measures, strategies, projects with org strategy and goals (clinical, patient, exp. financial and workforce) • Channel senior leadership attention and develop unit leadership • Engage physicians • Build improvement capacity and provide resources for improvement • Establish a just culture • Develop a competent trained and available workforce • Establish credentialing of core competency and training for all providers • Use American Congress of Obstetricians and Gynecologists (ACOG)/Association of Women's Health, Obstetric and Neonatal Nurses (AWHONN) guidelines for documentation and staffing • Develop a consumer advisory board

• Execute care that meets national standards (implement bundles, perinatal core processes) • Develop standard processes and protocols for response to obstetrical emergency • Design care process improvement based on trigger tool analysis, event detection, sentinel event • Standardise administration of high alert medications – oxytocin, magnesium sulfate, epidurals • Create an environment that supports care and healing • Consider segments of population and design reliable and appropriate processes for specific needs and characteristics of this segment of the population

• Adopt common language and interpretation of efm with multi-disciplinary training, i.e. National Institute of Child Health and Human Development (NICHD) criteria • Implement techniques for effective communication, i.e. SBAR • Establish reliable techniques for handoffs • Establish team response protocols • Implement huddles • Design simulations

• Design processes to support partnership in care between provider and patient and family • Develop with patient a customised interdisciplinary shared care plan • Design care process improvement based on information obtained about patient experience (interviews, assessments, focus groups, surveys) • Include patients and families on design and improvement teams • Communicate openly and honestly with family and patients at regular intervals • Do what you say, mean what you do

Perinatal community • Reducing harm • Improving care • Supporting healing

Perinatal leadership

Reliable design/reduce

variation

Effective peer

teamwork

Respectful patient

partnership

Key outcome

and process

measures*

* See perinatal community measurement strategy

45 C

hapter 14 M aternal and child health errors

Introduction Because most high-income countries provide universal access to care, much of the focus on improving maternal outcomes in these countries has been on improving the reliability of safe obstetric care practices. In low- and middle-income countries (LMICs), a much broader eff ort is underway to tackle the enormous problem at hand: Africa continues to record rates of maternal death that were last seen in ‘northern’ countries over a century ago. For exam- ple, in 2010, the maternal death rate in high-income countries was 15/100,000, and in sub-Saharan Africa, the rate was 500/100,000 – a more than 30-fold diff erence. But progress in wealthy countries is not uniform; mortality rates can vary signifi cantly depending on race and income levels. In the United States, the maternal mortal- ity rate (MMR) among non-Hispanic black women (28.4/100,000) was roughly three times the rate of that among non-Hispanic white and Hispanic women overall (10.5 and 8.9/100,000, respectively).

In all countries, both the supply (reliability of delivered care) and the demand (community and patient) aspects of the system need to be addressed, as well as linkage structures to connect improved demand with more reliable supply. Major opportunities exist to ensure that knowledge of what works at all levels of the health system is transferred between these diff erent settings. Th e ‘3-delays’ model, a comprehensive approach to obstetric safety, has become popular as an organising framework for improv- ing the care of pregnant women from the community to facilities (Figure 14.1). Th e model has informed work in Asia and Africa on improving decision making amongst pregnant women in Africa; the generation of novel community-facility structures that link pregnant mothers safely to the best place of care; and an adapta- tion of the Breakthrough Series quality improvement (QI) model (www.ihi.org), which is testing network thinking to inform a novel improvement team design that includes a web of players who are connected to a pregnant woman’s journey to safe childbirth.

Delay 1: Demand For many years, health systems have brought together pregnant women in high-income countries to educate and motivate good practice, and inform them of their options for safe and comfortable care during childbirth. While the social and educational benefi ts of these interactions are widely touted, the evidence of their eff ec- tiveness in improving decision making and safe outcomes is lack- ing. By contrast, community mobilisation through women’s groups has been shown to be eff ective and cost-eff ective in changing care and care-seeking practices, and in reducing neonatal mortality in a range of settings in LMICs. Th e Women’s Group intervention brings together women of childbearing age from a single village or group of small villages to undergo a structured, facilitated pro- cess of identifying, prioritising and deciding on an intervention

to improve local needs, and then undertaking the intervention together. Th e typical journey from initial discussion to women- led intervention involves monthly meetings of the group over a 9-month period. A recent meta-analysis of women’s groups in Asian and African countries has shown that these structures are associated with a signifi cant decrease in both neonatal and mat- ernal mortality. Th e exact mechanism of action that links women’s groups to this decrease is not yet fully understood (Prost A 2013).

Delay 2: Linkage Delay in reaching an appropriate and safe place of obstetric care is a signifi cant determinant of poor obstetric outcomes in LMICs. Numerous studies have attributed a large portion (up to 73%) of preventable maternal deaths to issues related to failure of referral systems that ensure pregnant women receive the care they need, when they need it. While planning for women at high risk of obstetric complications requires particular attention, the value of over-reliance on stratifi cation of pregnant women into ‘high risk’ and ‘low risk’ for obstetric complications has been questioned on the grounds that risk is hard to predict. Safe sys- tems ensure that all women have easy access to reliable obstetric care, that unexpected emergencies can be promptly managed or referred and that high-risk women (Figure 14.2) are at an appro- priate level of care at the start of labour. Innovative approaches that promote strong community facility linkages are being tested in countries with high rates of maternal mortality and unreliable access to care. In Malawi, health centres capable of basic obstet- ric care are being linked to communities through ‘task forces’ – committees, under the supervision of tribal authorities, which include village representatives and outreach nurses – that ensure safe passage to reliable antenatal and obstetric care for all preg- nant women in the surrounding communities. In Ghana, refer- ral and transport problems are being addressed through a novel adaptation of the Institute for Healthcare Improvement’s learn- ing networks. To ensure that community knowledge and inputs are maximised, the improvement team membership (which typi- cally includes facility-based members) was expanded to include representatives of the community who have key roles in ensuring that pregnant women get to the right place at the right time to ensure safe outcomes. As such, the improvement teams included conventional care team members (midwives, doctors and clinic staff ) as well as non-mainstream health purveyors (traditional healers, traditional birth attendants and chemical shop sellers), community leaders (spiritual and civic) and transport (taxi and ambulance) representatives. Th ese improvement teams use the same QI principles (data-driven decision making, and testing of local ideas) as with more conventional QI projects to achieve their aims.

Delay 3: Supply Improving the reliability of obstetric care has been key to eff orts to improve safety and outcomes of childbirth. Using QI methods to improve the reliable delivery of evidence-based obstetric care is the basis of a number of successful interventions in high-, middle- and low-income countries. In Europe and North America, the per- inatal collaboratives have brought together multiple hospitals and health systems around a set of concepts that provide the basis for systematic eff orts to improve outcomes. Th e pace of improvement is accelerated when these healthcare units are networked, as this enables learning between sites.

Aft er 2 years as an innovation community, IHI launched the IHI Improving Perinatal Care Collaborative in 2005 and it has con- tinued since then. Over 150 hospitals and hospital systems in the United States and elsewhere have joined to focus on the reliable delivery of safe care to the obstetric patient in labour and delivery. Th e collaborative is designed with four main primary drivers: peri- natal leadership, reliable design, eff ective teamwork and respectful patient partnership. Th e core theory, the Idealised Design of Peri- natal Care, centres on a safety model of prevention, identifi cation and mitigation which supports continuous quality improvement and having a system perspective (Figure 14.3).

Th e application of clinical bundles provides the infrastructure to the improvement methodology (Chapter 32). A care bundle is

a set of evidence-based practices that improve patient outcomes when applied collectively and reliably. Th e actual measurement of these care bundles is what sets them apart from traditional measurement – the care bundle is measured as a whole, rather than measuring the individual components. Th is is termed ‘all- or-none measurement’. As an example, if there are four accepted components that are necessary for appropriate care to occur, if one component of care is not delivered to the patient, the measurement is reported as 0%; the patient either gets all four (100%), or if any one is not completed the score is counted as zero (0%). Th is type of all-or-none measurement is a driver in changing the knowledge of a reliable delivery system of care.

Conclusion Improving the safety and reliability of delivery of care is a journey and a process of continuous improvement. It involves changing the current paradigm of the system, which in many places remains hierarchical, to one that involves all stakeholders (providers, nurses, patients and families, administrators and communities) as necessary and needed change agents who function together as a team with one outcome in mind – to do no harm.

P art 3 R

isks to patient care

47 C

hapter 15 S lips, trips and falls

Figure 15.1 Multifactorial assessment and intervention

Cardiovascular examination including lying and standing BP

Perceived functional ability and fear of falling

Visual impairment Environmental hazards Medical review Neurological examination

Cognitive assessment (dementia and delerium)

Urinary incontinence, frequency or urgency

Gait, balance and mobility

Oesteoporosis and fracture prevention

Slips, trips and falls15

P art 3 R

isks to patient care

Introduction Falls are defi ned as: ‘an event which results in a person coming to rest unintentionally on the ground or other lower level, with or without loss of consciousness’. Falls are a symptom of a number of interacting risk factors, not a disease, so when an older person seeks medical attention aft er a fall, there is an opportunity not only to prevent further falls and injury but also to reduce morbidity and mortality from treatable underlying conditions.

Why slips, trips and falls are a major concern In the United Kingdom, over 1.5 million older people fall in their own homes each year, and around 300,000 osteoporotic fractures, including 75,000 hip fractures, occur. Risks increase with age; a third of people aged over 65 years will fall each year, and half of those aged 85 years will fall. Rates of falls for older people living in care homes can be three times higher than for people living in the community. Th ese falls and injuries can have serious conse- quences: • One-third of older people who fall will develop fear of falling, aff ecting their activity, lifestyle and wellbeing • Two-thirds of patients with hip fracture do not regain their pre- vious levels of independence • One-third of hip fracture patients do not return to live in their own home • Every 5 hours, an older person will die as the result of a fall

Slips, trips and falls in hospital Many patients who fall are admitted to hospital – either for treat- ment of their injuries, or because the falls are symptoms of an underlying clinical condition. In the unfamiliar environment of the hospital, and with the eff ects of acute illness on top of long- standing illness, the risk of falls is even greater. Falls are the most commonly reported patient safety incident in almost all acute and community hospitals and mental health units, with over 280,000 falls reported in England and Wales each year and over 1,000 falls per year reported per acute hospital (of average size). Falls in hos- pital predominantly aff ect and harm older, frailer people; over 80% of falls in hospital occur in patients aged over 65 years, with the highest falls rates and the greatest vulnerability to injury seen in patients aged over 85 years.

Taking a falls history As part of every admission for all patients aged over 65 years, and for younger patients with conditions that aff ect their mobility, routinely ask about any falls in the past year. Older people can be reluctant to admit falling, seeing falls as unimportant, or believ- ing they could be seen as foolish or helpless. Phrasing the question as ‘Any falls, faints or dizzy spells?’ can help them to see that this

is an important part of their medical history. If falls are reported, explore: • Where did they fall? In their own home, or outdoors? What inju- ries? Have the falls left them fearful? • When did they fall? Is there any pattern of time of day or repeated falls, or any recent increase in the frequency of falls? • What were they doing? Were they walking, or had they just got up from their bed or chair? • Do they recall slipping or tripping? Do they remember hitting the fl oor? Did they suff er from transient loss of consciousness (T-LOC)? • Did they have any associated symptoms? • Did anyone witness the fall? What did they see, and does it match the patient’s account?

Ensure you have checked if they are cognitively able to give a reliable account of the fall. Even for cognitively intact patients, wit- nesses may be able either to corroborate or contradict the patient’s recall, or may have noticed signs the patient was unaware of.

‘Red fl ags’ for syncope and seizure Th irty per cent of blackouts in overs-65s present as unexplained falls. Th e patient should be investigated for syncope or seizure if: • Th e fall is not convincingly explained or recalled (with T-LOC, the patient may say, ‘I just found myself on the fl oor’) • Th e pattern of injury does not fi t the patient’s recall of how the fall occurred • Facial injury – conscious patients will not normally fall face-fi rst. • Th ey complain of dizziness, chest pain or palpitations

Remember that syncope is common (it aff ects up to 50% of people during their lifetime) and seizure is rare (around 1%)

No such thing as a ‘mechanical’ fall! Describing a fall as a ‘mechanical fall’ is meaningless and almost always a failure of clinical assessment – unless you’ve had sev- eral days to fully assess the patient’s vision, hearing, gait, balance, strength, coordination, cognition, cardiovascular, neurological and general health, and found them all to be perfect. Just because the patient describes a trip or slip hazard, don’t assume this was the sole cause of the fall – they may be unaware of other risk factors aff ecting them, and could be unaware that they lost consciousness. In the unu- sual scenario of an otherwise completely healthy and independent patient who has slipped or tripped, describe the fall as a slip or trip, and be clear that they have no underlying gait or balance disorders.

Numerical risk assessment, or identifying and acting on risk factors? Whilst some hospitals use numerical risk assessment scores to try to identify the patients most likely to fall, many tools in use have never had their predictive values tested, and even those scores

49 C

hapter 15 S lips, trips and falls

that have been tested have poor predictive values; NICE Clini- cal Guideline 161 says risk scores should not be used. Given the frailty, age and illness of the majority of hospital patients, com- pounded by the eff ects of treatment and the unfamiliar environ- ment, it can be better to simply consider all patients at risk unless they are independently and confi dently mobile, and have no his- tory of previous falls.

Multifactorial assessment and intervention NICE Clinical Guideline 161 requires that all older people with recurrent falls, or who have fallen and have problems with gait and mobility, should be given an individualised multifactorial assessment and intervention – that is, a systematic process of identifying and, wherever possible, treating, modifying or better managing the older person’s individual risk factors for falling. In a hospital, this will be provided by the multi-disciplinary team against a background of the general diagnosis and management of underlying disease. In addi- tion to a falls history, an individualised multifactorial assessment and intervention must consider: • gait, balance, mobility and muscle strength; perceived functional ability; and fear of falling • visual impairment, especially recent changes in vision, visual fi eld loss in patients with current or previous stroke, cataracts (early cataract extraction is of benefi t in reducing falls risk), out- dated glasses (if the frames look old, the lenses may be too) and diffi culty managing bifocal or varifocal lenses • cognitive impairment (including, in the hospital environment, assessment for delirium) and neurological examination • urinary incontinence, frequency or urgency • environmental safety (in hospital, this would include leaving personal possessions, drinks, call bells and mobility aids in reach; avoiding clutter; and acting on slip or trip hazards) • cardiovascular examination, including lying and standing blood pressure to detect orthostatic hypotension • medication review (see the ‘Medication review’ section for more detail) (Figure 15.1)

Medication review Falls can be caused by almost any medication that acts on the brain or circulation, and reducing or discontinuing medication where the risks outweigh the benefi ts can help prevent falls. Key types of medication that increase falls risk include: • Psychotropic drugs, especially benzodiazepines (but avoid abrupt discontinuation in habituated patients), antidepressants, medica- tion used for psychosis and agitation, opiates, anti-epileptics, phe- nothiazines, sedating antihistamines and muscle relaxants • Medication causing orthostatic or general hypotension, brady- cardia, tachycardia or periods of asystole, including angiotensin- converting enzyme (ACE) inhibitors; alpha, beta and angiotensin II receptor blockers; thiazide and loop diuretics, antianginals and acetylcholinesterase inhibitors

In hospitals, medication-induced hypotension is a particu- lar risk if previous doses of anti-hypertensives are prescribed to patients whose acute illness has resulted in their blood pressure dropping, as is ‘compliance hypotension’ (patients whose pre- scribed doses of anti-hypertensives do not refl ect what they have actually been taking before admission).

Osteoporosis and fracture prevention Whenever assessing for falls risk, bone health must also be consid- ered. Th e most serious falls-related injuries are fractures, especially hip fractures. Patients who have sustained a prior fracture must have a bone health assessment and appropriate treatment in line with NICE Technology Appraisal Guidance 161.

After a fall in hospital A National Patient Safety Agency Rapid Response Report iden- tifi ed that around 20% of patients who received serious injuries in a hospital fall had ‘insult added to injury’ through failures and delays in diagnosing and treating injuries from their fall. Following is some guidance on what to do aft er a fall: • Don’t become complacent – 99% of patients you see aft er an inpatient fall will not have serious injuries, but 1% will • Stop, think and examine on the fl oor before allowing the patient to be moved – using a sling hoist on a fractured hip can be agonis- ing, and using it on a fractured spine can be catastrophic • Ensure any investigations or referrals for possible injury are treated with the urgency they deserve – the odds of recovering from serious injury are stacked against patients who were already acutely ill before they fell • Consider pain relief, but be aware of the risk of masking symptoms • Request neurological observations for any patient whose fall was unwitnessed – if no one saw the fall, you can’t be sure they did not injure their head (even a cognitively intact patient could have retrograde amnesia from striking their head) • Don’t discontinue neurological observations too soon – especially in coagulopathic or anti-coagulated patients, in whom cerebral bleeds can take 24 to 48 hours or more to become symptomatic • Always consider: ‘Have they fallen because they are ill?’ A fall in a hospital patient is oft en an ominous sign of deterioration in their condition, including potentially life-threatening infections, delirium or cardiovascular events • Take the opportunity to prevent further falls – revisit multifacto- rial assessment and intervention

Final words Getting old is inevitable, but falling in older age is not. Th e risk of falls and fractures in older people can be reduced, and it is the responsibility of us all.

P art 3 R

isks to patient care

16 Patient safety in paediatrics Figure 16.1 National Institute for Health and Care Excellence (NICE) feverish illness in children guideline

The traffic light system: Children with fever and any of the symptoms or signs in the red column should be recognised as being at high risk. Similarly, children with fever and any of the symptoms or signs in the amber column and none in the red column should be recognised as being at intermediate risk. Children with symptoms and signs in the green column and none in the amber or red column are at low risk.

Child younger than 3 months old

Assess: look for life-threatening, traffic light and specific diseases symptoms and signs (see tables 1 and 2)

Child 3 months of age or older

If all green features and no amber or red

If any amber features and no diagnosis reached

If any red features and no diagnosis reached

Perform urine test for urinary tract infection

Assess for symptoms and signs of pneumonia

Do not perform routine blood tests or chest X-ray

Perform (unless deemed unnecessary): • Urine test for urinary tract infection • Full blood count • Blood culture • C-reactive protein

Perform chest X-ray if fever higher than 30°C and white blood cell count greater than 20 x 103/litre

Consider lumbar puncture if child is younger than 1 year old

Perform: • Blood culture • Full blood count • Urine test for urinary tract infection • C-reactive protein

Consider the following, as guided by clinical assessment: • Lumbar puncture in children of all ages • Chest X-ray irrespective of white blood cell count and body temperature • Serum electrolytes • Blood gas

Observe and monitor: • Temperature • Heart rate • Respiratory rate

Perform: • Full blood count • C-reactive protein • Blood culture • Urine test for urinary tract infection • Chest X-ray if respiratory signs are present • Stool culture if diarrhoea is present

Admit, perform lumbar puncture and start parenteral antibiotics if the child is: • Younger than 1 month old • 1–3 months old appearing unwell • 1–3 months old and with a white blood cell count of less than 15 x 103/litre

Whenever possible perform lumbar puncture before the administration of antibiotics

If no diagnosis is reached, manage the child at home with appropriate care advice. Advise parents/carers when to seek further attention from the healthcare services

Consider admission according to clinical and social circumstances and treat

If the child does not need admission to hospital but no diagnosis has been reached, provide a safety net for the

parents/carers

Green – low risk Amber – intermediate risk Red – high risk

Colour

Activity

Respiratory

Hydration

Other

• Normal colour of skin lips and tongue

• Responds normally to social cues

• Content/smiles

• Stays awake or awakens quickly

• Strong normal cry/not crying

• Normal skin and eyes

• Moist mucous membranes

• None of the amber or red symptoms

or signs

• Pallor reported by parent/carer

• Not responding normally to social cues

• Wakes only when prolonged stimulation

• Decreased activity

• No smile

• Nasal flaring

• Tachypnoea:

– RR >50 breaths/minute age 6–12 months

– RR >40 breaths/minute age >12 months

• Oxygen saturation ≤ 95% in air

• Crackles

• Dry mucous membrane

• Poor feeding in infants

• CRT ≥3 seconds

• Reduced urine output

• Fewer for ≥5 days

• Swelling of a limb or joint

• Non-weight bearing/not using an extremity

• A new lump >2 cm

• Pale/mottled/ashen/blue

• No response to social cues

• Appears ill to a healthcare professional

• Unable to rouse or if roused does not stay awake

• Weak, high-pitched or continuous cry

• Grunting

• Tachypnoea:

– RR >60 breaths/minute

• Moderate or severe chest indrawing

• Reduced skin turgor

• Age 0–3 months, temperature ≥38°C

• Age 3–6 months, temperature ≥39°C

• Non-blanching rash

• Bulging fontanelle

• Neck stiffness

• Status epilepticus

• Focal neurological signs

• Focal seizures

• Bile-stained vomiting

CRT = Capillary refill time, RR = Respiratory rate

Source: Reproduced with permission of NICE.

51 C

hapter 16 P atient safety in paediatrics

Introduction Th e delivery of safe care to children is a major challenge for health- care providers. Over the past 30 years, there have been major developments in medical science which have resulted in decreased mortality and better outcomes in all areas of paediatrics and child health. Th e aim now is for children to receive appropriate evidence- based care the fi rst time, every time. Th is is the essence of provid- ing highly reliable care to children, whether in the developed world or in less developed countries. Th e potential for this to happen in the less developed countries is great. Healthcare problems may dif- fer, but in essence the principles are the same. Th is chapter outlines a few concepts essential to paediatric patient safety.

How safe is healthcare for children? Since 1990, the global under-5 mortality rate has dropped from 87 deaths per 1000 live births in 1990, to 51 per 1000 in 2011. However, the majority are due to conditions that could be prevented or treated with access to simple, aff ordable interventions. Forty-three per cent of child deaths under the age of 5 take place during the neonatal period, so safe childbirth and eff ective neonatal care comprise an important focus. Th e key is the ability to deliver the care required to all children, whether this is immunisation, oral rehydration or safe water.

In the United Kingdom, the Confi dential Enquiry into Maternal and Child Health (CEMACH) report Why Children Die found pre- ventable factors in 26% of reviewed cases of child deaths, predomi- nantly related to poor communication and non-recognition of the deteriorating child. A review by the former National Patient Safety Agency (NPSA) also identifi ed diffi culty in recognition as a key safety issue, as well as a high rate of reported medication errors, and the need for improved communication between healthcare professionals.

Why children are different Patient safety is a young discipline, although it is central to healthcare. Paediatric patient safety is diff erent. Th e expression ‘Children are not little adults’ is particularly true in medicine; children are diff erent to adults in many ways. Table 16.1 shows key factors that infl uence risk.

Table 16.1 Factors that infl uence risk in paediatric patient safety

Diff erence in children Impact on safety Children go through diff erent physiological stages.

Wrong equipment for diff erent ages, or diff erent medications

Drug dosing is weight dependent. Medication error Children suff er from diff erent diseases.

Diagnostic delays

Children rely on adults and oft en cannot speak up for themselves.

Poor communication as a key factor

Oft en, children’s healthcare is provided in adult-oriented facilities.

Lack of training and equipment, or mental health harm

Adult-trained health providers oft en provide care.

No knowledge of disease spectrum and changing needs, missed diagnoses and no detection of deterioration

Children are at particular risk from natural and economic stresses.

Less able to survive as margin of error is smaller than in adults

Defi nition of harm Th ere are technical defi nitions of harm, but oft en these mask the real issue of what happened to the child. Th e NHS Institute for Innovation and Improvement (NHS III) trigger tool prefers the

defi nition ‘Anything that one would not like to happen to oneself, one’s own child or a member of one’s family’. Using such a broad, personal defi nition helps us understand how pervasive harm can be. Th e key is that in patient safety, we want to reduce harm to the minimum, or eliminate it.

Variation in healthcare Variation is evident throughout healthcare. We are all aware of patient characteristics that may explain some variation and diff er- ences in clinician capability. However, unwarranted clinical vari- ation, defi ned by Wennberg as ‘care that is not consistent with a patient’s preference or related to [their] underlying illness’, and var- iation in the way the system works are avoidable. Hollnagel defi nes ‘patient safety’ in terms of an organisation’s ability to respond to changing conditions and carry out normal business without harm- ing patients. Consider if, in a busy clinical department, there is an increased number of admissions and a member of staff is off sick. If safe, the unit would be able to manage to the same level of reli- ability and quality despite the stress under which it works. We all work in stressful environments where the predictable is regarded as unpredictable, and therefore crisis management is common. In less developed countries, if variation in delivery of simple inter- ventions was decreased, millions of lives could be saved.

High reliability Th e concept of high reliability comes from highly complex indus- tries and environments such as aviation, nuclear power and space travel. It means that there is very little chance of things going wrong. Key features of high reliability described by Weick and Sutcliff e are: • Preoccupation with failure – always asking what can go wrong before it does • Reluctance to simplify interpretation – not simply blaming but trying to understand how the system and individual interact in harm • Sensitivity to operations – always paying attention to what front- line staff and consumers of healthcare think and do • Commitment to resilience – a learning environment continually adapting to change • Deference to expertise – all decisions take into account the expe- rience of patients and frontline staff who can oft en solve the prob- lems that arise if given the authority to do so. Th e frontline staff (e.g. doctors in training) oft en know best how to solve problems and deliver safe care – we just need to make it easy for them to do it

Human factors Th ese refer to the interaction of humans and the environment in complex situations. Th e design of healthcare has only begun to understand the importance of this. To deliver healthcare is extremely complex, involving a number of steps that need to align. Design of the environment needs to anticipate how humans will act in these situations and how to avoid potential pitfalls (e.g. oxygen and air outlets with the same fi tting). Simulation train- ing with a focus on leadership and team communication is also essential for healthcare professionals to understand human fac- tors and develop their skills. Th is can be simple and applicable in all settings.

Interventions to decrease harm 1. Decrease variation Deming says that in order for the system to work to its optimum, it must reduce variation in how it delivers its outputs. In healthcare,

P art 3 R

isks to patient care

this means standardise where possible, keep to policy to provide evidenced-based care and specialise only where necessary. National Institute for Health and Care Excellence (NICE) guideline for feverish illness in children: a traffi c light system Feverish illness in young children is very common and usually indicates an underlying infection. It is a cause of concern for par- ents and carers. It is the second most common reason for a child being admitted to hospital and yet despite advances in healthcare in the United Kingdom, infections remain the leading cause of death in children under the age of 5 years. Fever in children can be a diagnostic challenge for healthcare professionals. It can be diffi cult to distinguish between a self-limiting viral illness and a more serious bacterial infection such as pneumonia or meningitis, especially when there is no apparent cause of fever, despite careful assessment. Due to a perceived need to improve the recognition, assessment and immediate treatment of feverish illnesses in chil- dren, the NICE feverish illness guideline was designed in 2007. Th e evidence-based guideline addresses the detection of fever, outlin- ing suitable thermometers and the importance of taking parental reporting of fever seriously; management by remote assessment; and the traffi c light system to predict the risk of serious illness (see Figure 16.1). Th e management of children with fever should be directed by the level of risk, as outlined in the guideline.

Th e implementation of guidelines, however, is poor. Th e use of improvement methods such as the Model for Improvement (see Chapter 23) is essential to achieve near-perfect compliance.

2. Detect harm Traditional ways of detecting adverse events in paediatrics have relied on voluntary reporting. However, research has shown that only 10–20% of errors are ever reported and, of those, 90–95% cause no harm to patients. It is important to focus on harm, which targets the system, rather than on individual error. Paediatric trigger tool Structured case note reviews by medical and nursing teams on a reg- ular basis help identify unintended harm within a service. By under- standing the types of harm that may occur, the organisation can focus its improvement work on the areas of greatest impact and track improvements over time. Th e Model for Improvement and rapid test cycles were used by the NHS Institute for Innovation and Improve- ment to develop the UK paediatric trigger tool. Co-production was crucial for its design: decision makers, experts, service providers and users worked together to create a tool that worked for them all.

Th e tool measures adverse events in district general hospitals, acute teaching hospitals and specialist paediatric centres, with high sensitivity and specifi city, through rapid (maximum 20 minutes) structured review of randomly selected case notes. Triggers listed in the tool are identifi ed (e.g. INR >5), then a closer examination of the notes determines whether an adverse event has occurred (e.g. bleeding). If an adverse event has occurred and harm has resulted, then the category of harm is assigned. Figure 16.2 illustrates a page of the tool with examples of triggers and Figure 16.3 shows an early warning scores chart. Th e purpose of the review is to identify harm, not to discuss whether it was preventable. Trigger tools typi- cally detect harm rates in excess of 30%.

3. Improve communication In most clinical incidents, poor communication is an underlying problem. Handovers are an important part of clinical care and there- fore need to be structured, consistent and reliable. Th e use of the SBAR (situation, background, assessment and recommendation) tool has become a fundamental part of safe care. All communication

should follow an SBAR-type structure – both written and verbal (see Figure 16.2). Reading this back is an essential component to ensure that the information transmitted is accurate and understood.

4. Improve access to healthcare Reducing newborn deaths globally: World Health Organization (WHO) safety tool to improve access to healthcare Th e Millennium Development Goals (MDGs) were devised in the year 2000 to provide targets for improving social indicators around the world. MDGs 4 and 5 provide global targets for improving the health and survival of mothers and children. In September 2010, the UN Secretary General launched a re-energised Global Strategy for Women’s and Children’s Health due to concerns around reach- ing these goals by 2015.

WHO estimates that nearly four million newborns die each year globally within the fi rst month of life; three-quarters of these deaths occur in the fi rst week. Delayed recognition of early warn- ing signs and delayed access to healthcare are the most common reasons for the majority of these deaths, which mostly occur in low-resource countries.

Th e Patients for Patient Safety team of WHO Patient Safety has created a patient-held safety tool that will help increase safety for both mothers and their newborn babies during the high-risk period: the fi rst 7 days of life. Tool development was via a ‘by patients, for patients’ approach, ensuring it is easy to use by all mothers globally in settings of varying literacy. Based on a check- list concept, it contains safety checks for common danger signs for mothers and their babies to lead to more timely and appropriate accessing of healthcare.

5. Responding to deterioration Early warning scores Early warning scores are a means of identifying deterioration to prevent intensive care admission or cardiorespiratory arrest. Th is is especially important in children since it is well recognised that children who have died or required intensive care have shown signs of physiological or behavioural disturbance in the period prior to collapse. One of the CEMACH recommendations for paediatric care in the hospital was a ‘standardised and rational monitoring system with embedded early identifi cation systems for children developing critical illness’. An example of an early warning score is shown in Figure 16.2. Th ere are important considerations to intro- ducing early warning scores: • Staff engagement is crucial. Recently, charts have incorporated nursing or parental concern as a score, recognising this importance • Th e accurate completion of observations and calculation of scores are crucial, and so appropriate training of all staff is required • Th e validity and impact of early warning scores are diffi cult to show, due to confounding factors such as rising standards of care and medical emergency response teams. Tibballs reviewed sys- tems to prevent in-hospital arrest and found that the key feature is empowerment of any staff , however junior, and parents to sum- mon help without deferring to senior colleagues or medical staff . Early warning scores may facilitate escalation of concerns using SBAR and therefore response to the deteriorating child

6. Decreasing medication harm Th e prescribing of medications is not always performed with a high degree of reliability. A key factor to improving this is service redesign. Zero-tolerance prescribing At Great Ormond Street Hospital PICU, a programme of zero tol- erance to errors has been instituted within a blame-free learning

53 C

hapter 16 P atient safety in paediatrics

environment. Prescriptions can be written in designated no- interruption areas which are fully equipped with all the material required to prescribe. Immediate feedback on any errors found is given to ensure that there is real-time improvement. Th is has resulted in a substantial decrease in prescription errors. Paediatric Emergency Drugs calculator app Formulas for calculation of emergency drugs and infusions required for the resuscitation and stabilisation of the critically ill child are based on weight. Th is can lead to medication errors or delay in treatment while waiting for doses to be calculated.

Th e Paediatric Emergency Drugs app combines clinical expe- rience from the Paediatric Intensive Care Unit of the Evelina Children’s Hospital, Guy’s and St Th omas’ Trust, London, who have developed formulas and guidance for paediatric emergen- cies, with handheld technology developed by UBQO Limited. It is a simple and easy-to-use off -line application for both doctors and nurses on intensive care units and at hospitals stabilising

and referring children. Aft er entering the child’s age, and weight if known, results of drug doses and over 80 infusions are quickly presented in a logical way. Physiological parameters based on the child’s age are also presented to help recognise deterioration, with less reliance on memory of the normal age values of heart rate, blood pressure and respiratory rate.

Nolan describes system changes to reduce errors, acknowledg- ing that many errors can be attributed to human cognition. He describes reducing complexity (e.g., availability of only paedi- atric concentrations of drugs), optimising information process- ing using automation and constraints (e.g. removing concen- trated potassium solutions from children’s wards) and mitigating unwanted eff ects of change (e.g. testing new equipment on a small scale with minimum risk). Drug calculators are an example of a system change to prevent human error. Dosing errors are consist- ently found to be the most common type of medication incident in children.

Figure 16.2 Paediatric trigger tool

PG1 EWS or baseline obs missing or incomplete OR score/observation requiring response

Tissue damage or pressure ulcer

PG2

PG3

PG4

Readmission to hospital within 30 days

Unplanned admissions

Full descriptions Trigger Adverse

event Severity of

adverse event Comment on this trigger

N/A E F G H I

No Yes No Yes

Patient age

Date of discharge

Length of stay

years, months

days Institute of Innovation

and Improvement

= Temporary harm to the patient and required intervention

= Temporary harm to the patient and required initial or prolonged hospitalisation

= Permanent patient harm

= Intervention required to sustain life

= Patient death

PAEDIATRIC TRIGGER TOOL www.institute.nhs.uk/safercare/portal

NHSTrigger Tool

Source: http://www.institute.nhs.uk/images//documents/SaferCare/Paediatrics/Paediatric%20Trigger%20Tool%20Form.pdf

Figure 16.3 Early warning scores

Ask receiver to repeat key information to ensure understanding

Early warning scores chart for 0–11 months

Situation: • I am (name), a nurse on ward (X) • I am calling about (child X) • I am calling because I am concerned that ... (e.g. BP is low/high, pulse is XXX, temperature is XX, early warning score is XX) Background: • Child (X) was admitted on (XX date) with ... (e.g. respiratory infection) • They have had (X operation/procedure/investigation) • Child (X)’s condition has changes in the last (XX mins) • Their last set of obs were (XXX) • The child’s normal condition is ... (e.g. alert/drowsy/confused, pain free) Assessment: • I think the problem is (XXX) and I have ... (e.g. given O2/analgesia, stopped the infusion) OR • I am not sure what the problem is but child (X) is deteriorating OR • I don’t know what’s wrong but I am really worried Recommendation: I need you to ... • Come to see child in the next (XX mins) AND • Is there anything I need to do in the meantime? (e.g. stop the fluid/repeat the obs)

0 1 Continue monitoring

Nurse in charge and Doctor

MUST review

Nurse in charge and Doctor MUST review and inform Consultant

Nurse in charge and Consultant MUST review

Source: http://www.institute.nhs.uk/images/Paediatric/PEWS4/PEWS_0-11months_v6.pdf

5 6

Remember: If you feel you need more help at any time, call for help – regardless of PEW Score

S B A R

http://www.institute.nhs.uk/safer_care/paediatric_safer_care/pews.html)

P art 3 R

isks to patient care

17 Technology in healthcare and e-iatrogenesis Figure 17.1 Functions of information technology (IT) systems in healthcare

Informing and supporting decisions

Supporting patients/citizens

On-line information on health/lifestyle or

illness/treatments

Tools for balancing risks/benefits/preferences

to aid choices

Targeted educational interventions

Personal health records and self-management

aids

On-line support networks

Supporting professionals Case-specific

diagnostic or treatment advice based on patient

data and expert knowledge/evidence

Automated prompts and reminders for guidleline compliant prescribing, screening or reporting

+ safety alerts Electronic guidelines,

research reports and CME tools

Storing and managing data Delivering expertise and

care at a distance

Data

Patient-specific records, supporting care of individuals

Population-based data, aiding research, policy

and planning. Administrative data, aiding organisational

and business processes. Integrated records, supporting mutiple

stakeholders. Medical images

Records systems

Clinical (e.g. for capturing, displaying, sharing,

linking or exchanging patient-specific data;

or populating decision support)

Administrative (e.g. for audit, purchasing,

billing, tracking service utilisation etc.)

Expertise and knowledge

Diagnostic or treatment advice

from subject experts (e.g. telepathology)

Medical conferencing, clinical email

Mobile access to records;

evidence; CDs

Care

Patient-provider email; internet consultations

Remote interventions (e.g. telepsychiatry,

telesurgery)

Home and ambulatory disease monitoring

and self-management

support

Figure 17.2 Assumptions underlying the introduction of IT in healthcare

Quality

Challenges surrounding the implementation of healthcare IT

• Under-developed technology

• Unintended consequences

• Changes to established work practices

• Technology taking time away from patient care and increasing workload

Efficiency

Improvements in ...

55 C

hapter 17 Technology in healthcare and e-iatrogenesis Introduction Healthcare systems internationally are increasingly struggling to cope with demand due to increasing population sizes and demo- graphic changes that are resulting in greater numbers of older peo- ple, who are oft en living with more than one long-term condition. Th ere is an ever-increasing amount of medical knowledge and array of new treatment options. Healthcare providers and policy makers are therefore seeking out new and innovative ways to cope with these challenges. One key development in this respect is tech- nological innovation, which has great potential to improve the quality, safety and effi ciency of care (Figure 17.1 and Figure 17.2). Technology has the potential to reduce the risks associated with care provision by improving communication, promoting the accessibility of knowledge and assisting with decision making. However, technological tools may also introduce new risks. For example, they may prove disruptive to established working prac- tices of users, and their introduction may inadvertently introduce new types of errors associated with their use.

Epidemiology of IT-related errors • Healthcare technology is used by all healthcare professionals in developed countries, and is increasingly being used in developing countries • Technical complications account for around 13% of adverse events in healthcare • In the United Kingdom, around 400 patients a year die from adverse events related to medical devices • Technology-related errors are expected to increase in line with new technological developments • Th ere is currently a lack of understanding around the measure- ment, analysis and categorisation of IT-related errors in healthcare settings

The nature of technology-related errors in healthcare settings Errors and adverse events resulting from the use of technology are referred to as e-iatrogenesis. Th ey most commonly occur in the fol- lowing scenarios, and can involve both hardware and soft ware: • Th e technology is needed but is unavailable for use: Unavailability can result in users being unable to access and/or transmit data. For example, a review of medical notes and past medications may not be possible if the computer network is down or portable devices are not suffi ciently charged. As a result, the user may not be alerted to potential drug allergies and contraindications, increasing the risk of inappropriate prescribing • Th e technology malfunctions during use: Th is occurs if the tech- nology is used as intended but the system does not perform the desired function correctly. For instance, a computer system may display incorrect information due to faulty algorithms, such as an X-ray result of another patient, potentially increasing risks of incorrect diagnoses • Th e technology is used in ways other than intended: Th is scen- ario tends to occur if the use of the technology is viewed as too cumbersome and time-consuming by users in a busy healthcare environment. As a result, users take the quickest route to enter

or access data in the system, which tends to compromise system reliability. For example, a user may fi nd it easier to input data in free text boxes, but this can result in the inability of a system to pick up and alert users to potential adverse events such as contraindications • Th e technology interacts with other technology in unintended ways: Th is can occur if essential information is not correctly transferred from one computer system to another. For example, a pharmacy system may fail to transfer important medication review-related information to the clinical ward system, resulting in the treating clinician not noticing a potential medication overdose

As can be seen, e-iatrogenesis can result not only from the design of the technology but also from the way it is implemented (or introduced) and used in individuals and organisations.

Design and implementation Th ere are several ways to mitigate against e-iatrogenesis. Th ese relate to both improved technological design and implementation- related activities. We will discuss some examples relating to both aspects in this section.

Design Technological systems should be designed with safety in mind and, in doing so, help users ‘to do the right thing’. Such design features can help to address surrounding latent error-producing conditions as well as human cognitive shortcomings that may lead to adverse events. Th ey are, however, not fool-proof and may bring their own inherent risks that users need to be aware of.

In relation to latent conditions, an example is bar-code verifi ca- tion technology which can help to reduce medicines administra- tion errors and resulting adverse events by identifying the right patient and the right medication. Th is is accomplished by scan- ning a patient’s wristband against the medication that is about to be administered, thereby minimising the risk of confusing patients and/or doses. However, although eff ective for some types of medi- cations, such systems do have limitations. Th ey can, for example, result in delays in administration if soft ware access for immediate administration is required and the soft ware is slow to load.

In relation to human cognition, an example is prescribing systems, which are commonly used to hold prescribing-related information and alert prescribers to potentially inappropriate doses, contraindications and allergies. However, the way infor- mation is displayed on the screen can signifi cantly impact user behaviour. For example, the use of a large number of pop-up alerts may become tiring for users, resulting in them ignoring potentially important messages. Th is is commonly known as alert fatigue.

Usability analysis can help to minimise the potential for design- related errors. Th ese can take a variety of forms, but the most commonly applied technique is heuristic evaluation. Th is involves assessment of a technological system based on an established set of usability indicators, including the following system features: the use of simple dialogue and language that are understandable for users, the ability to minimise users’ memory load, consistency and provision of feedback to users, the ability to provide shortcuts and eff ective error messages and the capability to prevent errors and facilitate documentation.

P art 3 R

isks to patient care

Figure 17.3 The need for human-centred design

System satisfies specific user and organisational

requirements

Evaluate designs against requirements

Understand and specify the context of use

Identify need for human-centred design

Specify the user and organisational requirements

Produce design solutions

Source: © Claudia Pagliari

Implementation In addition to design-related features, e-iatrogenesis can also result from the way that technology is introduced in organisations. For example, if users are not trained appropriately in how to use a technology, or if a system does not fi t in with individual work practices, unintended patterns of usage may result. Th ese are oft en conceptualised in the form of workarounds, or strategies employed by users to compensate for perceived inadequacies of the techno- logical system. If, for instance, a system requires a large number of clicks for users to navigate it, they may choose to delay data entry activities until they fi nd time later in the day. Th is in turn may result in records not being up-to-date and other healthcare profes- sionals not being able to access important information. Potential mitigating strategies can include extensive work process mapping, involving the assessment of existing individual workfl ows with the technology, and re-designing processes involving the technology around these. Th is should involve extensive consultation of users and subsequent changes in technological design according to needs (Figure 17.3).

Initiatives to reduce technology-related errors Th ere are several initiatives that have attempted to reduce errors associated with technological systems in healthcare. Th ese are mainly based on the design or implementation of best-practice guidelines and standards. We will outline a few examples in this section, but caution that the interrelated or iterative nature of design and implementation activities means that none of these should be viewed in isolation.

Safe implementation Toolkits for local organisations implementing healthcare technolo- gies aim to facilitate the process of change locally. Th ese tend to guide implementers through the whole life cycle of introducing new technol- ogies, from the conception of the project, planning, implementation and through to ongoing maintenance. One example is Stratis Health’s Health Information Technology Toolkit for Critical Access and Small Hospitals (http://www.stratishealth.org/expertise/healthit/hospitals/ htoolkit.html). Although based on the US context, many aspects of this work are transferable to organisations in other countries. Imple- menting organisations are provided with a range of resources that are adaptable to the particularities of individual contexts, including train- ing resources and advice on work process mapping.

National guidelines Th ere are also increasingly design-related initiatives that aim to reduce errors associated with healthcare technologies; for exam- ple, in 2010, the United Kingdom’s NPSA published guidelines for the safe on-screen display of medication information. Th ese include advice for designers on displaying text and symbols, drug names, numbers and units of measurement and other information. A good example is the tendency of busy users to misread numbers due to trailing zeroes. So, if a system displays ‘DOSE 5.0 mg’, the user may in fact read ‘50 mg’, resulting in an overdose. Th e NPSA recommends leaving trailing zeroes out, as a simple but eff ective mitigation against potential risks of overdoses (i.e. ‘DOSE 5 mg’). Th is non-compulsory guidance is intended for the use of system developers (who may want to design their systems accordingly) and implementers (who may want to purchase systems that fulfi l certain design-related safety standards).

57 C

hapter 17 Technology in healthcare and e-iatrogenesis A similar initiative is the NPSA’s guide to the design of elec-

tronic infusion devices. Th is includes both hardware and soft ware specifi cations and recommendations on how to address key safety concerns associated with these. For instance, some devices do not alert the user if the syringe and plunger are not adequately secured in the device, resulting in the potential for disengaging.

Safe implementation and design Th ere are also global initiatives targeting both implementation- and design-related sources of error. For instance, the World Health Organization’s Department of Essential Health Technologies (part of the Global Initiative on Health Technologies) oversees inter- national eff orts to develop policies and guidelines for the use and implementation of healthcare technologies based on available empirical evidence. In doing so, the initiative has four sub-streams

dealing with specifi c technological devices: Blood Transfusion Safety, Clinical Procedures, Diagnostic Imaging and Medical Devices, and Diagnostics and Laboratory Technology. Each mem- ber state is off ered assistance in strengthening existing technolo- gies within their economic capability.

Conclusion Technological systems have made signifi cant headway in ensuring that care is safer and more effi cient. However, their use can intro- duce new sources of error and potentially avoidable harm. Several design- and implementation-related initiatives can help mitigate the risk of error, but technology should be viewed as a tool as opposed to a fool-proof mechanism to eliminate human cognitive and latent errors.

P art 3 R

isks to patient care

18 Nosocomial infections Figure 18.1 Your 5 moments for hand hygiene Figure 18.3 12 steps to prevent antimicrobial resistance in

hospitalised adults

Figure 18.2 Surgical safety checklist

3 4

5 Before

patient contact

Before aseptic task

After patient contact

After body fluid exposure risk

After contact with patient surroundings

Source: Reproduced with permission of WHO. 1 Vaccinate

2 Get the catheters out

3 Target the pathogen

4 Access the experts

5 Practice antimicrobial control

6 Use local data

7 Treat infection, not contamination

8 Treat infection, not colonisation

9 Know when to say “no” to Vanco

10 Stop treatment when cured

11 Isolate the pathogen

12 Contain your contagion Prevent transmission

Use antimicrobials wisely

Diagnose and treat effectively

Prevent infections

Source: CDC

This checklist is not intended to be comprehensive. Additions and modifications to fit local practice are encouraged

Nurse verbally confirms:

The name of the procedure

Completion of instrument, sponge and needlecounts

Specimen labelling (read specimen labels aloud, including patient name)

Whether there are any equipment problems to be addressed

To Surgeon, Anaesthetist and Nurse:

What are the key concerns for recovery and management of this patient?

Before induction of anaesthesia Before skin incision Before patient leaves operating room

(with at least nurse and anaesthetist) (with nurse, anaesthetist and surgeon) (with nurse, anaesthetist and surgeon)

Has the patient confirmed his/her identity, site, procedure, and consent?

Yes

Is the site marked?

Yes

Not applicable

Is the anaesthesia machine and medication check complete?

Yes

Is the pulse oximeter on the patient and functioning?

Yes

Does the patient have a: Known allergy?

Yes

Difficult airway or aspiration risk?

Yes, and equipment/assistance available

Risk of >500ml blood loss (7ml/kg in children)?

Yes, and two IVs/central access and fluids planned

Confirm all team members have introduced themselves by name and role

Confirm the patient’s name, procedure, and where the incision will be made

Has antibiotic prophylaxis been given within the last 60 minutes?

Yes

Not applicable

Anticipated Critical Events

To Surgeon:

What are the critical or non-routine steps?

How long will the case take?

What is the anticipated blood loss?

To Anaesthetist:

Are there any patient-specific concerns?

To Nursing Team:

Has sterility (including indicator results) been confirmed?

Are there equipment issues or any concerns?

Is essential imaging displayed?

Yes

Not applicable

Source: Reproduced with permission of WHO.

Surgical Safety Checklist

59 C

hapter 18 N osocom

ial infections Introduction Nosocomial infection can be defi ned as an infection acquired in hospital by a patient who was admitted for a reason other than that infection. Typically the criteria for nosocomial infections are those occurring within 48 hours of hospital admission, 3 days post-discharge or 30 days of a surgical operation (e.g. an infection) following the insertion of an intravascular catheter (e.g. a central line). Whilst a central line is used to provide life-saving medica- tions and eff ective monitoring, prolonged use almost inevitably leads to infection which can have signifi cant consequences for the patient in terms of increased length of stay, treatment and, in a minority of cases, death.

Burden of nosocomial infections Th e prevalence of healthcare-associated infections (HCAIs) varies between 3.5% and 12% in developed countries, and between 5.7% and 19.1% in developing countries.

In the United States, two million patients will acquire a noso- comial infection during their hospital stay, of which approximately 90,000 will die. Th e estimated economic burden is estimated between $28 and $45 billion. Th e vast majority of nosocomial infections, however, are preventable, and a large corpus of work has been undertaken globally to reduce them; some initiatives have demonstrated signifi cant outcomes in reducing their prevalence.

Nosocomial infections are associated with increased morbid- ity and mortality, as well as increased admissions to, and length of stay in, intensive care, and the accompanying extra care costs from additional testing, drugs and procedures.

Case study A 65-year-old man is admitted to hospital following a myocardial infarction, and required emergency cardiac

surgery (a coronary artery bypass graft procedure). He made a good recovery over the fi rst 48 hours on a surgical ward, and he is on target to be discharged on day 5. On the third day, he developed a fever with rigours, and the on-call doctor noticed an erythematous sternal wound with a serous discharge, and prescribed broad-spectrum antibiotics and intravenous fl uids. Th e patient deteriorated further and was admitted to intensive care unit (ICU), where a Foundation Year 2 doctor inserted a central line and arterial line. He is diagnosed with sepsis secondary to a sternal wound infection.

Th e patient received additional antibiotics to treat a methicillin- resistant Staphylococcus aureus (MRSA) infection over the next 4 days. By day 7 of his ICU stay, he is scheduled to return to the ward, when the ICU nurse records a low-grade fever. Th e ICU doctor exam- ines the patient and notes erythema and pus at his 7-day-old cen- tral line insertion site. Th e doctor removes the central line and sends the tip to microbiology; Staphylococcus epidermidis is grown. Th e patient endures 2 further days of monitoring and antibiotics in ICU.

In this case, the patient developed a surgical site infection (with MRSA) and a catheter-related infection. His scheduled discharge date was originally 5 days post-op, yet he stayed for a total of 14 days. His ICU admission was costly in terms of extra medications, multiple specimens and tests as well as bed blocking other patients from being admitted to ICU and being admitted for elective surgery. Th e patient became depressed following his pro- longed ICU stay, and in the weeks following his hospital stay he was diagnosed with depression requiring antidepressant therapy.

Reducing HCAIs is not an easy task; its multifactorial nature makes it a diffi cult issue to contend with.

Initiatives to reduce nosocomial infections In the ‘Keystone ICU’ project, funded by the Agency for Health- care Research and Quality (AHRQ), 103 ICUs in Michigan, USA, participated in a state-wide safety initiative, which included insti- tuting evidence-based preventive strategies for reducing CRBSIs (catheter-related bloodstream infections). Th e project focused on changing provider behaviour through addressing safety culture, incorporating a centralised education programme for team lead- ers at each institution and closely collaborating with infection control staff . Th e intervention almost eliminated CRBSIs in most ICUs over an 18-month follow-up period, and 1500 lives were saved. Following the Keystone ICU project in Michigan, ‘Matching Michigan’ was a national initiative seeking to emulate the success and involved over 97% of acute NHS trusts in England. Th e results were promising, and a 60% reduction in the number of CRBSIs was reported. However, on closer analysis of data, it was diffi cult to determine whether the reduction in CRBSIs resulted from the Matching Michigan project or from a coinciding nationwide drive to reduce nosocomial infections, since many trusts were already implementing part of the 5-point strategy employed in the Michi- gan intervention. Th ere was also a decrease in other infections, which were not related to ICUs or CRBSIs.

Th e World Health Organization (WHO) has set up an inter- national campaign called ‘Clean Care Is Safer Care’ which aims to promote the essential role of infection control in patient safety. ‘SAVE LIVES: Clean Your Hands’ was a campaign within this to promote good hand hygiene practices of healthcare workers (see Figure 18.1 for ways to help reduce the spread of potentially life- threatening infections in healthcare facilities). Th e WHO Surgi- cal Safety Checklist also has elements in it (circled in Figure 18.2) which show the steps that need to occur to prevent nosocomial infections. Th e US Centers for Disease Control and Prevention (CDC) also has a 12-point checklist to prevent antimicrobial resistance (Figure 18.3).

Th ere is a lot of important work being carried out to try to reduce nosocomial infections, but clearly more needs to be done. Under- standing what nosocomial infections are, how they are acquired, what the repercussions and burden of them are and how they can be tackled are a critical part of any healthcare professional’s clinical knowledge.

P art 3 R

isks to patient care

19 Mental health errors Figure 19.1 Patient safety incidents reported between September 2010 and October 2010 based on care setting

Figure 19.2 Types of incidents reported from mental health and learning disability settings within one calendar year (September 2010–October 2011)

Acute/general hospital

Mental health service

Community nursing, medical and therapy service (including community hospital)

Learning disabilities service

Community pharmacy

General practice

Ambulance service

Dental or optician service

Type of incident (with illustrative example) Number of incidents

51,612

41,685

40,235

17,937

16,520

13,704

3381

2386

2101

2242

1651

1252

504

357

259

Patient accident – patient found on floor having fallen from bed

Disruptive, aggressive behaviour – patient verbally abusive towards staff

Self-harming behaviour – patient found to have inflicted a laceration to forearm using a broken CD

Access, admission, transfer or discharge (including missing patient) – delay in transferring patient to acute hospital for investigation of physical health problem

Medication – wrong medication administered

Other

Infrastructure (including staffing, facilities or environment) – not enough trained staff on ward to restrain a patient safely when required

Documentation (including records and identification) – wrong patient details on medication chart

Treatment and procedure – no physical health examination completed on admission

Patient abuse (by staff and third party) – allegation of physical abuse by staff member towards patient

Consent, communication and confidentiality – medication given without legal authority for detained patient

Implementation of care and ongoing monitoring and review – patient not observed according to the level of observations agreed by the clinical team

Infection control – outbreak of gastroenteritis on ward

Clinical assessment (including diagnosis, scans, tests and assessments) – patient not assessed prior to self-discharge

Medical device/equipment – resuscitation equipment not readily available

61 C

hapter 19 M ental health errors

Patient safety and risk in mental health Safety is at the centre of all good healthcare service. It is particu- larly important in mental health, but is also challenging. Patient autonomy has to be considered alongside public safety. A good therapeutic relationship includes both empathic support and objective assessment of risk. Eff ective care includes an awareness of a person’s overall needs as well as an awareness of the degree of risk that they may present to themselves or others. Risk management is a core component of mental healthcare. Assessing and managing risk are part of the daily work of mental health professionals.

Whereas signifi cant strides have been made in understand- ing the burden of error and harm in hospital specialties, limited work has been undertaken in mental health patient safety. Qual- ity improvement projects focused on safety are disproportionately focused on acute settings. Limited resources are dedicated to better understanding and working on safety issues within mental health or learning disability settings.

Patient safety incidents in mental health Th e National Reporting and Learning System in England and Wales was introduced in Chapter 6; it is a database of patient safety incident reports. Figure 19.1 shows the proportion of all reported safety incidents that come from mental health or learning disabili- ties services.

In 2010, there were 193,165 incidents reported in mental health and learning disabilities services – with 165,988 (86%) of these from mental health and 27,177 (14%) from learning disabilities services.

Figure 19.2 shows a breakdown of incidents by severity of harm and type of incident.

Th e fi ve most common types of reported patient safety inci- dents from mental health and learning disability settings: • Slips, trips and falls • Disruptive and/or aggressive behaviour • Self-harm – including incidents of attempted suicide, completed suicide and non-suicidal self-harm • Access, admission, transfer and discharge – including missing and absconding patients • Medication – including dosage errors, wrong medication admin- istration, allergy, adverse reactions and administration errors

Medication errors Between October 2007 and September 2008, the NPSA received 7419 reports of medication errors involving mental health patients, representing 9% of all reported medication errors. Of these, 96% resulted in no harm, but there were 100 cases of death or severe harm. Around 92% of patients in contact with mental health ser- vices are prescribed medication, and it is highly likely that many incidents involving medication are going unreported. People with mental health problems may be particularly susceptible to medica- tion errors due to various factors, including cognitive impairment, lack of insight into the nature of their mental disorder and need for treatment, or the complexity of mental health services and the number of diff erent interfaces between primary care, secondary care, social care, community and inpatient teams.

In mental health and learning disabilities settings, medication errors with particular risks include: • Omission of medication – for example, omission of methadone, anticonvulsant medication or clozapine can result in serious and rapid worsening of mental health conditions • Wrong dosing – drugs such as lithium have a narrow therapeutic range, so a wrong dose could lead to serious medical complications and potentially death • Failure of monitoring – drugs such as lithium or clozapine are subject to close monitoring. Th is is to prevent lithium toxicity or agranulocytosis for clozapine. A lithium-monitoring pack, similar to that used for oral anticoagulation therapy, has now been intro- duced which includes a patient-friendly information booklet, an alert card that patients should carry with them at all times and a record book to track critical information such as blood lithium lev- els and specifi c blood tests

Suicide and homicide Suicide and homicide receive the most attention in mental health safety research, with comprehensive data sets and analysis con- ducted by the National Confi dential Inquiry into Suicide and Homicide by People with Mental Illness.

Suicide Approximately a quarter of all suicides are in individuals who have been in contact with mental health services during the 12 months prior to death (termed ‘patient suicides’). Th e most common methods of suicide are hanging, strangulation and self-poisoning (overdose). Th e National Confi dential Inquiry has been collecting and studying data on patient suicides in the United Kingdom since 1997. Analysis of this large national data set has allowed the imple- mentation of national suicide prevention strategies, with the fi rst of these dating back to 2001 (‘12 Points to a Safer Service’). Recom- mendations included: • Th e removal of ligature points on wards • Follow-up within 7 days of discharge for all patient with severe mental illness or a history of self-harm within the last 3 months • Patients with a history of self-harm to receive no more than 2 weeks’ medication at a time • Development of assertive outreach teams to prevent loss of con- tact with vulnerable and high-risk individuals • 24-hour crisis teams in the community

Th e development of an evidence-based suicide prevention strategy with clear clinical recommendations and prominence in national policy has led to reductions in the numbers of patient sui- cides. From 2000, there was a 62% fall in the number of inpatients dying by suicide, and a 54% fall in the number dying by hanging. In addition to national strategies, there have been tools developed to support local services in assuring best practice on suicide pre- vention. Th ese include the Suicide Prevention Toolkits for mental health services, primary care, urgent settings and ambulance ser- vices, which provide audit tools and dashboards to monitor per- formance on areas that are core to reducing the risk of suicide. In 2012, a new cross-government suicide prevention strategy for Eng- land broadened the focus from patient suicides to reducing the sui- cide rate in the general population, with the impetus on reducing

P art 3 R

isks to patient care

Figure 19.3 The seven steps to improving patient safety in mental health

Step 1 Build a safety culture

• Create a culture that is open and fair

• Staff should have a constant and active awareness of the potential for things to go wrong

• Staff and the organisation should be able to acknowledge mistakes, learn from them and take action to put things right

• Information should be shared freely and openly with service users when things go wrong

Step 2 Lead and support your staff

• Strong leadership with clear policies in relation to safety and a willingness to implement best practice at service level

• Leaders should be visible and active in leading patient safety initiatives

Step 3 Integrate your risk management activity

• Patient safety is a key component of risk management, and should be integrated with staff safety, complaints management, litigation and claims handling, and financial and environmental risk

• The risk management system should be supported by a risk management strategy (involving a consistent approach to training, management, analysis and investigation of all risks), a programme of proactive risk assessments and the compilation of an organisation-wide risk register

Step 4 Strengthen reporting in mental healthcare

• A high level of reporting within an organisation indicates a better safety culture: the more aware staff are of safety problems, the more likely they are to report

• Key to improving reporting is ensuring that learning from the reports received are widely disseminated and then acted upon

Step 5 Involve and communicate with service users and the public

• Involving service users and the public in developing safer services at a strategic level

• Involving service users in their own care and treatment

• Encouraging an open, two-way dialogue between health professionals and service users when things go wrong

Step 6 Learn and share safety lessons

• Look at the underlying causes of patient safety incidents and learn how to prevent them from happening again

• Use Root Cause Analysis, and consider aggregating investigations to improve safety through addressing common issues

Step 7 Implement solutions to prevent harm

• Solutions need to be simple to implement and low cost

• Solutions range from designing out the potential for harm, designing systems which make it easy for people to do the right thing, to raising awareness and understanding

63 C

hapter 19 M ental health errors

the risk of suicide in key high-risk groups, tailoring approaches to improve mental health in specifi c groups, reducing access to the means of suicide and providing better information and support to those bereaved or aff ected by suicide.

Homicide When a homicide occurs involving a patient with mental ill- ness, there is signifi cant media and political interest. Mandatory independent homicide inquiries were introduced in England in 1994, but there has been much criticism of the value of these investigations in improving safety in mental healthcare. Th e com- mon preconception, oft en fostered by the media, is that people with a mental health problem are dangerous. It is tempting to think that fi nding ways of treating the mental illness will elim- inate risk and prevent the homicides from occurring. However, it is estimated that 21% of the homicides committed by people with mental illness were preventable by mental health services. Recent evidence suggests that although schizophrenia and other psycho- ses are associated with violence and violent off ending, particularly homicide, most of this excess risk appears to be mediated by sub- stance abuse comorbidity.

Th e number of homicides committed by people with a men- tal health problem has increased from 54 homicides in 1997 to 77 in 2005. Th e proportion of these committed by individuals with a psychotic disorder at the time of the off ence rose from 41% in 1997 to 62% in 2005. Of all those convicted of a homicide between 1997 and 2005, approximately 10% had been in contact with men- tal health services in the previous year. Recent evidence suggests that this upward trend in patient homicides has now reversed, fol- lowing a peak in 2006.

Independent inquiries aft er homicides have found a number of recurring themes, which should therefore be priorities for safety improvement work within mental health services. Th ese include inadequate risk assessment and management, poor communication between professional agencies, inadequate application of the Care Programme Approach, insuffi cient response to the patient’s non- engagement, failure to use the Mental Health Act appropriately, failing to listen to carers and non-compliance with medication.

Physical health monitoring Physical health and mental health are inextricably linked. Mental illness is associated with increased risk of physical illness, arising in part from a less healthy lifestyle and more frequent health-risk behaviours. Conversely, physical illness increases the risk of mental illness. For example, depression is associated with 67% increased mortality from cardiovascular disease, 50% increased mortality from cancer, twofold increased mortality from respiratory disease and threefold increased mortality from metabolic disease. Rates of depression are double in those with diabetes, hypertension, coronary artery disease and heart failure, and triple in those with end-stage renal failure, chronic obstructive pulmonary disease and cerebrovascular disease. People with schizophrenia and bipolar disorder die on average 25 years earlier than the general popula- tion, largely because of physical health problems. Schizophrenia is associated with increased mortality from cardiovascular disease (twofold), respiratory disease (threefold) and infectious disease (fourfold). Th erefore, as well as enhancing public mental health in order to improve population health and well-being, psychiatric services need to pay close attention to the physical health monitor- ing of patients.

Systems approach to improving safety Tackling patient safety issues such as those detailed in this chap- ter requires the adoption of a systematic approach to analysing, assessing and mitigating patient safety risks in order to achieve reliable and safe care for patients. Many innovative healthcare organisations have made important breakthroughs in the design and performance of safer systems by focusing on lessons learned in other high-reliability industries with a long history of using quality improvement methodology. Th is has led to the introduction of ini- tiatives such as standardising approaches, process re-engineering, decreasing complexity, incorporating human factors design and focusing on developing a safety culture. Th e ‘seven steps’ frame- work (Figure 19.3) aims to help healthcare organisations adopt a systems approach to improving patient safety.

20 Patient safety in primary care Figure 20.1 Patient safety learning in general practice

Figure 20.2 Example of drug interaction alert linked to electronic medical record

Case note review

Culture survey

Significant event

analysis

Staff feedback

Complaints

Incident reports

Suggestion book

Surveys

Patient groups/ open

meetings

Clinicians

All staff

Patients

Box 20.1 Example of patient safety incident report from general practice

“Patient on oral methotrexate admitted to hospital diagnosed with Severe Community Acquired Pneumonia. Was prescribed Cefalexin 250mg four times daily and Nystatin mouthwash for sore mouth, tongue, and throat pain by their GP surgery on morning of admission. Antibiotic/ antifungal therapy was given following a telephone consultation. Patient not seen in person and there was no urgent Full Blood Count taken”

Medication Alert

New medicine

Drug

08:00

16:00

00:00

25 mg

Current medications

OK Start date Stop date Dose

Recommendation

May increase risk of gastrointestinal bleeding. If possible, avoid taking paroxetine with diclofenac

Diclofenac

Drug

Carvedilol 6.25 mg

Paroxetine 20 mg

01-08-13

08:00 6.25 mg

08:00 20 mg

Stop order Overide alert Respond to alert and cancel order

P art 3 R

isks to patient care

65 C

hapter 20 P atient safety in prim

ary care Introduction Around one in 50 patient encounters in primary care involve a patient safety incident, with substantial patient harm occurring for one in 20 of those incidents. With 303.9 million consultations occurring per year in general practices in England alone (2008– 2009), approximately six million patients will experience a patient safety incident and around three hundred thousand are likely to result in substantial harm to patients. General practice consulta- tions represent a subset of the encounters in primary care, so the actual burden from errors in this setting, including pharmacy, nursing, dentistry, out-of-hours services and community hospitals, for example, is likely to be substantially higher than the estimates suggests.

Factors infl uencing the risk of patient safety problems General practice is the most common fi rst level of contact with patients, and is unsurprisingly the largest specialty in most health- care systems. Th e way general practice is delivered can vary in terms of models of delivery (e.g. via general practice surgeries, community hospitals and GP-led health centres); the number of patients being served; the availability of services off ered, and spe- cialist services and secondary care support; the diversity of patients (e.g. ethnic groups and chronic disease profi le); and the size and skill mix of the healthcare professional team providing care.

In recent years, there has been an emphasis on general prac- tice teams providing and facilitating more complex care. Th is oft en involves coordinating input from primary, secondary and social care sectors. Additional issues include the variety of undiff erenti- ated complaints that patients can present with, accepting diagnos- tic uncertainty, acting as the gatekeeper for further investigations, specialist opinion or emergency assessment and managing multi- ple comorbidities. As a result of these issues, the potential risk of patient safety problems occurring inevitably increases.

Building a safety culture in general practice General practice delivery oft en involves teams of healthcare profes- sionals (GPs, nurses, nursing assistants, community midwifes and psychologists), as well as others such as administrative staff , prac- tice managers and offi ce caretakers. Th e team’s awareness of the potential for things to go wrong, and their ability to identify and acknowledge mistakes, learn from them and introduce changes to improve the situation next time, are essential to increasing patient safety. A general practice team with a strong safety culture might demon- strate the following attributes: • Commitment: Buy-in to the philosophy that safer, better quality care is possible, and improvement tools exist to deliver this vision for their patients. • Teamwork: Everyone in the team – clinicians and non- clinicians – are clear on the contributions they can make to improving the quality of care delivery. Th is might include know- ing that a staff suggestion box exists to improve work processes in- house, knowing how to access the incident reporting system via a

practice computer and knowing who to discuss potential problems within the practice. • Leadership: Each team member’s personal commitment to patient safety is refl ected in their preparedness to assume own- ership when problems occur. Individual members might work to champion patient safety by ensuring it is a standing agenda item at all practice meetings, carrying out signifi cant event audits that are led around priority issues identifi ed by the wider team and having advocate training for all staff on methods to improve patient safety. • Accountability: Taking responsibility for your actions at a pro- fessional, legal, ethical and contractual level. Th is might include: writing a patient safety incident report when an incident occurs; leading signifi cant event audits in areas that could put patients at risk of harm; being open with patients when an error occurs, and recognising the importance of saying sorry; and responsibly refl ecting on complaints made by patients and service users. • Understanding: Recognising that ‘every system is perfectly designed to achieve the results it gets’. Th is requires team members to look beyond blaming an individual and seek to identify the sys- tem factors that led to an event occurring. Individuals should rec- ognise that when the workload is high, the risk of error increases, and work processes need to be fl exible and adaptable to meet the demands of the service. • Communication: Openness to share good practice and voice con- cerns or opportunities for improvement through recognised chan- nels of communication. Patient safety should feature as a standing agenda item in practice meetings and feature in the annual practice report.

Methods for safer practice Th ere are several ways to improve patient safety in general practice, which can include: • Make patient safety a standing agenda item at all practice meet- ings. • Refl ect on results from audits of medical records that seek to: identify avoidable acute admissions and preventable deaths (e.g. poly-pharmacy in elderly) or patients lost to follow up (e.g. patients on lithium or anti-coagulation therapy); and detect and measure harm (read more about the global trigger tool in Chapter 16). • Discuss the learning from signifi cant event audits and incident reports (see Box 20.1) at practice meetings and other meetings for local learning (with multiple other practices), and at a national level (e.g. the National Reporting and Learning System; read more about reporting systems in Chapter 6). • Use technology to reduce risk to patients, such as computerised decision support tools to minimise risk of prescribing errors (read more about medical errors in Chapter 11; see Figure 20.2 for exam- ple interaction). • Provide patients and all team members with ways to give their views on improving patient safety (e.g. patient safety, incident report forms and focus groups), and feedback the risks identifi ed to them, the changes subsequently implemented and the lessons learned. • Have a clear policy for providing a prompt, full, honest and compassionate explanation of any incidents, with an apology in a timely manner.

Part 4Quality improvement

Chapters 21 Improving the quality of clinical care 68 22 Science of improvement 70 23 Model for Improvement 74 24 Measurement for improvement 78 25 Spread and sustainability of improvement 84 26 Quality improvement tools: visualisation 86 27 Quality improvement: assessing the system 88 28 Patient stories in improvement 90 29 Leading change in healthcare 92 30 Public narrative: story of self, us and now 96 31 Planning an improvement project 98 32 Managing an improvement project 100 33 Quality improvement in psychiatry 104 34 Quality improvement in intensive care 106 35 Quality improvement in obstetrics 108 36 Quality improvement in surgery 110 37 Population health and improvement 112

P art 4 Q

uality im provem

ent

21 Improving the quality of clinical care

Figure 21.1 A patient’s journey through an improvement active hospital

Table 21.1 Aims of quality improvement: examples of what they mean to patients and healthcare professionals

Aim Definition Intervention What this can mean for the

patient and their family What this means for the healthcare professional

Providing high-quality care regardless of a patient’s characteristics

Reducing harmful delays

Providing care informed by the best available evidence which provides clear benefits

Avoiding waste

Providing care that is respectful of the needs and values of the patient and their family

To avoid unintentional harm from care provided to patients

Process mapping in operating rooms (See Chapter 36)

Patient stories to identify opportunities to improve the care experience (See Chapter 28)

Paediatric Early Warning Score (PEWS) (See Chapter 16)

Patient and family centred

Efficient

Effective

Timely

Equity

Safe

Source: Adapted from the Institute of Medicine 2001 report, Crossing the Quality Chasm: A New Health System for the 21st Century.

Central venous catheter care bundle (See Chapters 18 and 34)

Cardiotocography (CTG) sticker (See Chapters 13 and 35)

Suicide prevention strategies amongst mental health patients (See Chapter 19)

Fewer cancelled operations and shorter waiting lists

Opportunity to share experiences to improve the experience for future patients

A healthcare team member is regularly assessing the child

Shorter overall hospital stay

Baby in distress is identified and delivered as soon as possible

Support and care are provided in and out of hospital when assessed to be at risk of suicide

More elective procedures per day and increased capacity for emergency cases

Move from asking, ‘What’s the matter?’ to ‘What matters to you?’

A patient showing early signs of deterioration receives escalated care and intervention

The bundle is a reminder that simple tasks done reliably can decrease length of stay in ICU and overall hospital stay, and minimise risk of death

The visual aid supports complex decision making in potentially stressful situations

A coordinated, multidisciplinary action plan is initiated for those at high risk of suicide

Equity: All non-English

speaking patients to receive translation

services to permit explanation of any

procedure

Effective: Use of ventilator

associated pneumonia and catheter

associated infection bundles

(See Chapter 34)

Timely and efficient: 95% of blood test

results to be available within 90 minutes of

blood draw

Safe:

Use of WHO Surgical Safety

Checklist (see chapter 12)

Patient and family centred: Daily planning of patient’s needs

(meal options and visitor times)

Outpatient appointment

Intensive care unit

Ward stay Surgery

(oophorectomy) Diagnosed with ovarian cancer

69 C

hapter 21 Im proving the quality of clinical care

Introduction Despite the hard work, compassion and commitment of busy healthcare professionals, problems in the systems they train and work in oft en lead to a less than optimal experience for patients. Th e moral and ethical compass of those in the caring profession is shaped by the precept ‘First, do no harm’. So it is not surprising that statistics such as ‘1 in 10 patients acutely admitted to hospital are harmed’ alarm clinicians and expose a contradiction of the funda- mental expectations of healthcare.

Th e focus on improving safety in healthcare has been invalu- able. However, overall quality improvement in healthcare can be described through six aims (adapted from the Institute of Medi- cine; see Table 21.1). It is when patient care deviates from the six aims that healthcare professionals rightly become frustrated with an inability to ‘do the right thing’, and wonder what improvements could be made ‘if only we could change things around here’. It’s encouraging that many healthcare professionals no longer merely express frustration with the clearly evident problems they see in their systems. Th ese professionals act. Th ey understand that they have two jobs – their clinical job (i.e. working with patients to co-produce health) and their improvement job (i.e. working col- laboratively to improve systems and processes so patient care and outcomes get better). So we now have a healthcare workforce that is increasingly aware of the responsibility, the duty, to improve. But knowing how to make improvements is not something most peo- ple are trained in. So how might you act to improve healthcare?

The science of improvement and the Model for Improvement Everyone involved in the planning and delivery of healthcare must continuously seek ways to improve the systems within which they work. One proven approach to do this is the science of improve- ment. Th e science of improvement combines expert ‘subject matter knowledge’ (i.e. your insights and know-how of the clinical issue to be improved) with improvement methods and tools to create an approach that emphasises innovation, rapid-cycle testing in vary- ing contexts and spread. Th e science of improvement is derived from W. Edwards Deming’s system of profound knowledge which requires an understanding of systems, an understanding of varia- tion, a theory of knowledge and an understanding of psychology (‘profound knowledge’ is discussed in more detail in Chapter 22).

Healthcare is complex, but using expert subject knowledge and the elements of Deming’s ‘profound knowledge’ to build as com- plete a picture as possible of the problem you want to improve is logical. But once you have identifi ed issues to improve on, it can be diffi cult to operationalise your intentions into actions. Th is is why many to most well-intentioned eff orts result in either a lot of talk about change with no actual changes, or a lot of changes implemented but with very little actual improvement. One of the ‘improvement methods and tools’ at the heart of the science of improvement is the Model for Improvement, developed by Associates for Process Improvement (‘disciples’ of Deming’s) and described in more detail in Chapter 23. Th ree questions form the core of the Model for Improvement: 1 What are we trying to accomplish? 2 How will we know that a change is an improvement? 3 What changes can we make that will result in improvement?

Th e other core element of the model is a series of plan–do– study–act (PDSA) cycles that inform the degree of belief that the changes proposed in Question 3 have the desired infl uence (increase or decrease) on the measures identifi ed by Question 2, thus edging closer (or otherwise) towards achieving the project aim identifi ed in Question 1. PDSA cycles are meant to be short and rapid, with each cycle producing knowledge that updates the ‘theory of knowledge’ and the design of the improvement eff ort.

Measurement techniques and statistical analysis will help you to determine which changes (and when) are leading to the improve- ments you seek (more detail in Chapter 24). Th ere is a temptation to rush to generate a summary of the changes that led to improve- ments in order to spread the intervention to other settings. But the Model for Improvement emphasises the importance of always starting small in diff ering contexts in order to generate the neces- sary knowledge for those contexts (see Chapter 25).

Supporting your use of the Model for Improvement Th ere are many improvement tools to support your use of the Model for Improvement (e.g. to inform your answers to each ques- tion and design your PDSA cycles). A sample of tools is described in Chapters 26–32 to visualise, understand and assess the current system and plan changes. Patient stories have also proven to be a powerful method for accelerating improvement eff orts by high- lighting issues that would otherwise be overlooked by statistical methods (Chapter 28). Great societal leaders have demonstrated powerful methods of communicating urgency for change using stories. Th ose who are leading improvement projects should use stories to catalyse their improvement eff orts (see Chapters 29 and 30).

Successful improvement projects are oft en run by a multi- disciplinary team representative of those working within the sys- tem that is being changed (see Chapter 31 for tips on planning your project). Th e importance of engaging a multi-disciplinary team and giving every team member a role and a voice cannot be overstated. At the Institute for Healthcare Improvement, they call this essential approach ‘all teach, all learn’ – it is an acknowledge- ment that in improvement, everyone has something to teach, and everyone has something to learn.

Whilst planning is essential, managing lots of ideas and peo- ple, delivering the change as well as addressing barriers during the project can be challenging. Having a set of tools to visualise your theories for change (e.g. a driver diagram), as well as project management methods (e.g. a Gantt chart) to ensure you and your colleagues stay focused on the aim, can help keep track of your improvement progress (Chapter 32). Improving healthcare is hard work, and fi nding inspiration in the leadership eff orts of others is crucial. Borrow ideas, share the learning as widely as you can (via conference presentations, journal articles, blogging etc.) and be sure to support other colleagues who are also interested in partici- pating in, or leading, improvement (Chapters 33–37).

Figure 21.1 charts a patient’s journey through an improvement active healthcare setting; whilst an improvement project might focus on one specifi c quality aim, many will refl ect multiple aims.

P art 4 Q

uality im provem

ent

22 Science of improvement Figure 22.1 Lens of profound knowledge

Understanding variation

Human side of change

Theory of knowledge

Appreciation of a system

Source: Langley G, Moen R, Nolan K et al 2009. Reproduced

with permission of John Wiley & Sons Ltd.

Le ad

ers hip

ph ilo

so ph

an d p

rac tic

al va

lue s

Improvement capability

Subject matter knowledge

Profound knowledge

Figure 22.2 Subject matter knowledge, profound knowledge and improvement

Figure 22.3 Adverse drugs events per 1000 patients

01 J

an 08

J an

15 J

an 22

J an

29 J

an 05

F eb

12 F

eb 19

F eb

26 F

eb 05

M ar

12 M

ar 19

M ar

26 M

ar 02

A pr

A D

E p

er 1

00 0

pa tie

nt s 4.5

4.0

3.5

2.5

3.0

2.0

1.5

1.0

0.5

UCL

LCL

UCL

LCL

UCL

LCL

UCL

LCL

UCL

LCL

Outer 1⁄3

Inner 1⁄3

1. A single point outside the control limits 2. A run of eight or more points in a row above (or below) the centreline

3. Six consecutive points increasing (trend up) or decreasing (trend down)

4. Two out of three consecutive points near (outer one-third) a control limit

5. Fifteen consecutive points close (inner one-third of the chart) to the timeline

Source: Lloyd P. Provost and Sandra K. Murray (2011). The Health Care

Data Guide: Learning from data for Improvement. Jossey-Bass.

Reproduced with permission of John Wiley & Sons Ltd.

Figure 22.4 Special cause patterns

71 C

hapter 22 S cience of im

provem ent

Introduction • To improve healthcare, there are four components of knowledge that a leader can use to guide change: • Th eory of variation • Systems theory • Th eory of knowledge • Psychology Figure 22.1 shows how these four theories interact with each other. A leader’s philosophy and practical values guide the use of this knowledge. W. Edwards Deming commented that we don’t have to be experts in these fi elds:

One need not be eminent in any part of profound knowledge in order to understand it and to apply it. Th e various segments of the system of profound knowledge cannot be separated. Th ey interact with each other. For example knowledge about psychology is incomplete without knowledge of variation.

Th e system of profound knowledge provides a map to under- standing and optimising organisations. When leading an improve- ment project, you should understand the basic theories in each area, how the diff erent areas interrelate and why they are impor- tant for the improvement of quality.

Leaders combine subject matter knowledge and profound knowledge in creative ways to develop eff ective changes for improvement (see Figure 22.2). Subject matter knowledge is the knowledge basic to the things we do in life and daily work (e.g. how to drive a car, how to programme an infusion pump, under- standing paediatric dosing regimes and scheduling outpatient appointments), and profound knowledge describes the intellectual depth and insight gained from understanding theories of systems, variation, knowledge and psychology.

Attributes of a leader with profound knowledge

Variation Healthcare is dynamic and complex. Leaders understand that variation can help explain the changes seen in project measures because of common causes and special causes: • Common causes: those causes that are inherent in the process over time, that aff ect everyone working in the process and that aff ect all outcomes of the process. Figure 22.3 shows a control chart that is dominated by common cause variation. It shows a chart that is in statistical control (a stable process) and is averaging two adverse drug events per 1000 patients • Special causes: those causes that are not part of the process all the time or do not aff ect everyone, but arise because of specifi c circumstances. Figure 22.4 describes fi ve possible patterns that

would be identifi ed as special causes (explained with further rules in more detail in Chapter 24)

Graphical methods help one to learn from data and can be used by others to consider variation in their decisions and actions. Leaders understand the concept of stable and unstable processes, and the potential losses due to tampering. When stakes from mak- ing changes are high, advanced methods can be used to measure the capability of a process or system before changes are attempted.

Systems Leaders study and manage their organisation as a system. Th ey emphasise the importance of common purpose and interdepend- encies among groups in the organisation. Th ey understand that the performance of the organisation depends more on the interaction of the various parts than how the parts perform individually. Th ey understand both detail and dynamic complexity in a system. Th ey consider important systems concepts such as boundaries, feedback loops, constraints and leverage points, and use these concepts to develop, test and implement changes to optimise the system. Fig- ure 22.5 shows the shift in focus from describing the organisa- tion in a hierarchical organisation chart to the organisation being viewed as a system that is focused on patients.

Knowledge Leaders understand that management is prediction that comes from knowledge, and knowledge is built on theory. Th ey under- stand that people learn in diff erent ways. Th ey use the plan–do– study–act (PDSA) cycle for learning and improvement to learn, run tests on a small scale and make decisions (described in more detail in Chapter 23). In order to enhance learning, they make pre- dictions before changes are made. Figure 22.6 shows the iterative nature of learning and how deductive and inductive learning are built into the PDSA cycle.

Th e leader understands that organisational learning depends on a shared understanding of key terms used. Operational defi ni- tions are necessary to create shared meaning. An operational defi - nition should include: • A method of measurement or test • A set of criteria for judgement

For example, what do we mean by a ‘patient fall’? Th is seems simple enough. Consider the following defi nitions of a fall: • “….a sudden uncontrolled, unintentional downward displace- ment of the body to the ground or other object, excluding falls resulting from violent blows or other purposeful actions” • “….inadvertently coming to rest on the ground, fl oor or other lower level, excluding intentional change in position to rest in fur- niture, wall or other objects”

How a fall is defi ned can be very diff erent from the perspective of the patient or caregivers. Th e elderly may consider a fall to be the

P A

D S

P A

D S

P A

D S

P A

D S

PDSA cycle 1

PDSA cycle 2

PDSA cycle 3

PDSA cycle 4

Prediction (based on a hypothesis, conjecture, model, theory)

Real world Observation, carry out test, look for anomalies

D eduction In

du ct

io n

D eduction In

du ct

io n

D eduction In

du ct

io n

D eduction In

du ct

io n

Source: Maccoby M, Norman CL, Norman CJ & Margolies R 2013.

Reproduced with permission of John Wiley & Sons Ltd.

Need

Design and redesign (Sherman)

Plan to improve quality (Porter)

Production of product or service (Meade, Buford, Reynolds)

Distribution (Hunt)

Suppliers (Halleck)

Customer research (Sheridan) Patients

(Grant)

Source: Maccoby M, Norman CL, Norman CJ & Margolies R 2013. Reproduced with permission of John Wiley & Sons Ltd.

VP support administration

Halleck

VP research and development

Sherman

VP marketing Sheridan

VP services Meade

VP distribution and logistics

Hunt

Region 1 Buford

Region 2 Reynolds

Region 3 Hancock

Planning Porter

President Grant

Support processes (Halleck)

Figure 22.5 Change of organisational view to a systems view focused on patients

Figure 22.6 Plan–do–study–act (PDSA), deductive and inductive learning

P art 4 Q

uality im provem

ent

73 C

hapter 22 S cience of im

provem ent

‘loss of balance’. If you are interpreting data from your organisation on falls, then the defi nition becomes important. Have we created a defi nition that is focused on patients? Have the defi nition and the measure masked our safety relative to falls? Has the defi nition caused an overreaction resulting in wasted time by healthcare pro- fessionals?

How we learn with deductive and inductive learning theory, and the use of operational defi nitions, only scratches the surface of the ideas that could be placed under the description of ‘theory of knowledge’. Th ere is much more to learn, so this is a start.

Psychology Leaders appreciate how to create an environment that is conducive to unleashing the power of intrinsic motivation. Michael Maccoby has described the ‘Five Rs of Motivation’ for leaders: 1 Reasons: People are motivated to support change when the rea- sons for that change make sense to them. 2 Responsibilities: People want to know how change will aff ect their responsibilities 3 Relationships: People are motivated by good relationships with supervisors, collaborators and the people they serve. 4 Recognition: People enjoy having their contributions to the system recognised. Recognition enhances the dignity of the indi- vidual, one of the basic human drives. 5 Rewards: It is essential that people consider all pay and rewards to be fair. Giving out bonuses can strengthen a supervisor’s

authority, but it won’t motivate workers to do their work any bet- ter. Worse, they may be demotivated by the pressure to conform to a job that is not intrinsically motivating.

Leaders also have what Maccoby has called ‘personality intelli- gence’. Personality intelligence is essential for creating motivation. Personality intelligence includes both concepts and emotional understanding, both head and heart. Th e concepts include: • Talents and temperament – what we are born with • Social character – how we are like others brought up in the same culture • Drives – how we are like all people • Motivational type – how we are like some people within our culture • Identity and philosophy – how we are no one else

Summary Four parts of profound knowledge have been discussed: • Th eory of variation – special versus common cause variation • Systems theory – viewing the organisation as a system • Th eory of knowledge – appreciating the way in which the PDSA cycle builds theory for eff ective action • Psychology – the important role of motivation and personality intelligence in working with people; probably 80% of the eff ort in making eff ective changes

P art 4 Q

uality im provem

ent

23 Model for Improvement Figure 23.1 Model for Improvement Figure 23.2 Defi nition of improvement

Model for Improvement

• What are we trying to accomplish?

• How will we know that a change is an improvement?

• What changes can we make that will result in an improvement?

Act Plan

DoStudy

Source: Langley G, Moen R, Nolan K et al 2009. Reproduced with permission of John Wiley & Sons Ltd.

Table 23.1 How will we know a change is an improvement? Figure 23.3 The plan–do–study–act cycle

Act Plan

DoStudy • Complete analysis of the data • Compare data to predictions • Summarise what was learned

• What changes are to be made • Next cycle?

• Objective • Questions and predictions • Plan to answer the questions (who, what, where, when)

• Carry out the plan • Collect the data • Begin analysis of the data

Source: Langley G, Moen R, Nolan K et al 2009. Reproduced with permission of John Wiley & Sons Ltd.

Alter how the work is done ... Improvement is the result of some design or redesign of the system

2 Produce visible, positive differences in results relative to historical norms

And lasting impact

In fe

ct io

Week 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16

12 11 10

9 8 7 6 5 4 3 2 1

Before

Change

After

Source: Langley G, Moen R, Nolan K et al 2009. Reproduced with permission of John Wiley & Sons Ltd.

Decrease CAUTI rate

Decrease inappropriate catheter use

Maintain or increase staff satisfaction

• Time between CAUTI rate

• Catheter days/100 patient days

• % nurse compliant to new process on review of catheter

• Score of the pre and post test

• Nurse satisfaction score • Physician satisfaction score

Increase nurses’ compliance to new process on review of catheter use

Outcome

Process

Balancing

Increase nurses’ knowledge on criteria for catheter use

Goals Measures Type of measure

75 C

hapter 23 M odel for Im

provem ent

Introduction All improvement requires change, and change can involve the design or redesign of a process. However, you need to be confi dent that the changes you are making will lead to the improvement you seek. You need to consider: • Why these changes? • How will we know the changes presented have improved patient care?

By asking these questions routinely whenever making changes or applying new knowledge, learning can be generated about what really helped to deliver results.

Th e Model for Improvement (Figure 23.1) is a method used worldwide in healthcare when making changes that aff ect patients. It has three fundamental questions which will help focus on what exactly the change is seeking to improve. Th e three questions may involve a series of plan–do–study–act (PDSA) cycles, where you develop, test and implement changes. It is rare to be able to answer the three questions without running PDSA cycles – few projects are that simple. Figure 23.2 defi nes what the improvement is fol- lowing the change.

1. What are we trying to accomplish? Th e answer to the question must state what process, product, ser- vice or system will be designed or redesigned. Table 23.2 describes some useful and less useful responses to this question.

Table 23.2 What are we trying to accomplish?

Less Useful Response for Question 1

Better Response for Question 1

Reduce infections Redesign the existing urinary catheter review and removal of criteria to reduce CAUTI in surgical recovery wards.

2. How will we know that a change is an improvement? When answering this question, there are three levels of measure- ment to consider: outcome measures relate directly to the aim of the improvement eff ort, process measures determine whether

activities have been accomplished as planned and balancing measures ensure that changes are acceptable to those who must make the changes and ensures that changes do not cause unin- tended consequences. Table 23.2 provides examples of goals and measures that relate to reducing catheter-associated urinary tract infections.

3. What changes can we make that will result in improvement? To answer this question, new changes must be developed that can then be tested, or known changes that are successfully being used by others can be tested. For relatively simple systems, a list of changes may be developed and tested almost immediately. Table 23.3 identifi es changes that were tested to reduce catheter- associated urinary tract infections.

The plan–do–study–act cycle Once the three fundamental questions are answered, the PDSA cycle (Figure 23.3) is then the primary means for turning ideas into action and connecting action to learning. Th e PDSA cycle simpli- fi es the scientifi c method to test ideas and determine what action should be taken. Using the cycle eff ectively takes some discipline and eff ort. Th e PDSA form in Figure 23.4 provides some detail on what should be considered in each phase of the cycle.

Table 23.3 Changes to test (PDSAs)

Test new urinary catheter decision-making algorithm To test eff ectiveness of training on improving nurses’ knowledge of appropriate indication on urinary catheter use Test integrating the results of the nurses’ daily catheter review into doctor-led ward rounds Test installing a computer screen reminder

Th e purpose of each cycle is to build your degree of belief that changes you are making are leading to the improvements you pre- dicted. Th is will allow you to identify which changes were eff ective and reject those that failed to meet predictions. In Chapter 25, you will consider why this is important as you spread changes to other clinical settings, hospitals and even other countries.

Figure 23.4 PDSA form

Cycle #1 Objective: Test new urinary catheter decision-making algorithm

PLAN

Learning Questions:

1. How will this change reduce catheter utilisation and CAUTI? Why? Prediction: Nurses currently wait for doctors to identify when a catheter should be removed. We think this will reduce utilisation from 1–2 days less. The longer catheters are in, the more likely CAUTIs can occur.

2. How will this change increase nurses’ knowledge on criteria of urinary catheter removal? Why? Currently, doctors are solely responsible for determining when a catheter is removed. A catheter removal algorithm will guide the nurses.

3. How will this change increase nurses’ compliance to the new process on review of catheter use? Why? Nurses will help develop a catheter review form to be used by nurses and use it to test with all the nurses. Because they helped develop it, they will own it.

4. How will this change impact doctor and nurse satisfaction? Why? Doctors will like this since they will not be solely responsible for determining time to remove, & nurses will like this since they will be able to remove the catheter earlier.

Test Plan: 1. The algorithm that is being used in other locations will be reviewed by the Medical Director and Nurse Sponsor and altered to match the conditions. 2. A new review form will be developed to make it easy for the nurses to follow the algorithm. 3. The Medical Director and Nurse Sponsor will approve the review form. And pre-test it with one nurse and her patients on Day x. 4. All nurses on day shift will be tested for their knowledge, then trained (with a post-test) to understand the decision- making criteria and how to use the form, by a senior nurse supervisor. Doctors will be informed about the change on Day y. 5. Test will be conducted for Day z. Results will be reviewed, and the test will be extended if it is successful.

Data Collection Plan: Nurse knowledge (pre & post training); catheter utilisation rate (pre-post test); time between CAUTIs; new process compliance (for extended test); staff satisfaction.

DO: (Execute the Plan) Observations/Surprises: Doctors who asked nurses for their catheter review results caused those nurses to be more confi dent and open. After the fi rst day, the nurses asked to extend the test to more patients.

(continued)

P art 4 Q

uality im provem

ent

77 C

hapter 23 M odel for Im

provem ent

Learning questions:

1. No CAUTIs have been reported. On the first day, 5 catheters were removed from patients

2. Nurse scores have increased

3. Percentage of inappropriate indication with corrective action taken (compliance)

4. Doctors are positive and want to move catheter review to ward rounds. Nurses want to implement the new protocol

ACT:

S co

1. Test adding the catheter review results in daily ward rounds for doctors to continue to build confidence in nurses 2. Test adding an electronic reminder on the handover system to remind the team to conduct a catheter review

Pretest Post test AVG STD

STUDY: (what was learned?)

100

Se p

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18

Au g

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 4241 43 44

Nurses

Pre PDSA 1 Mean rate: 90.2 PDSA 1

Mean rate: 78.6

100

Pre -PDSA 1 Mean %: 0%

PDSA 1 Mean 96: 95.3%

Ju ly

2 3 24 25 26

A ug

1 2 3 4

O ct

1 0 27 28 29 30

S ep

3 4 5 6 7 8 9 10 11 12 13 14 15 16 17

PDSA 1: test algorithm

P art 4 Q

uality im provem

ent

24 Measurement for improvement Figure 24.1 The seven steps to measurement

Figure 24.2 The measures checklist

Top tip 24.1: A test of a good aim statement: if you were in a lift could you clearly and briefly describe your aim in a sentence, i.e. the time it takes to travel from one floor to the next?

Top tip 24.2: Measure the minimum. Only collect what you need; there may be other information out there but the aim is to keep things as simple as possible

1. Decide aim

2. Choose measures

3. Define measures

4. Collect data6. Review measures 7. Repeat steps 4–6

5. Analyse and present

Part 1: Measure setup

Measure name: DNA rate for clinic A

Why is it important? We need to ensure that the clinic is not disrupted by having unexpected gaps in the clinic schedule. The policy for this clinic is to offer another appointment which means that other patients may be disadvantaged if we have too many patients being rescheduled

Who owns this measure? The outpatient manager

Part 1: Measure setup

What is the definition? (spell it out very clearly in words)

The percentage of patients booked to attend clinic A who did not attend for their appointment and no warning was received at the clinic before it started

What data items do you need? The number of patients booked to attend clinic (B) and the number of patients who failed to attend without warning (F)

100 x DNA patients (F)

Booked patients (B)

Which patient groups are to be covered? Do you need to stratify? (e.g. are there differences by shift, time of day, day of week, severity, etc)

All patients booked into the clinic

What is the numeric goal you are setting yourselves? We want to achieve no more than 5% DNAs

Who is responsible for setting this? The Outpatient Improvement project team

When will it be achieved by? Within 9 months from the start of the project

G oa

l s et

tin g

M ea

su re

d ef

in iti

Part 2: Measurement process

Is the data available? (currently available/available with minor changes/prospective collection needed)

Both data items are readily available from the outpatient module of the Patient Administration System

Who is responsible for data collection? The Information Analyst attached to the project team

What is the process of collection? The analyst will extract the relevant data weekly using existing reporting sofware

What is the process for presenting results? (e.g. create run chart or bar chart in Excel)

The DNA rate will be plotted on a run chart with each point representing the actual DNA rate for each clinic. The clinic runs twice a week so there will be two data points plotted per week

Who is responsible for the analysis? The Information Analyst attached to the project team

How often is the analysis completed? New data will be added every 2 weeks

Where will decisions be made based on results? The project team meets every fortnight and the data is a standing item on the agenda

Who is responsible for taking action? The project team leader together with the outpatient clinic manager

R ev

ie w

A na

ly se

C al

cu la

te m

ea su

re a

nd p

re se

nt r

es ul

ts C

ol le

What is the calculation?

79 C

hapter 24 M easurem

ent for im provem

ent Introduction Measurement is a fundamental component of the Model for Improvement, by answering the question ‘How do we know that a change is an improvement?’ To answer this, it is useful to collect and use the right data, but good data do not happen by chance – they are the result of a sound process. Using a process called the ‘seven steps to measurement’ (Figure 24.1) should help you to col- lect data that are good enough to use for your improvement pro- ject. A measures checklist also accompanies the seven steps model (Figure 24.2).

Steps 1 to 3 – Getting ready

Step 1 – Decide your aim Th is is a direct reference to the fi rst question in the Model for Improvement. Unless you are very clear what you are setting out to achieve, then you are unlikely to be able to decide what to measure. Create an aim statement that is unambiguous, specifi c, succinct and measurable. Not sure how good your aim statement is? Try the exercise in Top Tip 24.1.

Step 2 – Choose your measures Your aim will probably have suggested an obvious measure or measures. However, there are other tools to help you: • Chapter 32 introduces the driver diagram. If you have created one of these, you can use it to identify likely measures. Measures attached to drivers will mostly be outcome measures, and those attached to interventions will be process measures • If you have undertaken a process-mapping exercise (see Chapter 26) as part of your improvement work, this can also be useful in suggesting possible measures

You need to ensure a balance between outcome and process measures. Th ink you’ve not got enough measures? Check out Top Tip 24.2.

Step 3 – Defi ne your measures Th is is an area that can cause problems, particularly when using a common measure such as ‘length of stay’ because there are many defi nitions that might be used. Will the length of stay in hospital or in a particular ward be included, and will it be measured in whole days or in hours?

Measures always require common agreement on defi nitions. Th is means specifying exactly what is meant and applying this defi nition consistently. Make sure that your measure is well- defi ned and has clear instructions that can be easily followed and repeated both by yourself at a later date and by others. • Repeatability: Can you, the person who created the defi nition, collect the right data consistently using it? • Reproducibility: Can the defi nition that you have created be reproduced by other individuals?

Figure 24.2 shows the checklist completed for an improve- ment project undertaken in an outpatient department. Th e pro- ject team wanted to reduce the number of patients who did not attend (DNA) their appointments and have chosen the DNA rate as one of their project measures. In part 1 of the checklist, they have set out their rationale for choosing this measure and also defi ned it quite precisely. Th ey have set themselves a goal, but they will not regard it as the only measure of success for their project. Th ey know that they will learn a lot about why patients do not attend and that will help them design a process that works for patients.

Note that as well as defi ning the measure clearly, they have specifi ed the data items required and also the calculation they will use to arrive at the DNA rate.

Sampling When should we collect data on 100% of patients, and when should we use a sample? If your numbers are small enough or the data are easy enough to collect that you can cover 100% of patients, then you should do it. If this is not feasible, then you should use a sample. Th ere is a lot written about how to select samples (see the ‘Further reading’ suggestions for this chapter at the end of the book), but essentially it is advisable to choose a sample that is rep- resentative of the overall population that you are measuring. Th is is so that you do not inadvertently introduce a bias into your results.

For example, you decide to choose fi ve patients ‘at random’ from an outpatient clinic and take the top fi ve case notes from the pile by reception. What you may not know is that the notes were placed in a particular order and you have chosen fi ve female patients in a mixed-gender clinic. Does this matter? In some circumstances, it will. So use your judgement based on your knowledge of the pro- cess (i.e. your subject matter knowledge) you are sampling from.

One way you can avoid bias is to use a random number gen- erator in spreadsheet soft ware packages.

Steps 4 to 7 – The collect–analyse–review (CAR) measurement cycle Measurement itself is a process. In its simplest form, it consists of three stages. First you collect the data, then you analyse them and present them in an appropriate way to convert them into useful information, and fi nally you review your information to see what decisions you need to make. Th e CAR cycle then starts all over again. Th e CAR cycle will help you to be explicit in your ‘plan’ about the data you need and how you ‘do’ collect and display it (Steps 4 and 5) to answer the questions and predictions you intend to review (Step 6) in order to ‘study’ and ‘act’ accordingly.

Step 4 – Collect your data Once you have decided what data you need, you are all set to start collecting them. However, unless you already have a system set up to give you exactly the data you need, you will need to plan how to collect them. And that plan will need to answer the following questions: • What are we collecting: data for every patient or just a sample? • Who is actually going to get the data? • How are they going to do it? • Where will they get it from: historically stored in a database or collected in real time?

Test your plan, using plan–do–study–act (PDSA) principles, to ensure the data you get back are reliable and what you were expect- ing. Worried that collecting data is a burden? See Top Tip 24.3. A word about baselines You will need to know your baseline (i.e. how your process is cur- rently performing) before you can track the progress toward your goal. To create a baseline, about 20–25 data points are ideal. Th is is because you want to see not just how you are doing ‘on average’, but also how variable your performance is and whether there are any cyclical trends to be aware of.

Th e number of points you need to demonstrate a diff erence will also depend on the size of the eff ect you expect to have (see Figure 24.4). Table 24.2 also gives further guidance on this

P art 4 Q

uality im provem

ent

Top tip 24.3: Aim to make measurement part of the daily routine. Where possible, use mechanisms that are already in place or amend them. This minimises the burden on staff and also maximises the chances of it being done reliably

Top tip 24.4: Remember the goal is improvement and not a new measurement system. It’s easy to get side-tracked into improving data quality, especially if you are confronted with challenges on the credibility of the data – just ensure it is ‘good enough’

Number of data points

Lower limit for number

of runs

Upper limit for number

of runs

10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 60 70 80 90 100

3 3 3 4 4 4 5 5 6 6 6 7 7 8 8 9 9 9 10 10 11 11 11 11 12 13 13 13 14 14 15 16 16 17 17 17 17 18 18 19 19 24 28 33 37 42

8 9 10 10 11 12 12 13 13 14 15 15 16 16 17 17 18 19 19 20 20 21 22 22 23 23 24 25 25 26 26 26 27 27 28 29 30 30 31 31 32 37 43 46 54 59

Source: Perla RJ et al 2013. Qual Manag Health Care.

Table 24.1 Tests for number of runs above and below the median

Ja n

F eb

M ar

A pr

M ay Ju n

Ju l

A ug

S ep O ct

N ov

D ec Ja n

F eb

M ar

A pr

M ay Ju n

Ju l

A ug

S ep O ct

N ov

D ec

2006 2007

Median

Figure 24.3 The number of complaints as a run chart

Table 24.2 Guidelines for number of data points to test a change (Source: Adapted from Langley G, Moen R, Nolan K et al 2009. Reproduced with permission of John Wiley & Sons Ltd.)

Total number of points Situation Fewer than ten Expensive tests, expesive protoypes,

or long time periods between available data points. Large eff ects anticpated.

Fift een to fi ft y Usually suffi cient to discem patterns indicating improvements that are moderate or large.

Fift y to one hundred Th e eff ect of the charge is expected to be small relative to the variation in the system.

Month

Value

Moving range

Jan Feb Mar Apr May Jun Jul

36 – 31 = +5 27 – 36 = –9

Figure 24.5 Calculating moving ranges

Figure 24.6 Formulae for upper and lower control limits

Upper control limit =

Lower control limit =

Mean +

Mean –

3 x Av MR

1.128

3 x Av MR

1.128

Figure 24.4 Attendances at an accident and emergency (A&E) department as a statistical process control (SPC) chart

A pr

M ay Ju n

Ju l

A ug

S ep O ct

N ov

D ec Ja n

F eb

M ar

A pr

M ay Ju n

Ju l

A ug

S ep O ct

N ov

D ec Ja n

F eb

M ar

10,000

9000

8000

7000

2011 2012 2013

A tte

nd an

ce s

Upper control limit (9222)

Lower control limit (7167)

Median (8195)

81 C

hapter 24 M easurem

ent for im provem

ent based on the availability of data given the circumstances. One way to get more points is to measure more frequently.

Sometimes the data you need are not currently being collected. If so, you should start collecting your data straightaway. You can still start testing small changes. Th ey will not aff ect your overall situation, so you can pursue those whilst creating your baseline.

Final word: do not go overboard on data collection. See Top Tip 24.4.

Step 5 – Analyse and present your data Special-cause and common-cause variation (described in Chap- ter 22) are concepts used to distinguish between whether we have made a diff erence (special cause) or if the changes observed in our data are simply a result of random variation (common cause). Both run charts and statistical process control (SPC) charts can help to distinguish between these two types of variation.

Data are plotted in time order, and the median is represented as a straight line. Figure 24.3 provides an example of a run chart illustrating the number of complaints received each month in an organisation.

Th ere are four statistical rules to identify possible special causes: Test #1: Six or more consecutive points above or below the median. Th is indicates a shift in the process. Values are still vary- ing, but they are doing so around a new median value. If this is a shift in the right direction, it is likely that the change you made is having a benefi cial eff ect. Th is is the most frequent ‘rule’ that you will see aft er introducing a change. Test #2: Five or more consecutive points all increasing or decreas- ing. Th is indicates a trend and suggests that the change you made is having an eff ect, but you don’t know yet when performance will become stable again. You need to keep measuring to fi nd out. Th is situation is more likely to occur if you are rolling out a change over a period of time. Test #3: Too many or too few runs. A ‘run’ is a number of con- secutive points on one side of the median. When counting, ignore any points that lie directly on the median. Use Table 24.1 to work out whether your variation is due to random causes. If the number of runs is inside the range, this is what we might expect by chance. If the number falls outside the range, then some external factor is having an eff ect. Too many runs suggest that the process has become less consistent, and it is possible that your change has had a detrimental eff ect. Too few runs suggest a more consistent process. Test #4: An ‘astronomical’ data point. You use your own judge- ment to assess whether the data value in question really is ‘odd’. Oft en, such markedly out-of-range results are caused by a data col- lection or data defi nition problem, so check that fi rst. If the data seem ok then try to fi nd out what might have caused such an odd result. It may cause you to think about creating a contingency plan for if such an occasion arose again.

Are any of these four rules present in the example given in Figure 24.3? Which months, if any, would you want to ask more questions about if these were data about a service you knew? Model answers are provided at the end of the chapter.

Statistical process control charts Run charts should be suffi cient for nearly all your measurement, but there may be occasion for you to need to understand how much variation the process of interest typically exhibits. For this, you will use an SPC chart. Th is chart (Figure 24.4) introduces the idea of expected variation. SPC charts still have a ‘typical value’ line (usually, the mean) but add two further lines, the upper and lower limits.

In Figure 24.4, we have plotted the monthly number of attend- ances at an accident and emergency (A&E) department over 2 years. Th e average number per month is 8195 attendances. Th e upper and lower limits show that attendances could vary by up to 1000 attendances on either side of the mean.

Th e purpose of these control limits is to show you that data points appearing within the limits, despite going up and down, are doing so as part of the normal variation that we see in eve- ryday life. If a data point spikes above or below these limits, then you know something diff erent has happened – a special event, hence this is called ‘special-cause variation’. When events like this are seen on a chart, you need to investigate what happened. Even though the event might be unlikely to occur again, it is still worth considering if there is anything you can do to minimise the impact if it did.

If our process exhibits just random variation, we can use the SPC chart to ‘predict’ what future performance would be like. We would expect any future data points to vary around the average and lie within the limits.

Calculating the upper and lower limits Th e upper and lower limits are derived from the actual data them- selves. Th e more variable the data, the further apart the limits will be. Th ere are several diff erent types of SPC chart, each with its own way of calculating the limits and each designed for a specifi c type of data. However, there is one chart, called the XmR chart, that statistical process control experts like Don Wheeler suggest works for almost any type of data.

To calculate the upper and lower limits for this chart, do the following: • Calculate the moving range for each pair of points in your data (see Figure 24.5) • Calculate the average of the moving ranges (‘Av MR’ in Figure 24.6) • Divide the Av MR by the bias correction constant, 1.128. What- ever your data, this is always the same value. Th is gives you one measure of variation (V) • Add three measures of variation (3V) to the mean to get the upper limit, and subtract three measures of variation (3V) from the mean to get the lower limit. Th e formulae are shown in Figure 24.6

Special-cause rules Th ere are several statistical rules to identify possible special causes when using SPC charts. Th e eight rules described by Lloyd Nelson will help you to do this. Th e fi rst three rules can be used without needing specialist SPC soft ware and so are the most commonly used. Note that rules 1–3, 5 and 7 have already been illustrated in Chapter 22, but we include them here for completeness and con- venience: Rule #1: (Figure 24.7) – For any point outside one of the con- trol limits, similar to an ‘astronomical point’ in run charts, SPC allows a statistical interpretation rather than just personal judgement. Rule #2: (Figure 24.8) – A run of eight consecutive points all above or all below the centre line. Th is is the same as rule #1 for run charts and indicates a shift in the process. Th e actual num- ber of points you take as a ‘signal’ (typically seven, eight or nine points are used) depends on your attitude towards risk. Th e more points you want to see, the less likely that sequence would happen by chance. For example, seven in a row occurs less than one in 100 times by chance.

P art 4 Q

uality im provem

ent

Mean

Lower control unit (LCL)

Upper control unit (UCL)

Mean

LCL

UCL

Mean

LCL

UCL

Mean

LCL

UCL

Mean

LCL

UCL

Mean

LCL

UCL

Mean

LCL

UCL

Mean

LCL

UCL

Mean

LCL

UCL

Mean +3V

Mean +2V

Mean +1V

Mean

Mean –1V

Mean –2V

Mean –3V

Mean +3V

Mean +2V

Mean +1V

Mean

Mean –1V

Mean –2V

Mean –3V

Mean +3V

Mean +2V

Mean +1V

Mean

Mean –1V

Mean –2V

Mean –3V

Mean +3V

Mean +2V

Mean +1V

Mean

Mean –1V

Mean –2V

Mean –3V

Mean +3V

Mean +2V

Mean +1V

Mean

Mean –1V

Mean –2V

Mean –3V

Figure 24.7 Rule #1

Figure 24.9 Rule #3

Figure 24.11 Showing measures of variation

Figure 24.13 Rule #6

Figure 24.15 Rule #8

Figure 24.8 Rule #2

Figure 24.10 Rule #4

Figure 24.12 Rule #5

Figure 24.14 Rule #7

83 C

hapter 24 M easurem

ent for im provem

ent Rule #3: (Figure 24.9) – Six or more consecutive points all increasing or decreasing. Th is indicates a trend and is similar to rule #2 for run charts. Rule #4: (Figure 24.10) – Fourteen points in a row alternating up and down. Th is is the classic sawtooth pattern that we oft en draw when sketching random variation freehand. In fact, this pattern is anything but random.

Th e following four rules now divide the space between the centre line (the average) and the control limits into three equal parts using the measure of variation calculated earlier (V). See Figure 24.11. Rule #5: (Figure 24.12) – Two out of three consecutive points more than two measures of variation on the same side of the centre line. Th e size of the change means that it can be identifi ed more quickly than rule #2. Rule #6: (Figure 24.13) – Four out of fi ve consecutive points more than one measure of variation on the same side of the centre line. Th e size of the change, although not as great as for rule #5, means that it can be identifi ed more quickly than rule #2. Rule #7: (Figure 24.14) – Fift een consecutive points close to the centre line that are within one measure of variation. Th ese points can be above or below the centre line. Th is happens when your process becomes more consistent or reliable. Rule #8: (Figure 24.15) – Eight points in a row either above or below the centre line with none within one measure of variation. Th is indicates that the process has split in some way if it is a new phenomenon or that the data come from two sources, perhaps a morning shift and an aft ernoon shift , if it is present from the start. Th e solution in this case would be to plot the two shift s separately.

Step 6 – Review your data It is vital that you set aside time to look at what your measures are telling you about your process and any impact changes that your processes may (or may not) have made. Remember that the

purpose of measurement is to guide you to make the right deci- sions about your improvement project.

Step 7 – Keep going! Repeat steps 4, 5 and 6 each month or more frequently. If you are measuring compliance with a process (e.g. compliance with handwashing or a care bundle), aim for a minimum of 95% for non-catastrophic processes. For a process where, if it fails, it will almost certainly result in serious injury or death, aim for 100%. Keep making changes until your data tell you this is so.

For outcomes (e.g. the surgical site infection rate or number of central line infections), you are aiming to consistently meet or exceed your goal. If you are using SPC charts, ensure the goal sits outside the appropriate upper or lower limit.

When do I stop measuring? Th e simple answer is you don’t. If you are consistently meeting your goal, you should still look to see if there are further improve- ments that could be made or monitor that things do not get worse again. Where you have a consistent performance in process meas- ures, you may decide to measure less frequently. Th is is fi ne as long as there is a related outcome measure that you are continuing to monitor. If this shows signs of deteriorating, you may need to revisit the process measures.

Some measures that you choose will specifi cally relate to a change proposed within the project. For instance, you might be introducing a new piece of equipment and staff will need conver- sion training. Whilst you are running a series of PDSAs to learn about your eff orts to convert staff , you might record data about compliance to key steps for safely using the new equipment. Th ese would be ‘PDSA-level measures’ (as opposed to ‘project- level measures’), and once you are satisfi ed that all staff are suc- cessfully trained, you do not need to continue measuring your progress.

Did you correctly interpret Figure 24.3? There is a run of seven points below the median at the end of 2007. This is rule #1. There are 2 months (March and May 2007) that could be considered ‘astronomical’. There are no special causes present before March 2007. You would want to be asking whether anything unusual was happening in March and May 2007 and did the organisation take any measures to improve services in May/June 2007?

P art 4 Q

uality im provem

ent

25 Spread and sustainability of improvement

Figure 25.2 Improvement phase by setting

Improvement phase

S et

tin gs

Innovation Testing Scale-up and spread

Scale-up

University teaching hospital

District general hospital

Rural medical centres

The figure illustrates how a new model may be developed in one or two similar organisations during an innovation phase. The model may then be tested in an increasing number and variety of settings during the testing phase. In the scale-up and spread phase, the model is then implemented in an increasing number of settings. In particular, there may also be a system of organisations that employ a deliberate scale-up strategy aimed at implementing the model in all associated organisations

Figure 25.1 Ideas and their results in various settings

C on

fli ct

Innovative sample

Learn which contexts it can be amended to work in as we move from innovation to ‘test and spread’

Effective Not effective

% e

ffe ct

iv e 100

0 100% 95% 90% 90% 93%

C on

fli ct

Innovative sample

Immediate wide-scale implementation

Effective Not effective

% e

ffe ct

iv e 100

Spread sample

100% 80% 70% 60% 50%

A possible result of taking an idea that has emerged from a small number of settings, with a narrow context, and immediately evaluated as a spread initiative as a fixed protocol. This idea was found to work in 100% of the ‘innovation’ sample settings and only 50% of the ‘spread’ sample

A possible result of taking an idea that has emerged from a small number of settings, with a narrow context, and learning which contexts it can be amended to work in. This idea was found to work in 100% of the ‘innovation’ sample settings and over 90% of settings where the learning suggested they matched the contexts in which it was likely to work

Source: Adapted from Parry GJ, Carson-Stevens A, Luff DF, et al (2013). Reproduced with permission of Elsevier.

85 C

hapter 25 S pread and sustainability of im

provem ent

Introduction Healthcare improvement aims to improve outcomes for patients by encouraging the adoption of improvement interventions that have solid evidence of eff ectiveness. However, in healthcare, in a review of 34 evidence-based interventions that had been replicated in other settings, 41% were found to have a smaller eff ect size or were not found to be eff ective. Th ere is, therefore, some uncer- tainty around how best to adopt and implement new interventions.

Improvement initiatives when fi rst established can be eff ec- tive, but as they became more widespread, a diminishing eff ect on outcomes can be seen. One explanation (Figure 25.1) is that the eff ectiveness of an intervention is oft en based on studies in a small number of settings. Th e complexity of the intervention may not be fully understood before it is spread to another hospital or setting. For example, the ‘Keystone ICU project’ (described in Chapter 18) where intensive care units in Michigan, USA, reduced central venous catheter bloodstream infections by introducing techni- cal interventions (changes in clinical practice) and non-technical interventions (linked to leadership, teamwork and culture change), has not always been successfully replicated elsewhere. Existing research suggests this is because the complexity of the interven- tion may not be fully understood. In this situation, a simple but intuitively appealing summary model of the changes needed to produce improvement becomes a fi xed protocol rather than the basis of multiple plan–do–study–act (PDSA) cycles for teams to adapt the interventions locally. Local adaptation could permit understanding about the training needs of diff erent staff groups and identify key clinical and managerial processes that would need to be in place for the intervention to be successful.

Th ere is a risk of rushing to generate a summary of the changes needed to produce improvement outcomes too quickly, with- out truly understanding which change(s) led to improvement. In healthcare improvement, a promising new intervention (or change model) can be found to be eff ective in a small number of settings, but when replicated as a fi xed protocol across a broad range of con- texts, it is found to be ineff ective. One way in which the improve- ment fi eld aims to address such issues is by encouraging the small- scale testing of new changes, fi rst in a small number of settings and then in an increasingly wide variety of settings.

Innovation phase All new models of care have to originate from somewhere, and in improvement this can be referred to as the innovation phase (Figure 25.2). Typically, in this phase, a new model of care (or intervention) may have been developed in one or two settings. For example, a university teaching hospital may have developed and implemented a checklist aimed at ensuring that a number of

evidence-based procedures are followed prior to a hip replace- ment. Th e innovation phase may aim to understand what impact the intervention has on patient outcomes and to describe the main changes that lead to improvement so that others can adapt it to their setting. Th ere may be some evidence that the intervention is eff ective in this single setting, but it may not be clear whether it will be applicable or eff ective in many other settings.

Testing phase Th e testing phase is used to help understand where an interven- tion can be adopted to work elsewhere. In this phase, improvement methods such as the Model for Improvement and PDSA cycles are used to undertake small tests of applying the new interven- tion in new settings. It is expected that improvement teams will undertake a number of PDSA cycles, adapting components of the intervention in order to learn whether or not it will work for them. Th e number of settings will increase and the types of settings will widen as evidence builds. For example, other university teaching hospitals may test the new model as well as non-academic, rural hospitals or hospitals with a diff erent patient case mix from the original setting.

Problems can occur if the testing phase is not well conducted in a wide variety of settings, or is not undertaken at all. For exam- ple, if a hospital system decides to scale up a checklist based on evidence obtained from a single setting in the innovation phase, they may not have fully identifi ed the key components of the inter- vention, and it may be that the checklist approach is only eff ective for certain patients in certain settings. Implementing the interven- tion across a wide variety of settings is likely to lead to poor over- all eff ectiveness, resistant surgical staff and little if any sustained improvement in patient outcomes (see Figure 25.1).

Scale-up and spread phase Th e scale-up and spread phase aims to broaden implementation. Scale-up occurs when there is a deliberate eff ort to push out or implement a new model on a long-term basis, across a par- ticular system, in order to achieve sustained improvement in patient outcomes. For example, a large hospital system may aim to implement a surgical checklist across all component hospi- tals. To do so, hospital leaders may hope to bring about sus- tained improvement by introducing policy changes requiring action by surgical teams as well as midlevel and senior clinical managers. Spread may occur when other settings wish to repli- cate the new model. For example, hospitals may have heard or read about a new surgical checklist associated with improved patient outcomes, and they may decide to try to introduce the checklist in their setting.

P art 4 Q

uality im provem

ent

26 Quality improvement tools: visualisation

Figure 26.2 Example of a swim lane process map

Figure 26.1 Common fl owchart symbols

DelayBasic process

DecisionStart/end point Patient flow

Information flow

Documentation

Wait Change

into gown Pre-op

assessment

Collect patient

information

Clinical aid

Nurse

Transporter

Secretary

Patient

NoYes

Yes

Outpatient

Patient arrives from floor

Pre-op interview

Pre-op assessment Orders?

Assist patient

with gown

Create chart for patient

Pre-op interview

Patient needs assistance

Patient checks in

When blown up to be something you could actually read, this swim lane process map is 12 feet long and 3 feet high! It details how each stakeholder contributes to a surgical procedure, including how the patient and information move throughout the day. All process maps certainly do not require this level of detail (see Chapter 36 for a worked example in surgery)

Suppliers

List suppliers here

Inputs

List inputs here

Process

Include process map or list the processes here

Outputs

List outputs here

Customers

List customers here

Figure 26.3 Example of a detailed swim lane process map for the day of surgery

Table 26.1 A suppliers, inputs, processes, outputs and customers (SIPOC) diagram template

87 C

hapter 26 Q uality im

provem ent tools: visualisation

Introduction Developing a better understanding of a process – rather than just your own tasks – is critical to improving quality and patient safety. Th is type of systems thinking is applicable beyond formal improve- ment work and can help clinicians to focus on healthcare qual- ity continuously. A process map is a quality improvement tool to graphically represent steps in a process, and it helps everyone on the team to be sure that they understand the fl ow, delays, tasks and responsibilities associated with that process. An incomplete picture of a process is like looking at a portion of a painting with a magni- fying glass without seeing what the whole painting actually depicts.

Creating a process map of the current system can help you to understand the process you are seeking to improve: how it works, who is involved and how it fl ows. It is also a useful visual aid for communicating the process to others who do not understand your process. Flowchart symbols (see Figure 26.1) are used to represent the sequence and fl ow of events throughout a process.

Th ere are multiple approaches for building process maps – the actual state, the intended state or the ideal state – and it is some- times helpful to consider more than one of these. If you want to map what is actually occurring (the actual state), you need to directly observe the process as the map is being created. What you observe can oft en be diff erent from what people think is (or should be) happening, and therefore the intended and actual pro- cess maps may diff er from one another.

Creating a process map 1. Determine the scope of the process to be mapped Choose a specifi c starting point (a process trigger) and a specifi c ending point. Th e starting and ending points are critical to avoid ‘scope creep’, where the process map slowly grows to become unmanageable. It may not be eff ective to develop a process map of a patient’s entire length of stay using a trigger of admission and an endpoint of discharge – that would be a very large process map! However, it could be helpful to map the patient experience during preoperative activities on the day of surgery to examine how the surgeon, anaesthesiologist, nurses and nurse anaesthetist interact with the patient. Additionally, the level of specifi city is an impor- tant component of ‘scope creep’ as the map can get very large if the process to be mapped is complex. For example, entering patient data into the patient’s notes may be considered a single process block (e.g. ‘enter patient data’), or each data element could be con- sidered a separate task (e.g. ‘enter allergies’ and ‘enter medications’).

2. Identify all of the stakeholders who have an effect on the process Choose representatives from the stakeholder groups – those directly or indirectly involved in the process fl ow or outcomes – to

help you develop the process map. It may not be feasible to have everyone involved in the process work together on the map, but observations and interviews can help fi ll in process activities. Th is activity can help to resolve misunderstandings about what is actu- ally occurring in a process.

3. Have everyone on the team list the individual steps in the process It is helpful to have stakeholders brainstorm and list all of the steps fi rst, without focusing on the actual sequence of events. If each member of the team writes each individual step on a sticky note or index card, then the entire group can sort through and sequence the steps on a wall or table.

4. Work as a team to agree on the actual sequence of events and the appropriate fl owchart symbols While you are working together, it is almost certain that you will generate possible solutions that could better the process that you are trying to understand. Your focus now is on creating the map, but be sure to document these ideas. In addition, you should docu- ment any assumptions you use.

5. Share your map with all stakeholders and make edits or additions as necessary Allow stakeholders to comment on your draft s, especially if they were not involved in its creation. Th eir insights can help clarify and refi ne the process map. Healthcare processes are complex, and everyone might not agree on the current state process map. Th is is an opportunity to identify areas for improvement or clarity.

Additional details A swim lane process map has an additional level of detail: ‘lanes’ guide the alignment of the steps in the map to visually represent who is responsible for certain aspects of the process (Figures 26.2 and 26.3). Th is type of process map may be particularly useful to illustrate patient or information handoff s.

A suppliers, inputs, processes, outputs and customers (SIPOC) diagram (Table 26.1) can help you frame a process map by fi rst listing these additional elements. For example, consider the pro- cess depicted in Figure 26.2. To determine what to map, it may be helpful to consider the diff erent providers (e.g. surgeon, anaesthe- sia provider and circulating nurse) who supply inputs to the pre- operative process (interviews, orders for additional tests etc.). Lab results, completed forms and other outputs are necessary before the patient and surgical team (the ‘customers’) can proceed to sur- gery. Once you’ve considered the suppliers, inputs, outputs, and customers, it may be much easier to create the actual process map of more complex systems.

P art 4 Q

uality im provem

ent

27 Quality improvement: assessing the system

Figure 27.1 Sample failure modes and effects analysis (FMEA) for potential failures to the magnetic resonance imaging (MRI) schedule from a patient arriving late to their appointment

Potential failure modes

Possible causes

Probably of occurrence (P)

Severity of effect (S)

Ease of detection (D)

Risk priority number RPN = (P*S*D)

Transporter gets lost

Slow transport

No transporter available

Patient not ready for transport

Elevator out-of-order

Patient does not arrive to the MRI

appointment on time

245

128

It would be difficult for a team to detect if a

transporter gets lost

Not having a transporter available may have a severe effect on the

MRI schedule

While this may not be likely, it is possible

This is the highest RPN, so we should

focus on this possible cause first

The elevator being out-of-order is a

very low probability

It would be easy to detect if the elevator

was out-of-order

Figure 27.2 Sample fi shbone diagram for late surgical procedures at the start of each day

Environment Personnel

Room set up late

Difficult airway

Patient late

Patient Policy Materials and equipment

Missing paperwork (consent or history

and physical)

No preadmission appointment

Implants or special equipment late

Late labs

Non-sterile equipment

OR team late

Physician late

Vender late

Late first case On-time start

Table 27.1 Failure modes and effects analysis (FMEA) template

Potential failure modes

List the ways your system

can potentially fail here

Possible causes

List potential causes for each

of the failure modes here

Probability of occurrence

Assign a score (1–10) for each

of the failure modes

Failure effect

List what happens when

each of the failures occur

Severity of effect

Assign a score (1–10) for each

of the failure effects

Detection method

How do you currently know

when the failure has occurred?

Ease of detection

Assign a score (1–10) for how easily failures are detected

Risk priority number

Multiply the probability of occurrence, severity of effect, and ease of

detection scores

89 C

hapter 27 Q uality im

provem ent: assessing the system

Evaluating your process Once a process map has been created, it can facilitate multiple strategies to evaluate potential failures, errors and breakdowns in the process. Two of these approaches are the failure modes and eff ects analysis (FMEA) and the why-why approach. Generally these brainstorming techniques do not generate solutions to issues, but rather serve as a means to identify and preliminarily prioritise potential quality improvement (QI) areas.

Processes are designed to work a certain way. When processes do not function optimally, they may contain potential oppor- tunities for system failures. Even if a patient is not harmed, a pro- cess breakdown could still be a considered a system failure. Th e two tools described below can be used to identify potential causes of a process breakdown.

Failure Modes and Effects Analysis Th e purpose of FMEA is to preemptively examine potential oppor- tunities for your system or process to fail, generate errors or break down, and, through a structured approach, prioritise improvement projects to address these potential failures. FMEA is not generally used as a reactive tool as its purpose is to help identify potential failures before they occur.

Once you identify an opportunity to use FMEA, choose a team that represents your stakeholders. Th ere are several variations on the FMEA method. We recommend the FMEA template shown in Table 27.1 to organise your brainstorming and guide the assign- ment of priorities for improvement opportunities using the follow- ing steps: 1 With the team, brainstorm all of the ways your process could fail. Each process in the process map (see Chapter 26) can fail if it does not do what it is intended to do. 2 For each of the process breakdowns you identify in step 1, list all of the potential causes that could contribute to that failure. Th is is where you consider how a failure could happen. Consider all of the potential causes for each of the potential failures, but note that there could be overlap. 3 Estimate the probability of occurrence for each of the poten- tial causes you identify in step 2. Assign a score of 1–10, where 1 represents something that is not very likely to occur and 10 repre- sents something that will almost certainly occur. 4 For each of the causes you identify in step 2, consider the eff ect of the failure as well as the severity of that eff ect. List what happens when failures occur, and assign a score of 1–10 to repre- sent the severity of the occurrence. Lower scores represent an out- come that presents very little risk, where a score of 10 represents the most severe outcome.

5 Consider the current methods to detect when a failure has occurred. What methods, controls or warnings are in place that would help you and your teams identify when a failure has occurred? Assign a score of 1–10, where 1 represents something that you would almost certainly detect and a score of 10 indicates a failure that is surely unlikely to be noticed. 6 Calculate a risk priority number (RPN) by multiplying the prob- ability of occurrence, severity of eff ect and ease of detection scores.

For several of the tasks above, the user assigns a score from 1 to 10 ranking the probability of occurrence, severity of the failure and ease of detection. Th ere is diversity in the QI literature about how each score should be selected and operationalised, and oft en in the healthcare setting, a 1–5 scale is used. Th e most important thing is that you and your team are consistent in how you select these scores. Very simply, lower scores indicate lower probabilities of occurrence, lower probabilities of severity and greater ease of detection. For example, Figure 27.1 is a basic example of an FMEA but demonstrates how scores can be selected for specifi c causes.

Potential causes with the RPNs should become high-priority improvement initiatives for you and your team. However, certain potential causes should be given highest priority. You should think carefully about the limits of severity scores and occurrence rates; generally, experts recommend prioritising failure modes with severity scores of 7–10 and events that can occur very frequently.

Why-why approach A why-why analysis is used to identify root causes that may lead to potential failures, errors and breakdowns in processes. Ask each QI team member to refl ect on why a potential problem could occur. Th is will likely lead to multiple ideas, and for each of these continue to ask, ‘Why?’ until a potential root cause has been identifi ed. Typi- cally, asking ‘Why?’ fi ve times will elucidate a potential root cause (hence this is sometimes called the ‘5 Whys analysis’). Th e results can also be displayed in a fi shbone diagram (named for its overall shape), shown in Figure 27.2, where the overall eff ect is listed at the ‘head’ of the fi sh. Th e large bones are used to classify possible cause categories, typically identifi ed as personnel, environment, policies, patients and materials and equipment. Each large bone can be divided into smaller bones representing sub-classifi cations of potential causes.

Other tools Understanding the processes and potential causes for problems is the fi rst step in quality improvement. Th e next steps in quality improvement use statistical tools to quantify, evaluate and monitor improvement and are not discussed here.

P art 4 Q

uality im provem

ent

28 Patient stories in improvement

Figure 28.2 An example of a patient story mind map

Tips for mind-mapping

1. Take a large, blank, landscape page of paper (i.e. turn the page so you can use the full width of the paper)

2. In the centre of the page place the name, identifying code or alias of your patient, thus representing the subject of your mind map

3. Use words and illustrations/drawings throughout your map. Wherever possible use single keywords in the early stages, circling them to ensure they are clearly visible. Each keyword should be written along its own line (known as a branch)

4. When a keyword is identified, make a note of the timing this occurred during the audio/video recording. This will assist you and your colleagues to re-listen to this section of the story again

5. Once you have completed listening and mind-mapping the story, listen to the whole story again, adding more details to the key words that were initially plotted

6. See if there are connections between words. If so, draw a line between them to make the association between emerging ideas as clear as possible

7. Experiment with different ways of linking and elaborating aspects of the patient’s story – use highlighters, codes and arrows

8. This process should be repeated until the map represents the totality of the story

Contact patient and provide introductory letter

Identify patient group

Agree time and date to meet patient

Obtain patient’s consent and record the story

Provide feedback to patient(s)

Themes from stories can inform an initial problem

statement for QI

Mind-map the story and share within your

team (advisable)

CT/MRI scan and uncertainties

No doctor seen for 2 days

Inappropriate placement on ‘geriatric’ ward

Patient moved from ward to ward

No records kept or available

Not alert to stroke symptoms

General lack of understanding of what was occuring

Patronising nurses

Nurses made things worse

Poor care = stress

Stress = sicker patient

Process issues

Staff issues

Consequences of poor care

Powerlessness

Alex takes own discharge

Alex

Figure 28.1 Summary of the process for collecting patient stories

91 C

hapter 28 P atient stories in im

provem ent

Introduction For centuries, stories have been told through the media of talk, writing or graphics. Stories are a valuable means of communi- cation between humans. However, with increasing value being attributed to objective scientifi c knowledge in recent decades, stories (which are by nature more subjective) have been somewhat overlooked as a resource for improving the quality of healthcare. Nevertheless, this situation is beginning to change. Th e UK NHS, for example, has recently taken steps to ensure the experiences of its patients, as well as those close to them such as family members, are listened to.

When patient stories are skillfully collected and utilised, they enable the health service to better understand what is working well and areas where patient and carer experience can be improved. As such, patient stories may become ‘mobilising narratives’ that help inspire healthcare teams to improve their service.

The process of collecting patient stories Th e process of collecting patient stories follows a series of steps (see Figure 28.1). Key steps are considered in this process: • Recruiting patients and obtaining consent • Ethically conducting and recording the story • Making sense of the story – mind mapping

Recruit patients and obtain consent It is important to contact potential storytellers who have recently experienced the service or patient journey you are aiming to improve. It may be tempting to only collect stories from patients who are easy to access and communicate with, while excluding patients with whom communication or access is more diffi cult. However, collecting stories from only one group of patients will not provide a fair picture of patient experience, and bias may be introduced into the project.

Stories should not be recorded unless storytellers are informed and provide valid written consent beforehand. All information that is necessary to storytellers to enable truly informed consent must be given in advance. Storytellers should be asked to sign a consent form, which will allow the use of the fi nal version of the story as a publicly available learning resource that aims to improve the qual- ity of healthcare.

Ethically conduct and record the story Th ere are three practical ethical behaviours that should be dis- played at all times when working with a patient story: • Respect: Storytellers and their stories should be treated with respect at all times. Th e intentions of the storyteller will be inter- preted accurately, thus preserving the integrity of the story • Support: Storytellers will be off ered emotional support during and aft er telling their stories • Confi dentiality and anonymity: Remind storytellers not to name other patients or staff members during the story. All

information collected from patients is to be kept in such a way that protects their identity unless they wish to be identifi ed, as many do.

Th e process of collecting the story involves good communi- cation skills that help the patient to feel at ease when recount- ing experiences. Establishing trust and rapport with the patient is important. For example, practising the opening introduction of the interview will help you confi dently explain to the patient why you are undertaking the project, in a manner that will reassure the patient but also establish trust and rapport. Now is also a good time to explain to the storyteller that stories are audio-recorded or (sometimes) video-recorded, thus ensuring that attention can be paid to listening rather than making notes. Recording also ensures accurate recall and allows the story to be listened to repeatedly with others in the team and beyond (n.b. ensure that patients consent to their stories being shared with others).

Following the introduction, the interaction moves to a series of questions about the patient’s experiences. Th e aim here is to allow the patient to speak as freely as possible within the time allocated. Th erefore, open questions should largely be used that will encourage patients to recollect and expand on their expe- riences. Closed or probing questions are justifi ed when specifi c information or clarifi cation is required. However, closed-type questions require caution and skill if the patient is not to feel rushed, or the conversation seems perfunctory rather than genu- inely participative.

It is generally unavoidable that, at certain times, you will need to add structure or direction to the interaction. However, in between these times, the patient will greatly value the experience of being listened to. Active listening is one of the most important ingre- dients for successfully collecting stories. During active listening, attentiveness is demonstrated to the speaker through a variety of actions, for example maintaining eye contact with the speaker or nodding at certain times to demonstrate understanding or that a certain point has been noted.

Making sense of the story – mind mapping To ensure that patient stories lead to improvements in healthcare, it is imperative that the key components of the story are dissemi- nated to colleagues. It is impractical to expect staff to read a whole patient story or even a lengthy précis of the story. Instead, the story should be summarised into the main learning or analytical themes, which are then used as themes and triggers for future ser- vice improvements.

A mind-mapping approach is oft en used as a means of trans- forming the totality of the story into a condensed graphical repre- sentation (see Figure 28.2 for a worked example). To make sense of the story, you should play back the recording of each story and ‘mind map’ the contents. Th e mind map should off er a means of reducing the totality of a story into main themes that can then be compared with the stories of other patients (see the ‘Tips for mind mapping’ box for more details).

P art 4 Q

uality im provem

ent

29 Leading change in healthcare Figure 29.1 25 reasons to change

Source: Adapted from a presentation slide given by patient safety leaders at WellStar Health System, USA

Medication error, fall

Wrong procedure

Fall Medication error

Deep vein thrombosis

Wrong site surgery

Burn

Delay in pain medication

Delay in diagnosis

Post procedural deathDelay in treatment

Medication error

Delay in pain medication

Wrong procedure

Fall

Medication error

Inappropriate touching

Deep vein thrombosis

Wrong site surgery

Burn

Delay in diagnosis

Post procedural death Delay in treatment

Ventilator associated pneumonia

Delay in pain medication

Figure 29.2 25 reasons not to change

It’s too ambitious

It’s too expensive

We didn’t budget for it

We can’t take the chance

It will take too long

They won’t fund it

It’s too radical

It’s too complicated

We have too many layers

It’s not our problem

Another department tried that

It needs more thought

It’s too political

It’s against tradition

There’s not enough time

They don’t really want to change

It needs a committee decision

This is just a fad

There’s no clear mandate

I’m all for it, but ...

It can’t be done

It won’t fly

We’ve always done it this way

It’s impossible!

We’re doing OK as it is

Source: Adapted from Bumsted P. Biocultural Science and Management; http://13c4.wordpress.com/2007/02/24/50-reasons-not-to-change/

93 C

hapter 29 Leading change in healthcare Opposing the status quo Leading improvements in healthcare can be challenging. However, once you start, you are likely to spend the rest of your career seek- ing to improve the systems that lead to poor patient experience and harm. Becoming a ‘radical’ against existing systems is not just about speaking up when bad things happen (which is very impor- tant), but also about taking action in your own area of infl uence to make high-quality care the norm. A radical is a courageous, pas- sionate person who is willing to challenge the status quo and take responsibility for change.

Figure 29.1 provides a powerful visual of the consequences of poor-quality healthcare and can be the catalyst for involving oth- ers in your ideas for change. Looking at the history of change and improvement, big change only happens in organisations because of radicals. However, there are many forces opposing the changes you will want to see. For example: systems that reward their workers for ‘keeping the trains running’ rather than seeking ways to improve care; people in powerful positions who have a vested interest in keeping the status quo; and colleagues and leaders who are scepti- cal, apathetic or scared of change (Figure 29.2).

So who are the most eff ective radicals in healthcare organ- isations? Research by organisational behaviour expert Debra Meyerson suggests that the people who get the best outcomes from change are those who have learnt to oppose and conform at the same time. Or, as she puts it, ‘Th ey are able to rock the boat and yet stay in it’. Th ese are the champions for change who stand up to challenge the existing order of things when they see there could be a better way. Th ey develop the ability to walk the fi ne line between diff erence and fi t, inside and outside. Th ese lead- ers are driven by their own convictions and values, which makes them credible and authentic to others in their organisations. Most importantly of all, they take action as individuals who, by involving others, ignite the kind of broader collective action that leads to big change.

Characteristics of radicals Radicals already exist in every healthcare organisation, in many diff erent roles and at many diff erent levels. Th ey are not typically the chief executives or senior clinical leaders, yet the impact of their change activities is oft en just as signifi cant.

Successful leaders oft en invest time and eff ort building eff ec- tive networks and relationships with others; continuously seek innovative new ways of delivering care and reducing harm; and commit to the patient-centred mission and values of the health- care organisation they work for. Th ey are driven by a passion for better, safer care for patients; are optimistic about the future and the potential for change; see many possibilities for doing things in diff erent ways; and generate energy for change which attracts oth- ers to collaborate with them for the shared common cause of better quality healthcare.

‘Troublemakers’ also challenge the way things stand but in a diff erent manner to that of radicals. Troublemakers complain about the current state of aff airs, but their focus tends to be around their own personal position rather than achieving the goals of the organisation. Troublemakers are angry about how things are and do not have much confi dence that things will get better in the future. Th ey alienate other people because if others link with them, troublemakers will sap their energy.

Figure 29.3 contrasts the characteristics of radicals and trou- blemakers.

Surviving as a radical Th ere are a few challenges for improvement leaders in this radical– troublemaker distinction. Firstly, many organisational leaders view anyone who challenges the status quo as a troublemaker. Th ere- fore, radicals can get unfairly labelled as troublemakers when they challenge existing systems and processes that create the potential for poor-quality care. Sometimes, this has the impact of making potential radicals keep quiet or conform when the most appropri- ate action would be to take action or speak up. Secondly, lots of people who care deeply about high-quality healthcare start out as radicals, but their voice does not get heard or they are ignored or ridiculed. As a result, they can begin to noisily question the status quo in an antagonistic and self-defeating manner, and eventually cross the line from radical to troublemaker.

History shows us that troublemaker tactics are hardly ever likely to get the results sought. As an improvement leader, you have a responsibility to watch out for this in your own behaviours and try to prevent it from happening by building relationships and forming alliances with others who are also willing to stand up and be counted for better patient care.

1. Make time to reflect on your own role as a healthcare improvement leader; what are the implications for acting as an agent or leader of change?

2. Seek out other radicals who share your mission and discuss tactics for rocking the boat and staying in it

3. Identify and support others who are at risk of crossing the line from ‘rebel’ to ‘troublemaker’

4. Think about how you adopt or build the ‘spirit of the student’ and how your role as an active learner can be a catalyst for others

5. Reflect on the extent to which you are knowing, doing, living and being healthcare improvement and patient safety; to what extent are you operating from your true self? How can you make your impact as a healthcare radical even more effective?

1. Big change only happens in healthcare organisations because of the courageous people (‘radicals, rebels, dissenters and heretics’)

who show there can be a different and better future 2. Being a healthcare improvement leader inevitably means challenging the status quo of how systems and processes of patient care are organised

3. Walk the fine line between difference and fit, inside and outside, conformity and rebellion; those who can rock the boat and stay in it are most successful

4. Change yourself first, rather than just expecting other people to change

5. Adopt ‘the spirit of the student’ – be open to continuous learning, embrace new ideas and approaches and be willing to challenge and change your own beliefs

6. It is important (but insufficient) to build knowledge and skills in improvement; to be truly effective as a leader, you have to be a role model for everyone else

Box 29.2 Key messages for leading change

Box 29.1 Some calls to action for improvement leaders leaders as healthcare radicals

Troublemaker Radical

Complain

Me focused

Anger

Pessimist

Energy sapping

Alienate

Problems

Alone

Create

Mission focused

Passion

Optimist

Energy generating

Attract

Possibilities

Together

Figure 29.4 Being a great change agent is about knowing, doing, living and being improvement

Knowing it Doing it Living and

being it

Use of self

D iff

ic ul

ty –

c om

pl ex

ity

Time

Knowledge

Skills

Figure 29.3 Comparing the characteristics of radicals and troublemakers

P art 4 Q

uality im provem

ent

95 C

hapter 29 Leading change in healthcare How to be a radical: start close to home fi rst Learning the methods and tools for delivering high-quality patient care will give you a new set of eyes for looking at the healthcare system. Whatever clinical setting you work in, you will be able to look at the world around you and see things that need improving; the medication process that should be reconciled, the preopera- tive system that creates the risk of infection or the labelling system which might lead to a drug error. However, if you are going to be a truly eff ective improvement leader, starting your own self- improvement work at an even earlier point is essential. Do not be tempted to launch into big eff orts to infl uence other people to change the way they currently think and work, before refl ecting on and changing your own practices fi rst. Th is is well understood by David Whyte when he says, ‘I do not think you can really deal with change without a person asking real questions about who they are and how they belong in the world’.

Improvement leaders are signal generators, and their words and deeds are constantly scrutinised and interpreted by the people around them in diff erent teams and departments. Th e amplifi ca- tion eff ect of what was done and said is far greater than we can imagine. Th us, perhaps the most powerful way to inspire others to change is to be the role model for that change. If you want other people to take a risk and change the way they think or carry out their roles for healthcare improvement, you have to take the lead. As Tanveer Naseer describes it:

You have to be the fi rst one up and off the high dive you’re asking others to leap from. Ask yourself: where am I playing it too safe and what is that safety costing me? Th en leap from your platform of safety into the cold water of change.

By taking action, the most eff ective radicals are able to achieve small wins that create bigger ripples in the organisation, building a sense of hope and confi dence that it is possible to make a diff er- ence. Th ey can then join forces with others who are also commit- ted to the cause, working as a collective force for changes that are commonly valued.

So, as an improvement leader, you should create your own goals for change, work out how and where to make a contribution to the bigger cause, seek and reach out to build alliances with oth- ers and demonstrate the will to move the change agenda forward. When you have the courage to act proactively like this, most line managers and leaders will value these behaviours, even where the organisation does not currently have a strong quality improvement culture.

Generating signals for success Fear is a major inhibitor of improvement leadership. People can be afraid to challenge or report poor-quality care because of fear of repercussion or reprisal. Fear is also a signifi cant barrier to learning – it is hard to learn when you feel fear. An improvement culture requires people to move away from a status quo that they feel comfortable with and into a brave new world of quality control (standardisation of procedures based on evidence-based practice), quality improve- ment (thinking about healthcare in terms of the six aims of quality and more) and quality planning (realising the steps needed to trans- late the quality improvement policy agenda into workable actions in the workplace), and that can be scary. As Peter Senge wrote in Th e Fift h Discipline:

When we see that to learn we must be willing to look foolish, to let another teach us, learning doesn’t always look so good anymore.… Only with the support and fellowship of another can we face the dangers of learning meaningful things.

Th is situation creates a specifi c call to action for healthcare radicals. As a signal generator at the leading edge of change, you must embrace the spirit of the student. Th is means taking respon- sibility for your own learning and being open to continuous learn- ing, embracing new ideas and approaches and being willing to challenge and change your existing belief systems. Your learning must move beyond knowledge and skills. For radicals, it is impor- tant, but not enough, to continuously build your knowledge of improvement methods and approaches. It is also important, but not enough, to take responsibility for your own development as a skilled leader or a facilitator of change. What sets healthcare radi- cals apart is the extent to which they purposefully seek to ‘live’ and ‘be’ improvement in the way they operate in the world and in their interactions and relationships with others. Figure 29.4 captures the relationship of those competencies.

History also teaches that organisational or system transfor- mation is always preceded by personal transformation. So if, as a healthcare radical, you want to play your role in radical changes for patient safety, you should focus deeply on your own perspective and the ways you interact with and infl uence others. Chapter 30 explores the role of narrative in public leadership; consider how you can use storytelling to radically involve others in your vision for change.

Th e more people you can infl uence in a positive way, the more that you and your followers (as organisational radicals) can unleash a powerful reservoir of energy for change, and the more diff erence you can collectively make for your patients.

P art 4 Q

uality im provem

ent

30 Public narrative: story of self, us and now Figure 30.1 Story of self, story of now and story of us

Source: Adapted with permission from Professor Marshall Ganz, Harvard University, Boston, MA, USA

Challenge

• What was the specific challenge you faced?

Choice

• What was the specific choice you made?

• Why?

• How did it make you feel?

Outcome

• What happened?

• What hope does it give us?

• Whay do you want to teach us?

• How do you want us to feel?

Purpose Story of self

Call to leadership

Story of now

Strategy and action

Story of us

Shared values and shared experience

Community Urgency

Introduction Stories can be powerful tools to inspire others to work with you to change existing practices. Social organisers – those who rally people towards shared common goals – use ‘public narrative’ as a method for cultivating the hearts and minds of those who can help to bring about change. Having a ‘public narrative’ to encourage others to work with you can be useful for building a team and gain- ing momentum to achieve your vision for change.

Public narrative has three elements: a story of why you person- ally believe a change is needed now – a story of self; a story of why you think those listening might believe a change is needed too – a story of us; and a story of what makes the challenge ahead so urgent to act upon – a story of now (see how the three elements connect in Figure 30.1).

Story of self Telling a ‘story of self ’ is a way to share the values that defi ne who you are based on your lived experiences. Th is can be made up of choice points, which are moments when you have faced a chal- lenge (e.g. the decision to speak out when you’ve seen a prob- lem in care), made a choice (e.g. the act of speaking up), experi- enced a positive or negative outcome (e.g. how those around you reacted and what they said) or learned a moral (e.g. speaking up for a patient when everyone else was ignoring her). Th ink about challenge, choice and outcome in your own story. Th e outcome might be what you learned in addition to what happened. Pow- erful stories leave your listeners with images in their minds that

shape their understanding of you and your calling. Articulating the decisions you make in the face of challenges ultimately com- municates your values.

Elizabeth is a 79-year-old woman who was admitted for investigations to try to explain her chest pain. Before the weekend, she was full of life and had a smile from ear to ear on the ward round. Today, she wasn’t smiling. She was wincing in pain and appeared confused. She had fallen on Saturday night by tripping over her telemetry leads. I was distraught by the fact we couldn’t ensure the safety of a patient who was in good health before her admission and now has a broken hip and is at a higher risk of mortality. Our team missed the opportunity to discontinue her telemetry before the weekend.

Story of Us Th e purpose of the ‘story of us’ is to bring alive the values your audience shares with each other that can inspire collective action. Your goal is to tell a story that (1) evokes shared values that can unite the group, (2) highlight the challenge(s) faced that makes action urgent, and (3) give hope that, if you work together a specifi c change is possible. If your team is developing a quality improvement intervention, then you must think of it in that con- text. What are the shared values of your team and those you are trying to engage?

Th e ‘story of us’ draws upon the values and shared experiences of the group. It becomes a moral resource (or a reminder of why change was needed) and can be called upon as the group moves

97 C

hapter 30 P ublic narrative: story of self, us and now

forwards to face uncertainty. Stories of us can inspire, teach, off er hope and advise caution. Th ey can shift power relationships (‘We’re all in this together’) to build new communities of passionate peo- ple committed to making change happen.

Our team had a meeting about how this could be avoided. I remember for the last 6 months, nurses have been asking us during the middle of the day, ‘Can we remove the urinary catheter?’ and ‘Can we remove the telemetry?’ As a junior doctor, we addressed the issue at that time or deferred it for a later time. In the meantime, we increase the risk of adverse events if the patient doesn’t need either. We had weekly conversations as a result that asked questions of how we could minimise errors, avoid interruptions, improve quality and decrease costs as they relate to telemetry and urinary catheters. Together, we created a review system that we would run with each patient at the conclusion of each encounter; we would ask about the need for telemetry, catheters, in fact any unnecessary attachment or intrusion on the patient. If they didn’t need them, we would discontinue them and the nurses would remove them. Using the Model for Improvement, we learned how to adapt the concept with other medical teams and eventually spread this across all the medical teams in the hospital.

Story of Now Th e story of now must articulate an urgent challenge that the com- munity (now known as the ‘us’) shares. Here, story and strategy overlap because a key element in hope is a theory of change – a

credible vision of how to get from A to Z. A meaningful choice is more like ‘We all must choose – do we commit to improving care until catheter-related infections are zero or not?’ A vision can begin by getting a number of people to show up at a meeting that you committed to run. You can achieve a ‘small’ victory that shows change is possible. A small victory can become a source of hope if it is interpreted as part of a greater vision.

In a public hospital where we are strapped for resources and inspired to provide the best care possible for our patients, simple interventions make a diff erence. Six months later, the rate of inpatient falls is down by 20%, the infection rate is down 5% and we have saved the healthcare system thousands of pounds. Th ese results got the attention of other junior doctors on the sidelines. As we continued to tell our story, others joined. Our eff orts moved from telemetry and catheters to cancelled blood results, health education, improved medicines reconciliation and redesigning surgical pre-assessment clinics. We were becoming a beacon of hope for high-quality patient-centred care.

Conclusion Th e narratives are linked. Th e now is linked to self through the pur- pose for which the authority will call upon others to join in action. Th e story of now creates urgency rooted in the values shared by the us who one hopes to move. Th e goal is not to have a fi nal ‘script.’ Th e goal is to learn a process by which you can generate your narrative, and call upon it, when, where and how you need to in order to motivate yourself and others to purposeful action.

P art 4 Q

uality im provem

ent

31 Planning an improvement project Figure 31.1 An overview of planning an improvement project

You get interested in quality improvement

Where can I learn more?

Pursue QI knowledge through on- line courses (e.g. IHI Open School),

connect with faculty, research

How can I find a project?

You see the problem first hand shadowing in the ED. Asthma advice leaflets are only made

in English and you see a Spanish speaking family return to the ED

several times for treatment because they didn’t understand

the information leaflet

Where can I get some guidance?

Your classmates suggest you speak with the QI director

for the ED

You decide that other students would be a great resource to

help with the legwork. It is especially valuable you have different perspectives, so you contact students in nursing, medicine, pharmacy, social

work, etc

Who should be a part of your improvement team?

You meet with the QI director, who is excited about the

project. She agrees to become your faculty advisor. Her first suggestion is to put together

a team

Your team meets with your faculty advisor, who helps you

identify the key stakeholders: the nurses, doctors, and

patients that will be affected by the project

Your faculty advisor helps you lead an introductory meeting

with the team OK, you have a team. Now what?

You invite doctors, nurses, the nurse manager, and several

motivated patient families to be a part of the team. Even if

they don’t all participate regularly, they’ll appreciate

being kept in the loop

Can I see it for myself?

You speak to a nurse manager in your hospital’s emergency department (ED). She tells

you that they need help with the number of children with asthma returning to the ED

after treatment

You Other students QI director/faculty advisor ED staff (technicians, nurses, doctors, attendings, etc)

Planning an improvement project

Together, you develop an AIM: decrease the number of patients who are returning to the ED within 72 hours for persistent asthma symptoms by 20% by June1, 2013

The team also comes up with some measures: The percentage of patients who represent within 72 hours, the percentage of patients filling medications after discharge, the percentage of patients who receive an asthma information leaflet in their primary language, and patient satisfaction with the process. Given seasonality of asthma, numbers might go up

and down independent of care delivered but simply because of variations in census

The team also comes up with two initial changes to the system: translating the asthma information leaflet into several primary languages of patients frequenting the ED and setting up a system for follow-up phone calls two days after discharge

Congratulations! You’ve taken the first steps towards improvement by planning an

improvement project. With all of these pieces in place, you and your team are ready to start testing changes and, hopefully, improving the quality

of care at the hospital

99 C

hapter 31 P lanning an im

provem ent project

Introduction Th ere are seven important steps to getting started on a successful quality improvement project (see Figure 31.1): 1 Learn about quality improvement. 2 Find a project. 3 Form a team. 4 Find a faculty advisor. 5 Identify the key stakeholders. 6 Develop an aim and measures. 7 Identify changes.

Learn about quality improvement See Chapters 21–27.

Find a project Th ere are several ways you could identify a quality improvement project, including: 1 Drawing on experiences as a student (or even if you’ve spent time as a patient); you might have observed things that did not work as well as they should have. Did patients have to wait for a long time before they are seen in a clinic? Was there a lack of com- munication between physicians and nurses at the bedside? 2 Talk to people who work in healthcare organisations. Front- line staff who take care of patients every day know which systems work well and which need improving. Ask them what they think is wrong. Or ask them what one thing they would improve if they could. 3 Routinely ask patients during your interaction, “What can I do to improve your stay?” Ask enough patients and it is likely that potential themes for improvement will emerge. 4 Identify and talk with those in charge of care quality in specifi c clinical areas – directors of quality, nurse managers or clinical directors. Staff members in these leadership positions oft en man- age a number of ongoing projects that strive to improve care for patients. You could join one of these projects and learn from experienced professionals about how to conduct quality improve- ment projects.

It is important to start small when beginning a quality improve- ment project. Do not try to improve the infection rates in the entire healthcare organisation. Start in one hospital, on one ward, at one bedside, with one patient. Use the Model for Improvement to determine what works, and then scale up.

Form a team Collaboration is essential. Building a diverse group of team members working towards a shared common goal is integral to success. In quality improvement, teams do not comprise just one type of professional – reach out to other disciplines, including medicine, nursing, pharmacy, public health and even engineer- ing. By having multiple disciplines, you can get a sense of how others view the healthcare system. And, through this broader perspective, you will be able to fi nd bigger problems and better solutions.

Find a faculty advisor It is critical to have someone who can advocate for you and your improvement team. Th is person may be the leader of a specifi c

clinical team or a member of faculty at your university. He or she will provide active coaching throughout the process and facilitate connections with your local healthcare setting.

Identify the key stakeholders Who is going to aff ect your project? And who is going to be aff ected by your project? Will you need to work with a certain physician? All the nurse educators? A few members of the pharmacy depart- ment? It is important to fi gure out this information as soon as pos- sible and then contact the appropriate stakeholders early and oft en as you begin a project.

It might help to outline the stakeholders in your project using a simple 2×2 matrix (see Table 31.1).

Table 31.1 2×2 matrix for stakeholder analysis

Low interest and involvement in process

High interest and involvement in process

High power and infl uence

Satisfy: Keep this group posted on important developments, and make sure they are satisfi ed at the end of the project.

Engage: Keep this group fully engaged from the start. Th ey will be instrumental in making your project a success.

Low power and infl uence

Monitor: Keep an eye on this group, but realise that your time and resources might limit your interactions with it.

Inform: Keep this group well informed throughout the process, as they will be aff ected just as much as the group above.

Develop an aim and measures Two mistakes that are oft en made when trying to improve a care process are starting a project without a goal and/or a measurement plan. A goal (or, in quality improvement terminology, an aim) tells you where you are going. It answers the fi rst question in the Model for Improvement: ‘What are we trying to accomplish?’ An aim statement needs to be specifi c, measureable and actionable. Here is a bad aim statement: ‘Our team wants to reduce medication errors’. Why is it bad? It is not specifi c and does not say how much, by when or for whom. Instead, say: ‘Our team aims to decrease the number of medication errors by 20% for patients in the intensive care unit by 5 December 2015’.

Similarly, without an eff ective measurement plan (refer back to Chapter 24 for more detail), you will not know if your interventions are leading to improvement. Th is answers the second question in the Model for Improvement: ‘How will we know that a change is an improvement?’ Typically, improvement projects include outcome measures, process measures and balancing measures.

Identify changes Th e last step before you start improving is coming up with some changes to test. In this stage, you will answer the question ‘What changes can we make that will result in improvement?’ For exam- ple, let us say you want to improve the percentage of patients who take their medications aft er a consultation. One change you might try is setting up a system to call them 2 days aft er they receive the prescription and remind them to take their medication.

100

P art 4 Q

uality im provem

ent

32 Managing an improvement project Figure 32.1 Initiating and executing a project plan

How can it be fixed/prioritised? • Consult the literature for best practices, guidelines, defined measures, etc

How will it be measured? • What will improvement look like?

How will the plan be implemented? • Educate and communicate with everyone involved/affected

How will changes be recognised? • Regular communication and status reports

How will aims be met? • Improvement based on measures identified in 3

Map out a plan

Execute the plan

Monitor progress

Evaluate

Identify possible causes

Tool: • Fishbone diagram

Tools: • Gantt chart

• Process flowchart

Tool: • Driver diagram

Define aim/purpose of project

Organise the team

Fix a problem

Identify strategy

Organisational priority

Key questions to ask:

101 C

hapter 32 M anaging an im

provem ent project

Introduction – developing strategies In an industry where costs are growing and resources are con- strained, good stewardship of healthcare is imperative. At the centre of activities to achieve better value for money are well- thought-out improvement projects that are successfully managed. Project management is the ‘application of knowledge, skills, tools and techniques to project activities to meet the project require- ments’, oft en when the solutions to a problem are known. It is a strategy used in most industries. Managing healthcare improve- ment projects does not necessarily require knowledge and skills unique to healthcare, but the tools and techniques may diff er from those used for project management in other fi elds. Improvement tools can be used when solutions are unknown and need to be developed or tested, for developing strategy and for monitoring progress towards the aim(s) of your project. Here, we consider how to merge best practices from project management (i.e. the role of the project manager) with the benefi ts of using improvement methods, to keep your improvement project on track.

Characteristics of a project In healthcare, improvement projects are planned and evaluated using short cycles. Each cycle has a beginning and an end, it is based on a defi ned time period of learning, and the end point is determined by specifi c criteria related to an anticipated result. Pro- jects develop as a result of a need to fi x a problem or an organi- sational priority (Figure 32.1). Planning the amount of resources that will be required to complete a project is essential, and a com- parison of costs and benefi ts helps project managers evaluate the feasibility of seeing the project through to completion.

A project typically develops out of a recognised need, whether it is to fi x or to improve, although the approach is essentially the same (Figure 32.1, Steps 1 and 2). Th e key diff erence is that projects related to a problem must fi rst investigate potential causes of the problem. A number of tools have been developed to work through this process (e.g. a fi shbone diagram, as discussed in Chapter 27) and are described in the context of project planning. When a pro- ject develops out of an organisational priority, on the other hand, the focus turns towards information gathering – what have others done, what does the literature indicate as best practices and how do guidelines and performance measures help guide organisations focused on the same priority?

Once the project strategy is decided, a plan can be mapped out to clarify who will be responsible for each aspect of the project. Important considerations include what will be done, by whom, by when and how it will be measured to evaluate progress towards the aim (Figure 32.1, Step 3; the plan–do–study–act (PDSA) form in Chapter 23 and the ‘Seven steps to measurement’ in Chapter 24 are used here). Next, the project team implements the plan, giving consideration to how the people involved will be impacted (Step 4). For example, if the aim was to reduce readmissions and the team is implementing a strategy to follow up with patients within 24 hours of discharge, the hospital providers, discharge planners, cardiologists, general practitioners and patient, and any providers involved in the transition of that patient, would need to be aware of the new process and their role. Once the strategy is underway, it is critical to monitor it regularly (Step 5); a run chart (or Shewhart chart), as described in Chapter 24, can serve as a visual display and update. A chart should exist for each project- related measure.

Finally, progress towards the aim(s) (Step 1 using the Model for Improvement) must be evaluated. Improvement is continuous.

Step 6 helps determine what is working and refl ecting on the team’s belief in the changes and resulting improvements could be useful. Modifi cations could be needed to the next PDSA cycles. Steps 4–6 are your regular opportunities to take a bird’s-eye view of your project to take stock of what has been achieved and what remains based on the changes you have developed and tested via PDSA cycles to date.

The role of the project manager Th e project manager is responsible for organising, managing and holding accountable the involved parties. Successful management of an improvement project is related to the participants having a sense of ownership of the process that is developed and tested, the outcomes that result, and how promising change models are eventually implemented in practice and sustained. Management is not, however, the same as leadership when it comes to working with a team towards a successful outcome. ‘Leader managers’ set themselves apart from other managers; whilst their projects could be based in short time periods, leader managers think over the long term and ask: ‘How will this project impact the future of the organisation, and will this project be the stepping-stone for other initiatives that have the potential to impact the organisation in the long term?’ Such projects are also planned and executed with the interests of the organisation in mind, not simply a single aspect of the organisation without consideration given to others. And when organisational politics begin to challenge the project’s impact, leader managers are equipped to manage the interests of multiple constituents. To that end, there is an emphasis placed on the vision and values of the organisation on the part of leader managers, and they by no means accept the status quo. Leader managers are focused on the aims of the project and leading a team towards suc- cessful outcomes.

Tools A number of tools developed for project management and others focused on healthcare quality improvement are outlined in Table 32.1. Each of these tools serves a diff erent purpose and can be useful at specifi c points in the project, from working to identify the cause of a problem, outlining the key steps in your project or mapping a timeline to ensure your project is completed in a timely fashion. When working to fi x a problem, tools such as the fi shbone diagram could be helpful to determine potential causes. If you were trying to improve a complex process like diabetic manage- ment, a process fl owchart could be useful. To progress the learning from ideas elicited by those tools, a change model like the Model for Improvement will next help to keep track and make sense of the impact that the ideas and concepts (which you have identi- fi ed from searching the literature or borrowing from elsewhere, or are developing and testing de novo) are having on achieving your goals.

A driver diagram is most useful when you have a specifi c aim in mind; that is, you know what you want to achieve (Figure 32.2). Let us continue with the example of diabetic management with the aim of HbA1c test results being available at the time of a patient visit. We need to explore the factors that may infl uence completing the test before an appointment, such as the doctor or nurse order- ing the test, patients knowing that the test is needed and taking the initiative to have the test done, and the results being communi- cated to the healthcare practitioner who is managing the patient’s diabetes. Th ese are considered primary drivers which are either achieved together or addressed as targets towards achieving the

aim. Th e secondary drivers impact the primary drivers and may include ongoing monitoring of blood test results, access to a phle- botomy service and a process for sending results to practitioners. Each of the ‘drivers’ outlined as contributing towards the aim should be measurable so that the diagram provides a complete picture of a strategy for achieving an aim and how to measure progress.

If you are beginning with a problem or an ‘eff ect’ and your aim is to better understand the ‘causes’ to understand how to fi x it, you could use a fi shbone diagram. If the problem was incomplete dia- betic A1c tests, you could think through the potential factors that are contributing to the tests not being completed. To categorise the potential causes, consider the methods, the people and the systems involved in completing an HbA1c test. Th ere is no standard for categorising the potential causes, but some categories to consider include the 4 M’s – methods, machines, materials and manpower; the 4 P’s – place, procedure, people and policies; and the 4 S’s – surroundings, suppliers, systems and skills. Some of the ‘meth- ods’ we could explore relate to test ordering, patient awareness

of the need for an A1c test, and communication of test results (Figure 32.3). Th e ‘people’ involved in the process may include practitioners, administrative staff , lab personnel and the patients. Th e ‘systems’ involved would be related to communication between the practitioners and the patient, the patient and the lab where the

Table 32.1 Project management and improvement tools

Tool Purpose Elements Product Driver diagram

When: Step 2

Visual representation of an aim and the factors and activities that impact the aim

Aim: Primary factors – need to be infl uenced to achieve the aim Secondary factors – activities that impact the primary factors

Th eory of change

Fishbone diagram

When: Step 2

Outline of the factors that infl uence an outcome

Branches may include people, processes, policy, methods, materials and environmental factors

Hypothesis of causes of a problem

Gantt chart

When: Step 3

Illustration of a schedule for project completion

Project milestones and elements, time and status

Progress made towards an aim

Process fl owchart

When: Step 3

Visual representation of a process and the people involved

Symbols used to depict start and end, steps in the process and decision points

Pictorial representation of a process

Plan–do– study–act (PDSA)

When: Steps 2–6

Model for Improvement intended to test change by planning it, trying it, observing the results and acting on what is learned

Plan – the test or observation, including a plan for collecting data Do – try out the test on a small scale Study – set aside time to analyse the data and study the results Act – refi ne the change, based on what was learned from the test

Knowledge about whether a strategy led to improvement

Figure 32.3 Putting a fi shbone diagram to use

Problem

Materials Machines

Measurements Environment People

Methods

Time

Purpose • Explore possible causes of a problem

Characteristics • Potential causes to consider – people, environment, materials, methods and equipment

It looks like:

Purpose • Guide implementation of a change

Characteristics • Three or more levels that outline the 1. Aim 2. Factors that influence the aim 3. Activities that act on factors

It looks like:

Outcome desired

Primary driver of change A

Secondary driver A1

Secondary driver A2

Primary driver of change B

Primary driver of change C

Secondary driver C1

Secondary driver C2

Secondary driver C3

Secondary driver B1

Figure 32.2 Putting a driver diagram to use

102

P art 4 Q

uality im provem

ent

103 C

hapter 32 M anaging an im

provem ent project

A1c test is completed, and the lab and the practitioner. As you are identifying each of these factors, consider how each is contributing to the problem and why it is happening. Items that appear in more than one category can be considered most likely causes and can be further discussed and prioritised as contributing factors. Th e prod- uct of a fi shbone diagram is a list of potential causes that can then be used to identify an improvement strategy.

A process fl owchart can be useful at various points in your exploration of a problem or a solution, or in the management of a project through to completion. Figure 32.1 is a process fl ow- chart that outlines how to execute a project plan. Th is tool out- lines activities and sometimes decision points along a project cycle. For example, if we mapped out a plan for diabetes management, completing A1c tests before a practitioner visit would be one of the steps explored along the process from ordering the test to the practitioner receiving the results. Th is visual representation of a process can help identify each step and how to proceed at various decision points. Th e result is a mapped-out process and a plan for achieving your project aim(s), which help to maintain focus along the project cycle. See Chapters 26 and 36 for examples.

A Gantt chart can be used to monitor progress towards project aims. Th is tool, similar to a timeline, outlines project deliverables on the y-axis and time on the x-axis. A Gantt chart is useful for monitoring deliverables and communicating the status of a project. It complements a process fl owchart when time is associated with each project element represented in the fl owchart (Figure 32.4).

Both tools present project managers with a framework for organising elements of a project and a mechanism for communi- cating with the individuals engaged in the project.

Conclusion Healthcare improvements must be achieved by leader managers – those dedicated to the organisation and its long-term success. Becoming a leader manager begins with recognising that, as a health

professional, you have two jobs: job 1 is doing the work, which is providing healthcare; and job 2 is improving it. Armed with the knowledge of how to manage a project and the tools presented in this chapter, you can become a leader manager and be successful in job 2.

Purpose • Outlines project deliverables and assigns a timeframe to each

Characteristics • Project deliverables and time

It looks like:

Deliverables

Deliverable 1

Deliverable 2

Deliverable 3

Deliverable 4

Deliverable 5

Deliverable 6

Deliverable 7

Deliverable 8

Deliverable 9

Deliverable 10

Target date Week 1 Week 2 Week 3 Week 4 Week 5 Week 6

Figure 32.4 Putting a Gantt chart to use

104

P art 4 Q

uality im provem

ent

33 Quality improvement in psychiatry Figure 33.1 Plan–do–study–act (PDSA) cycles for increasing use of the Alcohol Use Disorders Identifi cation Test – Consumption (AUDIT-C)

AUDIT-C questions

1. How often do you have a drink containing alcohol?

2. How many drinks containing alcohol do you have on a typical day when you are drinking?

3. How often do you have six or more drinks on one occasion?

Concept 1 PDSA cycles (see graph 1)

Concept 3 PDSA cycles (see graph 3)

Concept 2 PDSA cycles (see graph 2)

Figure 33.2 Root cause analysis on low rates of alcohol screening

Urgent issues

Too disorganised

Alcoholic

Don’t want to push

Will ask later

Psycodynamic focus

Packet not sent

Packet not completed

Packet not transcribed

Patient factors Alliance Packet

Current approach OK

Unaware of AUDIT

Unsure of scoring

What if positive???

Not part of routine

Forget to ask

Forget questions

Too busy

Afraid to document

Forget to document

Not in LMR template

Assessment DocumentationEducation

100

Mean

UCL

LCL

PDSA Ramp 1 led to a rapid

increase in use

PDSA Ramp 2 took it to the

next level

PDSA Ramp 3 helped keep useage high as the project

expanded clinic-wide

P er

ce nt

ag e

Feb 2008 PDSAs for concept 1

May 2008 PDSAs for concept 2

June 2008 PDSAs for concept 3

Figure 33.3 Control charts for monitoring improvement in the rate of Alcohol Use Disorders Identifi cation Test – Consumption (AUDIT-C) utilisation

105 C

hapter 33 Q uality im

provem ent in psychiatry

Introduction Th e nature of psychiatric disorders and their treatment present a unique set of challenges for quality improvement. Stigma about mental illness, a high degree of inter-provider variability in the approach to diagnosis and treatment, and a relative lack of objec- tive tools for screening, diagnosis and outcome measurement are common challenges. Th e World Health Organization (WHO) esti- mates that harmful use of alcohol accounts for approximately 3.8% of all deaths (2.3 million) around the world, and studies show the mortality rate is increased threefold when alcohol use disorders are comorbid with mental illness. Despite these data, the use of validated alcohol use screening tools remains low in the majority of outpatient psychiatric clinics.

Th e aim in this example is to increase the use of a scientifi cally validated tool to screen for alcohol use disorder in a general psy- chiatry outpatient clinic. Th e goal was to increase use of the tool during new intake evaluations from 4% to 50%. Let’s look at how the Model for Improvement was used to achieve this goal.

Strategy for change Th ree key concepts that resulted in the biggest changes during the improvement project are outlined; the plan–do–study–act (PDSA) cycles that developed and tested the change ideas are summarised. Th e three concepts informed secondary drivers in a larger driver diagram. Concept 1: Provide training and support for tool use – Education and ‘cheat sheets’ will increase the rate of use of a scientifi cally vali- dated screening tool for alcohol use disorders. Summary of plans made in PDSA cycles: 1 Create a team comprising department leadership, substance abuse specialists and frontline clinicians from the outpatient men- tal health clinic. 2 Use a fi shbone template to identify potential causes of low rates of alcohol screening (see Figure 33.2). 3 Establish a baseline from previous patient notes in the past 3 months (see the control chart in Figure 33.3). 4 Expand fi shbone categories through team discussion to include specifi c elements felt by the team to contribute to low rates of screening. 5 Identify a pilot group of mental health clinicians that agree to pilot a simple-to-use and scientifi cally valid screening tool (the Alcohol Use Disorders Identifi cation Test – Consumption (AUDIT-C)). An educational session was convened; clinicians were provided with a ‘cheat sheet’ (a prompt with key details) to guide use of the tool during the initial assessment. Each clinician asked to trial the tool in their practice and provided feedback at a follow-up educational session. Summary of learning from collected data and feedback from PDSAs run for Concept 1: Only 4% of charts showed evidence of use of a standardised screening instrument. In one month, the rate of screening usage increased to 40%, but the improvement

plateaued, and towards the end of the month there was evidence of decreased usage. Th is prompted the team to determine why; the issues identifi ed in the fi shbone diagram combined with clinician feedback guided further tests in subsequent PDSA cycles. Concept 2: Regular feedback to clinicians – Implementing an audit and feedback process to understand barriers to embedding the intervention will help sustain our gains. Summary of plans made in PDSA cycles: 6 Schedule audit and feedback sessions with the pilot clinicians to explore why some clinicians were exhibiting more consistent uptake of the screening tool, while others appeared to lag behind, and to better understand their individual barriers. Th e sessions would reiterate the importance of tool use, present data on the uptake by other colleagues, and use the fi shbone diagram to guide discussion to help recognise the multiple factors that may be pro- hibiting tool use. Summary of learning from collected data and feedback from PDSAs run for Concept 2: One-to-one feedback and discussion with lagging clinicians comprised a more powerful method for improving uptake than group feedback sessions. Control charts were a consistent source of guidance to judge the impact of our feedback eff orts and identify potential clinicians to target. Use increased from 40% to 90% in 4 months. New mean and upper control and lower control limits were created for the project. Consistent pushback from clinicians about fi nancial compensation for completing these tools with patients was noted. Concept 3: Financial incentives – Including the use of incentives will encourage other staff (not in the pilot group) to partake in education and use the tool during emergency psychiatric admis- sions. Summary of plans made in PDSA cycles: 7 Determine whether the uptake of AUDIC-C can be increased through a targeted incentives programme. All clinicians were informed of the AUDIT-C project, and the benefi ts were identifi ed by clinicians in the trial group. Details about the new incentives programme were included via email messages and announcements at department-wide meetings. Incentive payments were provided to clinicians who partook in education sessions and then used AUDIT-C in initial intakes. Summary of learning from collected data and feedback from PDSAs run for Concept 3: Th e 93% screening target set by the US Veterans Administration was exceeded. Financial incen- tives advised by the trial group proved popular for wider uptake. Additional work was needed to improve the incentives programme and to embed AUDIT-C into electronic templates. Furthermore, a set of brief treatment algorithms was developed to support clini- cal decision making using the AUDIT-C scoring. Th e team felt that a continuous learning opportunity was generated from using the Model for Improvement to develop changes and test-specifi c change ideas based on the concepts underpinning key primary and secondary drivers in the driver diagram (see Chapter 34 for an example of a driver diagram).

106

P art 4 Q

uality im provem

ent

34 Quality improvement in intensive care Figure 34.1 A critical care driver diagram

Outcome Primary drivers Secondary drivers

Critical care

Provide appropriate, reliable and timely care to critically ill patients using evidence-based therapies

Integrated patient and family into care so they receive the

care they want

Develop an infrastructure that promotes quality care

Create a highly effective and collaborative multidisciplinary

team and safety culture

Improve critical care outcomes

(reduce mortality, infections and other adverse

events)

• Reduce complications from ventilators • Reduce complications from central venous catheters • Optimal glucose control • Prevent healthcare associated infections and cross contamination • Proper sepsis recognition and treatment

• Proper sepsis recognition and treatment • Involve patient/family in daily goal setting process • Promote open communication among team and family • Ensure clarification of care wishes and end of life care planning

• Ensure appropriate infrastructure and leadership to provide consistent, reliable, evidence-based care • Improve ICU throughout • Ensure competent staff with knowledge in improvement work

• Reliable care planning, communication and collaboration of a multi-disciplinary team

Source: Scottish Patient Safety Programme 2008

Figure 34.2 Intensive care unit average length of stay, (August 2007–August 2012)

1.6

1.4

1.2

0.8

0.6

0.4

0.2

Au g

De c

Ap r 0

Au g

De c

Ap r 0

Au g

De c

Ap r 1

Au g

De c

Ap r 1

Au g

De c

Ap r 1

Au g

1st meeting

1st reduction in length of stay

No VAPs for 7/12

Reliablility in all 8 elements of workstream

No CLABSI for 2 years

No VAPs for 3/12

SMR

Median 2nd reduction in length of stay

7 element reliability

107 C

hapter 34 Q uality im

provem ent in intensive care

Introduction Patient safety is core to all aspects of intensive care training, edu- cation and standards. Th e strong history of nurturing a safety cul- ture in the speciality, learning from error, preventing harm and working as part of a multi-disciplinary team all contribute to the disciplines of safety and critical care. Bundles, processes and checklists are all terms now familiar to those working in intensive care units (ICUs).

Problem An audit in an ICU revealed the incidence of ventilator-associated pneumonia (VAP) to be 17 per 1000 ventilated days and the central line–associated bloodstream infection (CLABSI) rate to be 22 per 1000 catheter days. Improvement was needed in terms of reduc- ing waste, harm and unwarranted variation within clinical practice and healthcare delivery.

A multi-disciplinary team was established that comprised medical and nursing staff , a pharmacist, a physiotherapist and a dietician. Th e team met regularly in person or through online discussion in order to problem-solve issues and guide testing of change.

Intervention Following an initial brainstorming at a weekly multi-disciplinary team meeting, a critical care driver diagram (Figure 34.1) and a Pareto chart (Figure 34.3) were used to hi ghlight the main chal- lenges and consider how the change package could address these challenges. A Pareto chart is used to graphically summarise and display the relative importance of the diff erences between groups of data; the lengths of the bars represent frequency of the issues and are arranged in descending order to visually depict which situ- ations are most signifi cant. A driver diagram helped to clarify our understanding of how our own ICU system worked and identify the changes we thought were needed for improvement.

Th e aim of the project was to reduce ICU standardised mor- tality ratio by 10% between December 2007 and December 2012. Th e change package comprised eight key processes that could assist in delivering reliable critical care: 1 Ventilator-associated pneumonia bundle 2 Central venous catheter (CVC) insertion bundle 3 Central venous catheter maintenance 4 Peripheral venous catheter (PVC) maintenance 5 Multi-disciplinary ward rounds 6 Daily goals 7 Glycaemic control 8 Hand hygiene

Patient stories, both good and bad, were used to describe areas of good practice and establish a burning platform when subopti- mal care had occurred (see Chapter 28). Care bundles to ensure good compliance with evidence-based practice as well as check- lists for high-risk or complicated procedures (e.g. rapid sequence induction and intubation, and CVC insertion) were also used.

Change strategy and measurement Th e Model for Improvement was used to initiate a continu- ous series of small-scale tests of change performed by a multi- disciplinary team. First it started small with one patient, one nurse and one doctor on one shift . Once some understanding about the process was gained and it was being delivered reliably, it was then tested with three, then fi ve, patients, at diff erent times of the day, prior to its rollout to the whole ICU.

Motivation and communication with members of the team were facilitated by twice-daily safety briefi ngs and an open and transparent display of infection rates in the middle of the clinical area in the ICU. Real-time assessment of data was displayed on daily and monthly run charts.

Results Signifi cant reductions in VAP and CLABSI rates were achieved with more than 250 days and 730 days between events, respec- tively. Th is was associated with a 0.5 day reduction on time spent on a ventilator, as well as a 1.1 day reduction in ICU length of stay. Th e 5 Whys (see Chapter 27) were used to explore whether the reduction in length of stay could be explained by reasons other than a reduction in hospital-acquired infection rates (i.e. less sick patients). Th e analysis revealed that despite an increase in the com- plexity and severity of cases, the ICU average length of stay had still reduced by 1.1 days with a corresponding 0.23 reduction in the standardised mortality ratio from 0.92 to 0.69 (see the run chart in Figure 34.2).

Conclusions Th e public display of the ICU’s infection rates helped the ICU transition to a transparent and safety-focused culture. Multiple small-scale tests of change, guided by the Model for Improvement, were integral to changing practice in this high-risk environment. Bundles of care, daily goals and checklists also helped to produce high-quality, reliable healthcare.

15 13 8 7

4 1 2

VA P

C LA

BS I

C om

m un

ic at

io n

M ul

tid is

ci pl

in ar

y w

ar d

ro un

ds PV

C in

fe ct

io n

H an

d hy

gi en

e G

ly ca

em ic

c on

tro l

Fr eq

ue nc

100

P er

ce nt

CLABSI, central line-associated bloodstream infection; PVC, peripheral venous catheter; VAP, ventilator-associated pneumonia

Figure 34.3 Pareto chart displaying ICU challenges

108

P art 4 Q

uality im provem

ent

35 Quality improvement in obstetrics

Figure 35.3 Cardiotograph model of improvement

How will we know that a change is an improvement?

What are we trying to accomplish?

To redesign the current process of cardiotograph (CTG) interpretation to reduce the number of cases misclassified and resulting in morbidity

Goal Measure Type of

measure

Decreased the number of cases of CTGs that are misclassified

% CTGs correctly classified Outcome

Increased staff knowledge of National Institute for Health and Care Excellence (NICE) CTG classification criteria

Score of the pre-and post-test Process

Increased staff compliance with the new process % compliance to new process Process

Increase midwife confidence Midwife confidence score Balancing

Figure 35.4 Percentage of cardiotographs (CTGs) misclassifi ed on Ward 1 (p-chart)

Day

P er

ce nt

UCL = 74%

Mean = 41%

LCL = 8%

28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54

Chart laminated on CTG machine

Presented at weekly team meeting

Introduced to Ward 1

staff

Training for all Ward 1 staff

No: Date: Time:

risk

High

Low

C: 10 Temp: Pulse:

Feature 0 Reassuring

1 Non-reassuring

2 Abnormal

<100

>180

100–109110–169

161–180

Rising baseline

<5 bpm for 40–90 min>5 bpm

BRA Start = Now =

Plan

Who

Name

Sign

Case midwife Snr midwife Fresh eyes Doctor

Continue

Conservative, e.g.

antipyretics, adjust synto

1 = Suspicious0 = Normal ≥2 = pathological

FBS Expediate delivery if FBS impossible or

acute event, e.g. brady >3 mins

Atypicals >50% contractions

>30 min

Late >30 min

Brady >3 minBrady up to 3 min

Typicals >50% contractions

> 90 minTypicals <50% for <90 min

None

Present Absent*

<5 bpm for >90 mins

Sinusoidal >10 mins

Figure 35.1 CTG score sticker

CTG deceleration algorithm

Look for the four S’s

Does the deceleration have ...

... Shoulders? (slight rise in baseline before and after decel)

... a Shape like a V?

... a Slow return baseline?

... an overShoot to a higher baseline?

Typical Atypical

Yes No

– a W or U shape

No Yes

Yes

C BRA

V A D O

= Contractions = Baseline rate = Variability = Accelerations = Decelerations = Overall Impression

Figure 35.2 The 4 S’s sticker

109 C

hapter 35 Q uality im

provem ent in obstetrics

Introduction Th e cardiotograph (CTG) is a method of assessing foetal heart rate to identify babies at risk of complication (which can include death) as a result of the birthing process. Th e interpretation of CTGs was acknowledged to be poor at a large university teaching hospital, despite regular CTG training and the use of standardised stickers. An audit by a senior midwife showed defi ciencies in classifi cation (52%), inappropriate action (56%) and inadequate documentation (24%) of the CTGs of 25 women in November 2012. National medical malpractice claims supported a strong argument to make this a focus for a quality improvement project because mistakes in CTG interpretation can result in cerebral palsy, resulting in over £2 billion of medical negligence claims in the United Kingdom over a 10 year period.

Aim Ninety per cent correct classifi cation and appropriate management of the CTG.

Intervention Introduce a CTG classifi cation sticker that was developed and suc- cessfully implemented at another hospital.

Timeline of change Days 1–4: Th e specialist registrar obstetrics doctor (SpR) shows sen- ior colleagues and peers a prototype of the CTG sticker. Feedback is negative as they feel that there is already a sticker system in place and ‘another piece of paper’ would not be helpful. Days 7–11: SpR undertakes a brief survey with 34 midwives and 16 doctors, and identifi es several key reasons for CTG misclassifi - cation: • Th e CTG terminology and accompanying guidelines kept chang- ing and had too many features to remember in an acute situation • Decelerations (a feature of the heart rate slowing down) were the most confusing feature to interpret (95%). Many (72%) were unfamiliar with the diff erence between typical and atypical decel- erations despite training day interventions • Two stickers were already in use. Neither helped with accurate classifi cation or management • Frequent need for a second opinion as unsure, and CTG was over-classifi ed as suspicious or pathological

SpR presented a new sticker to staff and received unanimous positive feedback about the layout (aided easier classifi cation), credibility (incorporated both classifi cation systems) and use- fulness (served as a reference and an educational tool, and also prompted appropriate communication and management). Sug- gestions were made to include a chart or algorithm on the CTG machine to help distinguish between atypical and typical decel- erations. Days 10–11: Findings presented to midwifery manager. Day 14: Findings presented by manager at the midwifery senior management meeting. Need for new sticker accepted, and col- our coding suggested. Mock-up of CTG deceleration algorithm drawn up. Day 19: Patient safety manager liked the tools and presented them at the quarterly organisation-wide Clinical Improvement in Mater- nity meeting. Both rejected by neighbouring hospitals in the same group as ‘another piece of paper’. Day 23: Manager and supervisors felt a pilot should go ahead and agreed to implement it as a printed CTG sheet before committing to sticker production.

Testing in practice Days 24–25: SpR tests tools with six midwives on labour ward to use with one patient each per 12-hour shift . Feedback at end of each day shows tools to be eff ective and easy to use, but decelera- tion chart could be simpler. Day 26: SpR analysed fi ndings, and CTG sticker text strengthened by minimising wording. Deceleration algorithm honed down to a simple-to-remember ‘4 S’s’ (see Figures 35.1 and 35.2 ). Day 27: Tools were updated, and fi ndings were presented to team. Day 28: Midwifery manager and supervisor laminate deceleration chart and attach to all CTG machines on labour ward. Day 33: Tools and fi ndings presented by SpR and patient safety manager to obstetricians at weekly CTG meeting to raise aware- ness. Initially rejected as ‘another piece of paper’ by consultants. Accepted by trainees and students as valuable – they took copies and photographs on their telephones for their own personal ref- erence. Evidence and rationale for tools recognised aft er demon- stration of correct classifi cation with CTG error cases brought to meeting. Participants agreed to expand the new system with a trial week for all staff on the labour ward. Days 39–43: Staff introduced to pilot and use of tools at handover. Ongoing education with tools in the form of informal reviews dur- ing the ward rounds to eliminate gaps in teaching and utilisation. Feedback and suggestions for improvement gathered daily. Days 44–continuing: Project ongoing and continuing to build momentum and popularity with midwives and junior obstetric doctors.

Measures of improvement Figure 35.3 outlines the process, outcome and balancing measures used during the improvement project. Run charts were made for each process and outcome measure.

Effects of change Process measures improved in a short time. Staff adapted to the change. CTG stickers are now present on every machine in every room on the labour ward. Senior doctors and midwives have sup- ported the project and have assigned a junior doctor to undertake a weekly analysis of 25 sets of notes to inform a run chart pre- sented at the CTG meeting each week. Incorrectly analysed CTGs are used for teaching purposes.

Figure 35.4 is a control chart to display the percentage of CTGs misclassifi ed on one of the participating wards in the improve- ment project. Th e Ward 1 team felt strongly that no CTGs should be misclassifi ed if the rules within the tool were replied reliably; thus, they decided to chart the percentage of misclassifi cations, in which an overall downward trend was positive. Note that the team has annotated key project events on their chart. Can you iden- tify special-cause rules 2, 5 and 6 (see Chapter 24 for a recap) in Figure 35.4?

Learning for others Redesign of an existing tool can encourage buy-in and acceptance, and can be cost-eff ective, especially if stakeholders with authority are involved early on to ensure leverage and sustainability.

110

P art 4 Q

uality im provem

ent

36 Quality improvement in surgery Figure 36.1 OR process improvement map

Circulator nurse

Environment services/ patient

transport

Scrub technician

Anaesthesia

Surgeon

Control board

Yes

Environmental services/patient transport

Problem: Should be available when patient being wheeled out of the OR to begin process immediately

Change: The OR nurse in charge would designate a named auxiliary nurse to make this call with 15 minutes notice

Circulating nurse

Problem: Responsible for communicating an update

Change: Call the front desk and patient holding area at time of closing to notify them that the case is finishing

Control board

Problem: Should be notified 20 minutes before case completed

Change: the auxiliary nurse calling environmental services will also call to update the control board

Anaesthetist

Problem: Evaluate the next patient before the turnover time to ensure that the workup is complete and no other steps need to be done prior to surgery

Change: A nurse anaesthetist to undertake all necessary checks 90 minutes prior to the procedure

Call for transport

and cleaning

Transport patient

out of room

Clean room

Pre-op patient

Transport to OR

Set up new equipment

Complete case

breakdown

Set up new equipment

Take patient to

PAW or ICU Handoff

Clean, check equipment Get meds

See patient and bring

to OR

Call Pre-op holding to get patient ready

Complete surgery

Pre-op next

patient

Remove case from

board

Surgeons

Problem: Ensure consent forms are properly filled out and signed by patients and pre-op workup is completed before patients come to the holding unit

Change: Junior doctors are required to document a pre-op note noting all workup is complete the day prior to surgery

OR Scrub nurse

Problem: Breakdown case set as soon as operation is complete and set up new case after room is cleaned

Change: Ensure proper equipment is pulled for each case night before surgery

Wheels out

Wheels in

Turnover time

111 C

hapter 36 Q uality im

provem ent in surgery

Problem Operating room (OR) turnover time can be a cause of ineffi ciency in surgery. Th is has many implications, including delays in sched- uled operations, staff overtime costs and patient and healthcare professional dissatisfaction.

A busy surgical department in a large US teaching hospital wanted to improve the quality of the operative experience for sur- gical patients. When the process was audited, turnover times were noted as double the national average. A multi-disciplinary team approach was initiated to engage all stakeholders in a process- mapping exercise to better understand the current system and determine why the process was failing. Insights from this learning process were used to redesign existing processes.

Experience of running a process-mapping activity A process-interaction map shows steps performed at key inter- actions between functions (inputs and outputs) and activities (actions to move the process along that are oft en performed in parallel). Th is allows team members to understand areas of waste and redundancy and fi nd ‘hot spot’ areas to focus improvement eff orts.

Step 1: Defi ne the problem Th is was defi ned as: ‘To reduce OR turnover time (the time between a patient leaving and another patient being brought in as the next case) by 50%’.

Step 2: Establishing the team Th e OR turnover process requires a team eff ort, involving sur- geons, anaesthetists, scrub nurses, auxiliary nursing staff , porters and ward nurses. Having key opinion leaders from each group was vital for project credibility. Bringing multiple professionals to the table resulted in some disagreement about the perceived processes – each member of the team had a unique view of the system and valued processes diff erently. When one group was not represented, it was diffi cult to raise questions about their contributions since the answers were not in the room (e.g. who schedules the porters? And who communicates with them if there is a delay?).

A neutral facilitator oversaw the process-mapping activity, helped to focus attention on the problems being discussed, pre- vented meetings from running off target and kept the group on track to accomplish the goals initially agreed at the outset.

Step 3: Mapping the ‘as is’ process A process map was constructed to visually depict the way the OR was currently being run. A large-scale image of the process was drawn, and small, coloured sticky notes were added to highlight value-adding and non-value-adding activities.

Th e sticky notes focused on problem areas that required atten- tion and decision making. Each step was outlined in terms of who was involved, inputs and outputs, and an estimate of the time it took for each step.

Step 4: Establish measures for improvement and propose changes Th e next goal was to fi nd measurable changes for each group to help improve the system. In Figure 36.1, each key professional role was identifi ed with their respective responsibilities during the turnover time highlighted. Problem areas involving each group were identi- fi ed, and the team proposed agreeable changes collectively.

Step 5: Mapping the ‘should be’ process and implementation Once improvements and changes had been identifi ed and agreed upon by the group, a ‘should be’ map was created. A pilot was launched in one operating room. A named team member was responsible for documenting the times for each process during each turnover. A team debriefi ng occurred every day aft er the second turnover to discuss what went right and wrong during the turnovers and the challenges that the team faced. Frequent plan– do–study–act (PDSA) cycles brought light to individual processes that required some adjustment.

Changes guided by the process map By the end of the fi rst day, the turnover time had been reduced by 50%, in keeping with acceptable national standards. Th e team imme- diately realised that a major reason for change was the improvement in communication between the teams and individual ownership of the steps taken by each team member in the process fl ow. In order to sustain the change, factors that resulted in reverting back to the old processes (i.e. staff who missed the initial trial period continued to use the old method) were identifi ed using a fi shbone diagram. Th e Model for Improvement was used to test change ideas to miti- gate the impact of those factors. A run chart of turnover time is now kept on display in the OR to ensure that the turnover time remains at a time deemed reasonable by team members.

112

P art 4 Q

uality im provem

ent

37 Population health and improvement Figure 37.1 A population health view of cardiovascular disease (CVD): socio-economic contributory factors, physical consequences and fi nancial burden

CVD = No.1 killer

Manifest as CHD, CVA, PVD

Increase predicted

Lifestyle and

socio-economic factors

contribute to CVD

Direct care

Loss of productivity

Costs UK £30.7 billion p.a.

Cardiovascular disease management

Population health management of cardiovascular disease

1st event can be fatal

1 in 3 deaths in UK

113 C

hapter 37 P opulation health and im

provem ent

Introduction Looking beyond the smaller-scale improvement considered so far, here we will consider how visions of quality care are delivered to an entire population. ‘Population health’, similar to public health, explores factors that determine health and disease in populations and also works to develop and evaluate the impact of interven- tions to improve the health and wellbeing of the public. To think at a population level, a group of patients who share a common risk factor, health goal or condition need to be defi ned. Th en, the health outcomes of the patient population can be considered, the health determinants that infl uence them can be identifi ed and the policies and interventions that impact those determinants can be brought together to best understand and develop improvement ideas for potential targets to improve health outcomes. Th is strategy relies on access to information about a variety of factors that infl uence health, as well as a willingness of healthcare stakeholders at all lev- els to put the information to use in order to devise an approach to result in sustainable improvements for patient populations.

Prevention of cardiovascular disease Th e World Health Organization (WHO) preventive strategy for reducing cardiovascular disease (CVD) focused on a population- based approach for managing people at high risk of either develop- ing or worsening already-established CVD. To tackle mass disease, reducing risk factor levels in those at highest risk either without CVD or with established CVD is achieved by targeting medical risk factors and unhealthy lifestyle behaviours. For example, given the causal relationship between salt intake and blood pressure, reducing dietary sodium intake could be benefi cial. WHO esti- mates that 49% of all coronary heart disease (CHD) and 62% of all strokes are attributable to high blood pressure. In most Western countries, the habitual salt intake is about 10 g a day – twice the level recommended by WHO. Th us, it is predicted in the United States that a 3 g dietary salt reduction per day would reduce the annual number of new cases of CHD by 60,000, stroke by 32,000 and myocardial infarction by 54,000 and the number of deaths from any cause by 44,000. A population-based approach requires prevention to be focused on groups, rather than individuals, such as those with CVD or those without, or those with a particular risk factor or those without. Research studies have demonstrated that it is possible to see up to 50% reductions in CVD mortality due to changes in risk factors, with 40% due to improvements in treatments.

Barriers to implementation To successfully achieve this larger scale thinking, buy-in at the clinical level is essential. However, there can be barriers to imple- menting CVD prevention guidelines, including time constraints, a sense that preventive strategies are not eff ective, confusion over which risk assessment tool to use, inadequate training and the potential for adverse infl uences on clinician–patient relationships.

A large UK study (ASPIRE-2-PREVENT) showed that there was a wide gap in the implementation of evidence-based medi- cine in vascular management in both hospital and primary care.

Amongst asymptomatic high-risk people, there was a large pro- portion not achieving lifestyle, risk factor level and therapeutic targets. Where guidelines are adhered to, the risk factor and thera- peutic targets may be diff erent to what the evidence base regards as optimum. For example, in the United Kingdom, a reason for not reaching the blood pressure and lipid targets in high-risk patients could be because general practitioners use the Quality Outcomes Fr amework audit targets of blood pressure of <150/90 mm Hg and a total cholesterol target of <5.0 mmol/l, which is higher than the optimal therapeutic targets of <140/80 mm Hg and <4.0 mmol/l advised by research respectively.

A large survey conducted by the British Heart Foundation, a UK charity, found that a third of people with high cholesterol and a quarter with high blood pressure do not take their medication correctly. Th is may be due to side eff ects and poor information about the benefi ts. Traditional advice giving can create resistance to change among patients. Merely telling someone they are at risk of developing a disease is rarely suffi cient to change behav- iour. Th eories about behaviour change and research associated with them have clarifi ed the need to look beyond approaches that are based simply upon delivering expert opinions. Patients pre- fer a more patient-centred approach to consultations rather than more directive advice giving, for example a nurse-co-ordinated multi-disciplinary prevention programme for high-risk individu- als and those with CVD in hospitals and primary care practices. Th e approach of eight visits over a 16-week programme, involv- ing dieticians, physical activity specialists and cardiologists, led to healthier diet, improved physical activity and more eff ective control of blood pressure for patients and their partners in eight European countries.

Opportunities to reduce CVD require population-based strat- egies including legislative changes to help reduce, for example, smoking, trans fat consumption and dietary sodium consumption. Th e United Kingdom, Portugal, Finland and Japan have reduced population-wide salt intake through a combination of regulations on the salt content in processed foods, food labelling and public education. Legislative changes to increase the facilities and oppor- tunities for exercise can further augment the population-based preventive approach.

Conclusion Tackling the burden of preventative disease is an important clini- cal and societal goal. Healthcare professionals can play a big role in disease prevention since many patients remain undiagnosed and untreated, and there are treatment gaps even among patients with established disease. Barriers to population-based eff orts are candidate areas for improvement projects locally (e.g. in general practice, or prior to patient discharge in hospitals). Th e improve- ment methods and tools in Chapters 21–36 can be used to develop improvement initiatives that focus on disease prevention (e.g. promoting better cardiovascular health to South Asian men) and health promotion (e.g. healthcare student-led projects to promote healthy eating choices to primary school children). Th e Model for Improvement could be used to learn how to successfully adopt evidence-based interventions into clinical practice.

114