Health Care Research

11 Combining Research Methods: Case Studies and Action Research

Rebecca Jester

Introduction

In Chapters 7 and 8 , we focused on the unique features of quantitative and qualitative research. In this chapter, we aim to demonstrate how research methods can be integrated and combined to address specific research questions. The chapter will provide an overview of two specific research designs: action research and case studies, together with examples from research projects conducted by the author. This chapter does not aim to provide an in-depth philosophical debate related to case study and action research approaches, but rather a practical discussion of the merits, limitations and application of these two approaches. We begin by discussing the concepts of ‘mixed methods’ and ‘triangulation', first introduced in Chapter 2 .

Mixed methods approaches

Traditionally, within health and social research, individuals have aligned themselves with either the quantitative or qualitative paradigm. However, in reality, many real world research projects benefit from mixing or combining methods. Mixed methods research can be accomplished either by using specific approaches to research, such as action research or case study, as discussed within this chapter, or by adopting a phased approach within a study. This might involve the first stage being exploratory within the qualitative paradigm, and the results from this being used to form specific hypotheses for testing within an experimental design, such as a randomized controlled trial. Equally, a quantitative approach (say, a questionnaire) might be used to gather data from a wide range of people, with the results being used to develop a qualitative interview schedule for use with a small sample of respondents.

Triangulation

Very often a research study is undertaken with multiple datasets, mixed methodology or with different researchers, such as at different sites. Triangulation is a very useful technique that enables you to enhance and verify concepts. As Ramprogus ( 2005 , p. 4) suggests, ‘triangulation … tries to reconcile the differences of two or more data sources, methodological approaches, designs, theoretical perspectives, investigators and data analysis to compensate for the weaknesses of any single strategy towards achieving completeness or confirmation of findings’. However, triangulation must be exercised with caution; it is no substitute for robust and well-established methods of establishing validity. Rather it is used to confirm or affirm, and therefore verify findings.

Combining methods in practice

The author's research interests focus on applied clinical studies related to the care and treatment of trauma and orthopaedic patients. Applied clinical research often develops from the identification of clinical problems and the need to develop robust evidence to support practice, policies and protocols. Often, clinical research questions necessitate the use of multiple methods, which may include both quantitative and qualitative approaches. Furthermore, if research findings are to be used to underpin practice, and managers and health care professionals are to be convinced to change practice based on research evidence, it is important for them to be active partners in the research process instead of being passive recipients of research findings and decisions. Healthcare practitioners and managers are bombarded with policy directives, efficiency targets and new policy on a daily basis and often feel that the pace of change is overwhelming. From observation, many practitioners and managers are reluctant to engage in the process of implementation of research findings because they do not understand the research process or data analysis, or how to interpret and implement the findings. We must strive, therefore, to support healthcare professionals to be at least competent and confident appraisers of research relevant to their practice. Action research and case studies are two practical approaches that put the patient/client and/or healthcare staff at the centre of the research process, and indeed necessitate their active participation. Being involved should help to demystify research and improve understanding and translation into clinical practice.

However, mixed methods approaches are not without their challenges. Collins, Onwuegbuzie and Qun ( 2007 ) identify four challenges to mixed methods research: representation, legitimization or validity, integration and politics. The challenge of integration of qualitative and quantitative approaches is the most complex to tackle, and Collins et al. (p. 269) pose the question: ‘Is it appropriate to triangulate, expand, compare or consolidate quantitative data originating from a large, random sample with qualitative data from a small, purposive sample?'

As researchers, we need to weigh the benefits of mixing approaches within a single study against the challenges and complexities that it involves. Most importantly, the approaches taken must be justified as the most effective method of answering the research question/s, and we must not permit our individual biases toward particular paradigms to influence those decisions. We now illustrate two ways in which methods can be integrated within a single research design: action research and case studies.

Action research

Action research has been used within the social sciences since the 1940s. Kurt Lewin (1890–1947) is generally thought to be the originator of the term ‘action research’ as a method of researching a social system, while at the same time attempting to change the system being studied. Many students and healthcare practitioners are familiar with Kurt Lewin's work related to change management theories and, specifically, force field analysis (Value Based Management, 2009 ). This begins to highlight the symbiotic relationship between generating new knowledge through the research process and using this knowledge to change practice or systems.

Action research is a form of enquiry that describes, interprets and explains a social situation while at the same time implementing an intervention aimed at change, improvement and involvement. It is problem-focused, context-specific and future-orientated, and founded on partnership between all researcher(s) and participants (Waterman, Tillen and Dickson, 2001 ).

Action Point

From the description above, identify key words and concepts that indicate action research may be a useful approach in contemporary healthcare research. Think about the words you have selected, and why.

Characteristics of action research

The words you have selected may relate to the key characteristics of action research presented below, which are adapted from Waterman and Hope ( 2008 ):

· • collaboration and participation;

· • developmental;

· • change and improvement;

· • practical and context specific;

· • cyclical;

· • involves reflection and reflexivity.

Within the context of contemporary healthcare research, most funding bodies (such as Research for Patient Benefit; National Institute for Health Research (NIHR), 2006 ) strongly encourage or indeed insist upon research designs and approaches that include the active participation of patients/service users/carers and practitioners. There is an increasing realization that many research studies do not influence clinical practice or policy, and there is a need to increase emphasis on implementation of good quality research. The cyclical nature of action research also lends itself to healthcare, as there is a constant need to evolve our evidence based practice and policy. There are several models of action research, including the early action research cycle developed by Kemmis and McTaggart ( 1982 ) which includes: identification of the problem; investigation of problem concepts and study of the related literature; designing the plan of action to solve the problem; putting the plan into action and monitoring changes to practice; reflection, changes and modifications to the solution. This model is illustrated within the case study of applied action research, later in this chapter. Subsequent models include that of Hart and Bond ( 1995 ), which includes: the stages of identifying a problem; using research methods to assess the problem; planning and implementing change; and evaluating the outcome of change.

Types of action research

Hart and Bond ( 1995 ) identify four types of action research:

· • experimental;

· • organizational;

· • professionalising;

· • empowering.

A number of authors, including Grbich ( 1999 ), purport that these four types of action research run along a continuum, whereby at the experimental end the researcher leads in all aspects of research design and process and usually conducts the study at the behest of senior managers within an organization who perceive the need to ‘fix a problem’. The organizational approach remains researcher-centric, but does include greater negotiation and consultation with those working at grass roots, that is, practitioners. Professionalising approaches begin to embrace approaches to the design and process of the research, where often the need to change practice through research is identified by the practitioners and/or service users/carers themselves, with the researcher acting as a guide and facilitator. The empowering approach views the participants as co-researchers who contribute equally to decisions about what the problem is and which approaches of design and process are needed. Ultimately, the empowering approach aims to lead to emancipation of grass roots practitioners, service users and carers, enabling them to break free from top-down approaches to improve services and instigate change to policy and practice.

The empowering/emancipatory approach aims to foster a culture of collective enquiry, reflection and problem solving. These are important benefits, as once the action research project has ended, change and continual improvement need to be sustained by practitioners, managers and service users/carers.

Case Example 11.1: Pre-operative Fasting

The author, whilst engaged as a lecturer practitioner within a specialist orthopaedic hospital, was approached by a number of ward based nurses who were concerned about pre-operative fasting policy and practice within the organization. The custom was to withhold all fluids and food for patients from midnight prior to surgery the following day, irrespective of what time the patient was scheduled to have their surgery. A critical incident where a patient's surgery was cancelled due to severe headache bought on by dehydration prompted the nurses to raise their concerns. The issue of concern was raised with the senior consultant anaesthetist and the author, although a novice to the technique, believed action research was the best approach to be taken.

Kemmis and McTaggart's ( 1982 ) model of action research was chosen to guide the process.

Identification of the problem

Patients were being starved for excessively long periods prior to surgery. There were also frequent changes made to theatre lists by surgeons, and often these changes were not communicated in a timely manner to ward staff and patients. Nurses perceived that post-operative nausea and vomiting were in part due to the excessive fasting practices.

Investigation of problem concepts and study of the related literature

The literature review revealed that it was safe for healthy adult patients to drink clear fluids up to two hours prior to surgery and consume food or milk up to six hours before. The researcher then conducted a survey of fasting times, with a convenience sample of 110 elective adult orthopaedic patients over a period of one month. This survey revealed that the mean fasting time was 11.94 hours and far exceeded the recommendations from the literature review. The literature review had also highlighted reasons why patients are starved for excessive periods pre-operatively, including:

· • lack of knowledge by nurses about the physiology of gastric emptying in healthy adults;

· • practice based on custom rather than on evidence based research;

· • changing of theatre lists leading to uncertainty about what time patients would be anaesthetized.

Designing the plan of action to solve the problem

In collaboration with the senior anaesthetist, three interlinked actions were agreed. Firstly, anaesthetists were asked to start prescribing the time patients could consume both food and fluids. Secondly, the author implemented an educational package for ward based nurses to improve their knowledge of evidence based fasting practice (Jester and Williams, 1999 ). Thirdly, pre-admission letters to patients informed them of the new fasting practices and they were encouraged to drink clear fluids up to two hours prior to their surgery, for example, at 6.30 a.m. for those on the morning list and 11 a.m. for those on the afternoon list.

Putting the plan into action and monitoring changes to practice

Following a three month period during which nurses were asked to report problems with the implementation of the new fasting practices, a follow up survey was conducted on 106 adult elective orthopaedic patients within the same clinical areas as before. The follow up survey revealed that the mean fasting time had reduced by 5.40 hours from the initial survey, but this still meant that patients were, on average, not taking clear fluids for around six hours prior to their surgery. The survey also demonstrated that 60 of the 106 patients were prescribed a last drink by an anaesthetist, while 46 were not. Of the 60 patients who had a prescription for the last time of clear fluids, nursing staff did not give the drink at the prescribed time to six patients for the following reasons:

· • three patients were asleep and were not woken up for a drink;

· • one patient refused the drink offered;

· • one patient had their position on the list moved up;

· • no reason was given for one patient.

(Jester and Williams, 1999 )

Reflection, changes and modifications to the solution

This stage was used to explore why only 60 of the 106 patients had been seen by an anaesthetist and had the last time for clear fluids prescribed. This highlighted that for patients admitted on the day of surgery, the anaesthetist often did not arrive at the hospital in time to see patients before he/she needed to be in theatre. Discussions with nursing staff also highlighted the fact that frequent changes to theatre lists made nurses fearful of encouraging patients to drink clear fluids up to two hours prior to surgery. This was because surgeons would be unhappy if they called for patients who could not go to theatre because they had been given a drink. The reasons for frequent changes to the list were investigated and this revealed that it was often because of surgeons’ preferences, or occasionally because the correct instruments were not available in theatre.

In this case illustration, it can be seen that action research was truly cyclical because more work was needed to improve communication between theatre staff and ward based staff. Furthermore, improving efficiency around availability of equipment was required, as was the need for the empowerment of nurses to advocate for their patients’ rights.

Action Point

Reflecting on Case Example 11.1 , what approach to action research was adopted? (Take into consideration there may be an amalgam of approaches.)

The approach taken in the worked example was predominantly professionalising, as the clinicians identified the problem, namely excessive pre-operative fasting periods for elective orthopaedic patients. However, it can also be viewed as adopting elements of the organizational approach, as the research methods were determined by the researcher, rather than in consultation with the clinicians.

Action Point

If you were replicating the study, what would you do differently, and why?

To answer this question, you need to consider possible limitations of the study and how these might be addressed. Limitations appeared to include the following:

· • A further element of data collection was required at the stage of ‘problem concepts investigated and related literature studied', as the author did not explore the unique issues affecting fasting practices within the organization. That is, there was an over-reliance on the literature;

· • Not all staff groups were represented in the action research process; specifically, there were no surgeons in the project team and, because they were highly influential within the organization, their support was needed to minimize changes to theatre lists and to avoid conflict when nurses adopted the new fasting policy;

· • There was no patient involvement; it would have enriched both the data collected and the implementation of improved policy to elicit patients’ views and preferences related to preoperative fasting practice.

Limitations of action research

There are a number of criticisms of action research, which include the emphasis on action and outcomes and the need to instigate change, rather than to generate theory or a new body of knowledge. Also, the researcher may have a conflict of interest if the action research is at the behest of senior managers (e.g. at the experimental/organizational end of the continuum). However, action research is one of the few approaches that offers the potential for clinicians, users and carers to be engaged in all aspects of research design and process, which greatly increases the chance that the findings of the research will be implemented and sustained within the practice setting.

Case studies

Robson ( 2002 , p. 178) defines a case study as ‘a strategy for doing research which involves an empirical investigation of a particular contemporary phenomenon within its real life context using multiple sources of evidence’. Case study methodology involves the in-depth recounting and analysis of experience. It can be used to study phenomena in health and social care at an individual level, such as one client or patient, or part of or a whole organization, such as the trauma and orthopaedic unit or an acute hospital. The researcher can adopt either a retrospective (looking back) or prospective (looking forward) approach to the case study. Yin ( 2003 ) proposes that case studies are most useful when the researcher wants to ask how and why questions. Case study is not an approach usually used to test a hypothesis or determine cause and effect; it is mostly concerned with the subjective richness of individual or organizational experiences of a phenomenon (Zucker, 2001 ). Thus, qualitative approaches to data collection and analysis are often used. However, case study approaches can also adopt quantitative approaches; for example, single-subject experiments involve the individual or unit of research being exposed to an intervention/treatment and data being collected both before and after implementation of the treatment. Furthermore, case studies can be used for exploratory, descriptive or explanatory purposes; the specific approach taken will depend on the nature of the research question(s) and the degree of control the researcher has over the phenomena being studied. Case studies are strong in reality and allow the researcher to study the phenomenon within its social, political and economic context.

Action Point

Suggest some data collection methods that would be suitable for both prospective and retrospective approaches to a case study of an individual or part of or a whole healthcare organization.

Some of these are summarized in Table 11.1 .

Characteristics of case studies

Whether the case study involves an individual or part or all of an organization, there are certain characteristics that must be identifiable, which include:

· • clearly articulated research questions;

· • propositions;

· • unit(s) of analysis;

· • a determination of how the data are linked to the propositions;

· • criteria to interpret the findings.