Reflective Project

9/25/22, 4:40 PM Print

https://content.uagc.edu/print/McNeill.2947.17.1?sections=part03,ch11,ch12,ch13,ch14,ch15,ch16&content=all&clientToken=c8cf3d54-c96f-25b0-db… 1/27

Part III Healthcare Analytics Implementation Methods

9/25/22, 4:40 PM Print

https://content.uagc.edu/print/McNeill.2947.17.1?sections=part03,ch11,ch12,ch13,ch14,ch15,ch16&content=all&clientToken=c8cf3d54-c96f-25b0-db… 2/27

11 Grasping the Brass Ring to Improve Healthcare Through Analytics: Implementation Methods

Dwight McNeill

The first two parts of the book provided an overview of the challenges, opportunities, and fundamentals for analytics to improve healthcare. The next two parts of the book provide solutions (Part III (http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/part03#part03) ) and examples of best practices (Part IV (http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/part04#part04) ). The six chapters in Part III (http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/part03#part03) provide state of the art and science solutions to some of the most vexing analytic challenges facing healthcare. These solutions directly address the Healthcare Value Framework to reduce costs, improve outcomes and revenues, and transform the business.

Using the EHR to Achieve Meaningful Results One of the most important challenges for healthcare analytics is to support healthcare reform through the Affordable Care Act (ACA) in at least three key areas: Insurance reform (especially health insurance exchanges), the Center for Medicare and Medicaid Services (CMS) innovations (especially accountable care organizations (ACOs) and the consumer oriented and operated plans CO-OPS), and health information technology (HIT) (especially meaningful use). All of these areas require advancements in analytics. The most critical barrier to the full expression of analytics is the need to digitize and connect the data “pipes” and integrate new and diverse data. Digitizing the medical record, that is, the electronic health record (EHR), finally took off in 2009 with the Health Information Technology for Economic and Clinical Health (HITECH) Act. HITECH, through Medicare and Medicaid, provides incentives to physicians and hospitals that adopt and demonstrate “meaningful use” of EHR systems. According to a 2012 National Center for Health Statistics (NCHS) Data Brief, more than 50% of all physicians had adopted an EHR system by the end of 2011, and of the remaining 50%, half plan to purchase or use one already purchased within the next year. Similarly, a 2012

survey of U. S. hospitals indicated that EHR adoption increased from 16% in 2009 to 35% in 2011.1

(http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/ch11#ch11end01)

Meaningful use focuses providers on using the EHR information to improve clinical practice, not just to comply with regulations by installing EHR systems. There are many compliance demands of the ACA in addition to meaningful use. Core administrative IT systems need to be ramped- up to provide basic reporting. For many providers, keeping up with the compliance issues consumes most of their analytical time and money. So, it might be hard to see beyond the present demands to an analytics horizon of possibilities.

Full EHR implementation will help in the delivery of care by providing just-in-time information and facilitating coordination among key providers. But after these data are used for the specific clinical purposes, they become digital exhaust and are seldom repurposed. So, the EHR should not become another siloed data bank. The healthcare system needs to move into an integrated information management system that combines the EHR data, and all of its unstructured data challenges, with other person-based data to improve outcomes and reduce costs.

Deborah Bulger and Kathleen Aller, in Chapter 12 (http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/ch12#ch12) , “Meaningful Use and the Role of Analytics: Complying with Regulatory Imperatives (http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/ch12#ch12) ,” make the case that “meaningful use” is much more than a compliance issue and that IT adoption and reporting requirements associated with it form a strong foundation on which to build a new approach to managing care. They state that the meaningful use framework seeks to “revolutionize clinical quality measurement” by making it an automatic, low-cost byproduct of the care process itself. An IT infrastructure that encourages the use of technology is a critical competency to facilitate actionable intelligence across the enterprise and distribute it to stakeholders when and where they need it to make decisions.

Improving the Delivery of Care Improving the delivery of care to achieve outcomes and efficiencies that take the U.S. healthcare system out of last place when compared to other wealthy nations should be a top priority. The voltage drop between what is known (treatment guidelines) and what is done (actual practice) results in the right care being delivered only 55% of the time. Analytics can and must show the way. Making use of the data by embedding it in clinical decision making, by turning it into useful, accessible, timely, and user-focused information is where the analytic payoff occurs.

Improving clinical decisions through analytics can occur in three ways:

• Shaping care through decision rules. These include rules for care protocols, drug interactions, diagnosis, and order sets, which can be included in EMRs.

• Monitoring and optimizing performance through balanced scorecards and dashboards, which are used for management review and interventions.

• Supporting physicians and care givers with tools for clinical decision making at the point of care.

Real change in healthcare takes place “on the ground” at the physician/patient level. So, changing physician behavior to improve the delivery of care is the challenge for analytics and healthcare leadership. One might reasonably ask, however, “if you build it, will they come?” Physicians do not use information optimally for a variety of reasons. First is the seemingly impossible task to keep track of all the emerging research about diseases and new treatments. Second, it is not always possible for doctors to get the information they need on a given patient because they cannot find it due to the digital “pipes” problem. Third, even if this information were available it might not be used because physicians are

9/25/22, 4:40 PM Print

https://content.uagc.edu/print/McNeill.2947.17.1?sections=part03,ch11,ch12,ch13,ch14,ch15,ch16&content=all&clientToken=c8cf3d54-c96f-25b0-db… 3/27

trained in and have a strong preference for intuitive thinking. Many do not have the predilection to sort through a lot of information and decision maps to make data-driven decisions. One indicator of the need for more data-driven decision making in medicine is that diagnostic errors occur

about 20% of the time.2 (http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/ch11#ch11end02)

So, the analytics task is to get the information, guidance, and insights in the hands of physicians, “their way,” just-in-time, and in their preferred delivery mode.

Glenn Gutwillig and Dan Gaines, in Chapter 13 (http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/ch13#ch13) , “Advancing Health Provider Clinical Quality Analytics (http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/ch13#ch13) ,” develop the case that clinical quality analytics must move beyond a focus on transaction to the ability to measure a health provider’s compliance to established clinical standards of care as well as to analyze the relationship between compliance and clinical outcomes. The authors present their version of a “next generation clinical quality analytics solution” and a “Clinical Quality Workbench.” They deconstruct clinical protocols into process components that describe the clinical setting, protocol entry criteria, diagnostic steps, evaluation criteria, key decision points, the treatment steps, the evaluation criteria as well as related time intervals, exit criteria, and the needed outcome measurements. They present a case study on sepsis and demonstrate how the pinpointing of noncompliance and subsequent action can improve quality and reduce costs.

Medical errors continue to be seemingly intractable to improvement. According to the Commonwealth Fund, one in three adult Americans

reported a medical mistake, medication error, or lab error in the last two years.3

(http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/ch11#ch11end03) This is the highest rate among other wealthy countries that collect the data, and the magnitude of the difference is that it is almost twice as high as the best performing countries including France, Germany, and the Netherlands.

Dean Sittig and Stephan Kudyba, in Chapter 14 (http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/ch14#ch14) , “Improving Patient Safety Using Clinical Analytics (http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/ch14#ch14) ,” concentrate on the detection that an error has occurred that is not solely reliant on a clinician’s decision to report it. They discuss the use of “triggers,” or automated algorithms, to identify abnormal patterns in laboratory test results, clinical workflows, or patient encounters. They describe algorithms to identify different types of errors including medical diagnosis, medication administration, and misuse of EHRs. They detail the analytic infrastructure components required for a successful triggers program, including an advanced EHR system, a clinical and administrative data warehouse, a set of clinically tested algorithms or triggers, and a team of clinicians responsible for investigating and managing the incidents identified by the triggers.

Managing the Health of Populations Managing the health of populations is much different from the medical management of individuals. Medical management in the healthcare system is largely about managing sickness. That’s the medical model. Population health is about the production of health, both by preventing illness and by limiting its impact on healthy functioning. In fact, the determinants of health are largely outside the healthcare system. Individual behavior is the strongest predictor of health accounting for 40% of good health, whereas healthcare contributes only 10%. Therefore the approach to producing health is different and includes a raft of other interventions than just what the doctor orders, including social interventions and a reliance on behavior change at the individual/patient/member level.

The incentives for investing in population health are getting better, including changes in payment policy from fee-for-service to global payments, a focus on outcomes and payment accordingly, and changes to the way health is produced. Population health has its roots in public health and has been the province of governments. But with the payment and outcome policy changes under way, including the regulations for health insurance exchanges and reform of insurance practices such as no denial of coverage for preexisting coverage, health plans will need to manage their collective members’ health and demonstrate their performance in a transparent way. This will require health plans to take their members seriously because they are their new customers in addition to their stalwarts, employers.

The analytics of population health management are different from clinical management. On the one hand it requires a great deal more knowledge about and the relationships with people to engage them in the coproduction of health. Claims and traditional clinical data are not enough. The various components of health including the World Health Organization’s definition physical, mental, and social well-being will require different domains of measurement and at various levels of aggregation, from clinical practices, to healthcare systems, to communities and states. Imagine what would happen if municipalities were held accountable for the health of their citizens just as they are for education, roads, public safety, and jobs.

Stephan Kudyba, Thad Perry, and John Azzolini, in Chapter 15 (http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/ch15#ch15) , “Using Advanced Analytics to Take Action for Health Plan Members’ Health (http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/ch15#ch15) ,” detail the difficulty of developing, implementing, and managing population-based care programs. They present a conceptual framework, based on “hot spotting” techniques, that defines the information requirements, analyses, and reporting that will lead to actionable results. They concentrate on the need for proactive predictive analytics that can identify likely future poor health or high cost candidates, who can be optimally impacted by programs, and who will get sufficiently engaged with the care management process to make it a success. They emphasize the foundational need for diverse, robust, integrated, and “clean” data.

9/25/22, 4:40 PM Print

https://content.uagc.edu/print/McNeill.2947.17.1?sections=part03,ch11,ch12,ch13,ch14,ch15,ch16&content=all&clientToken=c8cf3d54-c96f-25b0-db… 4/27

Adopting Social Media to Improve Health Social media is about people using online tools, such as Twitter and Facebook, along with platforms such as mobile, to share content and information. It’s the combination of the tool and platform that makes social media so “combustible.” The combo is creating a social revolution. On the one hand, it satisfies consumers’ long-held wish for convenience, simplicity, immediacy, autonomy, and technology that works for them. On the other hand, it democratizes data by changing the locus of control, emphasizing the power of networks, and how products and services, including healthcare, can be purchased, evaluated, and improved. In healthcare, it could enable more patient/people engagement in the decisions about their personal information and their health.

The implications for healthcare analytics have not totally emerged at this time. Other industries have used social media for marketing, sentiment analysis, and brand management, and healthcare is following their lead. The data for this purpose of analytics have largely come from “scraping” data from digital sources including social media sites and using them for marketing purposes. But, the real potential may lie in gathering much more relevant data from individuals with their consent and engendering their partnership to engage in data sharing activities that help them improve their life.

David Wiggin, in Chapter 16 (http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/ch16#ch16) , “Measuring the Impact of Social Media in Healthcare (http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/ch16#ch16) ,” provides an overview of current and emerging uses of social media specifically for healthcare and proposes an analytical model to measure its impact. He focuses on two general areas of impact, including provider collaboration/education and patient health, including patient education, patient affinity groups, patient monitoring, and care management. He explores how to quantify the value of social media and asserts that its real contribution is in improving population health and that the best source of data may come from patients themselves in the form of surveys. For example, he notes that patient affinity groups collect self-reported data from its members about what works and does not work, which can contribute to comparative effectiveness studies.

Notes 1 (http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/ch11#ch11end01a) . E. Jamoom, P. Beatty, A. Bercovitz, et al., “Physician

Adoption of Electronic Health Record Systems: United States,” 2011. NCHS data brief, no 98. Hyattsville, MD: National Center for Health Statistics, 2012.

2 (http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/ch11#ch11end02a) . Pat Croskerry, “Clinical Decision Making and Diagnostic Error,” presentation at Risky Business London, May 24, 2012, www.risky-business.com/talk-128-clinical-decision-making- and-diagnostic-error.html (http://www.risky-business.com/talk-128-clinical-decision-making-and-diagnostic-error.html) .

3 (http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/ch11#ch11end03a) . Commonwealth Fund, “Why Not the Best? Results from the National Scorecard on U. S. Heath System Performance,” 2011, www.commonwealthfund.org/Publications/Fund- Reports/2011/Oct/Why-Not-the-Best-2011.aspx?page=all (http://www.commonwealthfund.org/Publications/Fund-

Reports/2011/Oct/Why-Not-the-Best-2011.aspx?page=all) .

9/25/22, 4:40 PM Print

https://content.uagc.edu/print/McNeill.2947.17.1?sections=part03,ch11,ch12,ch13,ch14,ch15,ch16&content=all&clientToken=c8cf3d54-c96f-25b0-db… 5/27

12 Meaningful Use and the Role of Analytics: Complying with Regulatory Imperatives

Deborah Bulger and Kathleen Aller

The Health Information Technology for Economic and Clinical Health (HITECH) provisions in the 2009 American Recovery and Reinvestment Act (ARRA) created a tremendous opportunity for physicians, hospitals, and health systems to adopt electronic health record (EHR) systems. The legislation includes significant financial incentives designed to accelerate EHR use and ultimately reduce healthcare costs by improving quality, safety, and efficiency. However, the incentives are tied to demonstrating meaningful use of certified EHR technology based on specific measures and milestones that must be documented and reported.

On July 28, 2010, the Department of Health & Human Services published two companion rules finalizing Stage requirements for healthcare

providers and for certified technology.1 (http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/ch12#ch12end01) Under the final rule, eligible providers and hospitals need to report the results of a core set of measures and a menu set of measures as part of the demonstration process. These measures are paired with meaningful use objectives—such as using computerized provider order entry (CPOE) or recording smoking status—and apply to eligible providers, hospitals, or both. One of the core objectives is to report automatically computed quality measures to the Centers for Medicare and Medicaid Services (CMS). Within this one objective are 15 clinical quality measures for hospitals, evaluated for all patients regardless of payer. The same objective for eligible providers breaks down further into core and specialty measures, but there are many fewer for a given provider.

Since the ARRA legislation became law, there has been a flurry of activity, including the federal rule-making process for meaningful use and certification. Eligible providers and hospitals that plan to qualify for incentives must demonstrate meaningful use; health information technology (IT) vendors are responsible for achieving EHR certification. Vendors responded rapidly to the requirements to optimize EHR objectives and measurement. At this writing, the Office of National Coordination for Health Information Technology (ONC) lists 160 inpatient

and 363 ambulatory applications from various vendors that are certified as either complete or modular EHRs for Stage 1 objectives.2

(http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/ch12#ch12end02)

To receive full EHR incentives, hospitals must be meaningful users of certified technology by Federal Fiscal Year 2013. Eligible hospitals and providers are investing significant financial and human resources in assessing the current status of EHR technology in their organizations and devising strategies to mitigate any potential gaps. A report published by Accenture in January 2011 estimates that 90% of U.S. hospitals will need to install or upgrade EHR technology during the next three years and that approximately 50% are at risk of not achieving meaningful use

by 2015 when penalties will begin to take effect.3 (http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/ch12#ch12end03) A recent HIMSS Analytics survey reports that only 27% of hospitals responding to the survey since the final rules were published expect to meet Stage 1 meaningful use

requirements by July 2012.4 (http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/ch12#ch12end04)

The findings from these and other studies suggest that achieving meaningful use will require an accelerated investment in technology but will also demand a strategic commitment to the use of information to drive behaviors and ensure EHR adoption across the enterprise. This convergence of technology, information, and behaviors is a critical predictor of successful business performance as described by Marchand, et

al.5 (http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/ch12#ch12end05) In this regard, “meaningful use” may be one of the most influential developments in recent history to drive analytics maturity in the healthcare market.

Supporting EHR Adoption with Analytics A basic HITECH assumption is that measuring the quality of clinical care and IT usage should flow automatically from using an EHR system during patient care. For this automation to occur, several prerequisites must be in place:

• Certified EHR functionality must be deployed throughout the provider organization. An example of IT functionality is nursing documentation.

• That functionality must include the necessary clinical content to support the required data collection. For example, to record the smoking status of a patient, a single structured documentation element needs to be in place.

• The functionality must be deployed using a prescribed workflow and methodology to help ensure that the data collected are consistent and comprehensive. To build on the previous example, the prescribed workflow would embed collecting smoking status for all patients age 13 and older in the admission assessment conducted by a caregiver, and it would cue the caregiver that documentation is missing or unconfirmed until properly collected or updated.

Within the framework of the meaningful use rule, healthcare providers must achieve and report specified results for each of the IT functionality measures associated with meaningful use objectives. In the previous example, incentive payments depend on documenting specific findings for more than 50% of the applicable patients treated during the measurement period. No organization wants to reach the end of a reporting period only to tally up its results and find them deficient. It is therefore essential that EHR functionality be supplemented with a way to continuously measure the adoption of key EHR components and track internal performance against the full set of IT measures. Certified EHRs must be able to calculate metrics associated with the objectives to quantify capabilities and adoption levels. The intent is to record whether:

• Features are activated

• Communication functionality has been tested

9/25/22, 4:40 PM Print

https://content.uagc.edu/print/McNeill.2947.17.1?sections=part03,ch11,ch12,ch13,ch14,ch15,ch16&content=all&clientToken=c8cf3d54-c96f-25b0-db… 6/27

• EHR components are in use by a percentage of users

• Basic data of interest are being collected for a portion of the patient population

To be effective, information must reach the people charged with improving performance in a timely and appropriate fashion. For measures related to patient care activities, reporting must reach caregivers in real-time. For this reason, there are meaningful use objectives to implement drug and allergy checking, and to align that checking with CPOE-based medication orders. By regularly monitoring the measure across individual units, caregivers, or shifts, the organization can address issues of adoption or care processes that affect patient care outcomes. For measures that affect payments such as meaningful use incentives, managers can ensure that performance does not drop below required levels. This aggregated, trended reporting, supplied through an organization’s enterprise intelligence solution, will increase visibility to meaningful use objectives across all stakeholders.

A well-designed enterprise dashboard should support the analysis of multiple aspects of the meaningful use objectives. For example, it is not enough to simply track the percentage of orders placed by authorized providers using CPOE. It is also necessary to know who is ordering, what they are ordering, and more importantly, who is not using CPOE. This information is essential to put in place the necessary coaching, training, and system modifications to support adoption, and the attainment of meaningful use. As an organization progresses along the adoption path, it may decide to set a more aggressive goal than the Stage 1 meaningful use requirement for CPOE.

Shifting the Quality Analysis Paradigm Clinical quality measurement is already a prominent component of The Joint Commission certification process and CMS incentive payment mechanisms such as Hospital Inpatient Quality Reporting Program (Hospital IQR). However, the computation and submission of measures today is typically a cumbersome, costly, manual process that is almost entirely retrospective, while doing little to drive behavior change. Several organizations have tried to automate the data collection with varying degrees of success. Mapping the data elements needed to calculate the current measures to the technology required to capture them suggests that, assuming a completely digital environment, approximately 60% of the data could be captured electronically but that over half of those elements would require human intervention for validation and quality control. Primary barriers to automated data collection fall into three categories:

• Measure specifications that are incompatible with automated data capture and that specifically require human intervention.

• Lack of standardized documentation across the healthcare system. As an example, smoking status may be documented by different caregivers in multiple locations using a variety of text fields and lookup tables.

• Low adoption rates of technology from which key data must be extracted. Gartner reports that only 20% of hospitals use clinical documentation in the emergency department, while only 5% use perioperative charting and anesthesia documentation as part of an

EHR.6 (http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/ch12#ch12end06)

Besides the barriers to data access, the descriptive nature of the reporting has prevented organizations from developing a competency for analytics much beyond ad hoc reports and query drilldown. Furthermore, the intense focus on regulatory requirements relegates the healthcare constituent to a secondary audience with access to results only after they have been submitted to CMS months after the patient has been discharged (see Figure 12.1 (http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/ch12#ch12fig01) ).

The meaningful use framework seeks to revolutionize clinical quality measurement by making it an automatic, low-cost by-product of the care process itself. Conceptually, this goal assumes that the required data for measure calculation are captured within the EHR during care and then flow seamlessly to reporting and data submission mechanisms. To accomplish this, existing measure specifications must be completely redesigned, or retooled, to transform them from manual measures to so-called eMeasures.

The initial work on measure redesign was done by the Health Information Technology and Standards Panel (HITSP), under contract to CMS, to modify three sets of existing quality measures to use EHR-generated data directly. Fifteen of the original 16 measures, covering Emergency Department (ED) throughput, stroke, and venous thromboembolism (VTE), were adopted in the final rule for use by eligible hospitals.

9/25/22, 4:40 PM Print

https://content.uagc.edu/print/McNeill.2947.17.1?sections=part03,ch11,ch12,ch13,ch14,ch15,ch16&content=all&clientToken=c8cf3d54-c96f-25b0-db… 7/27

Figure 12.1 Changing the quality reporting paradigm

Managing quality measurement under this new framework represents a significant paradigm shift for healthcare organizations. Historically, quality measurement has been treated as a separate business process, employing a host of abstractors to collect data after the fact. The new paradigm treats data as a core clinical competency driving information to and from the point of care. Using data as a by-product of patient care will eventually reduce dependency on manual chart abstraction but will also demand new analytics skills on the part of clinicians. While extrinsic drivers may define quality measurement in this new model, organizational stakeholders clearly become the primary audience for the results, creating intrinsic motivation for organizational improvement. Although measure calculation will continue to be retrospective, there should be less lag time for reporting with electronically generated measures. Additionally, the measure design begins to lay the foundation for more prescriptive clinical workflows, and hence more prescriptive analysis—that is, the ability to use evidence-based guidelines and subsequently capture data that facilitate better outcomes and predictive analysis. This approach drives hospitals up the analytics maturity curve by enabling use of real-time alerts and query tools to ensure that more uniform care is delivered in a timely manner.

In Stage 1, CMS has not specified achievement targets for clinical quality measures. However, the measures lend themselves to statistical analysis and forecasting aspects of the care process. For example, measures of ED admit to departure times facilitate the use of statistical process control tools to evaluate variation, correlate root cause, forecast capacity, and anticipate throughput barriers. By measuring incidence of preventable venous thromboembolism (VTE), the VTE measure set strongly correlates process of care (i.e., adherence to evidence-based guidelines for VTE prevention), with patient outcome. As a result, organizations can capitalize on the propensity of these measures to change behaviors and drive improvement.

Driving Analytics Behaviors CMS has clearly adopted the principle of defining measurable goals and aligning them with incentives. It now lies with participating eligible hospitals to manage performance against those goals. Given the specificity of those goals, one may question whether meaningful use will help or hinder an enterprise journey toward analytics maturity. To answer that question, we organized measures along three aspects of care:

• Primary stakeholder accountability

• Action required to achieve the measures

• Method by which the action is taken

By combining the objectives-based and quality measures in this manner, patterns emerge that enable targeted coaching and change management.

9/25/22, 4:40 PM Print

https://content.uagc.edu/print/McNeill.2947.17.1?sections=part03,ch11,ch12,ch13,ch14,ch15,ch16&content=all&clientToken=c8cf3d54-c96f-25b0-db… 8/27

Accountability

There has been considerable focus on CPOE implementation and adoption in hospitals, with the physician as the primary “owner” for approximately 30% of the combined IT functionality and quality measures (see Figure 12.2 (http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/ch12#ch12fig02) )—for example, the use of CPOE for medication orders or the appropriate discharge medication orders. Sixteen percent of the measures are primarily directed at nursing care, such as charting vital signs or providing stroke education. However, 34% of the measures are based on care that is typically interdisciplinary, such as maintaining up-to-date problem lists or the occurrence of preventable VTE. This suggests a need for a cross-functional approach to improvement and shared accountability for the results. Organizing measures along the lines of accountability and using analytics to create transparency empowers stakeholders to take appropriate action to ensure continued success.

Figure 12.2 Accountability for measures

Action

Viewed another way, when one considers the action (see Figure 12.3 (http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/ch12#ch12fig03) ) that the measure describes, 28% of the measures evaluate activities that occur during transitions of care, such as medication reconciliation or ability to provide a summary of care at discharge. Care coordination—“insuring that patients receive well-coordinated care within and across all healthcare organizations, settings, and levels of care”—is one of eight key objectives developed by the National Priorities Partnership. Aimed at improving quality and reducing disparities in care, it will likely be a continued focus of meaningful use in Stage 2 and 3. While compliance is measured at the individual objective level, it is important to view measures holistically. The ability to analyze disparities in care depends on the availability of descriptive data (such as race,

ethnicity, language need, and socioeconomic status) for populations at risk for poor quality care.7

(http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/ch12#ch12end07) Therefore, objectives such as recording patient demographics play a significant role in monitoring potential disparities in care in underserved populations. In other words, failure to achieve meaningful use in one objective may have a direct impact on another.

Figure 12.3 Action provided by accountable stakeholder

9/25/22, 4:40 PM Print

https://content.uagc.edu/print/McNeill.2947.17.1?sections=part03,ch11,ch12,ch13,ch14,ch15,ch16&content=all&clientToken=c8cf3d54-c96f-25b0-db… 9/27

Method

How the objectives are achieved represents the most complex aspect of measure management (see Figure 12.4 (http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/ch12#ch12fig04) ). Impressively, 23% of the objectives should be achieved electronically, such as the incorporation of lab results or the e-exchange of key clinical information at care transition points, and are a primary function of the IT infrastructure. Objectives such as ensuring overlap thrombolytic therapy for stroke patients requires intervention by multiple stakeholders—the physician, clinical pharmacist, and patient. Nearly 40% of the objectives are achieved through either assessment of the patient or decision points—that is, drug formulary checks or decision to admit from the ED. EHR technology augments and enables care decisions but does not replace critical thinking and clinical expertise. Analytics provides access to knowledge about the care delivery model and the ability to dissect patterns and trends that may impede progress. For example, is a delay in the decision to admit a patient from the ED to an inpatient bed a function of communication, knowledge, or capacity? Each root cause commands a different response. Understanding the “how” of measure compliance combined with the “who” and “what” enables an organization to manage improvement with laser focus while ensuring overall enterprise success.

Figure 12.4 Method by which action is taken

“Wringing every drop of value”8 (http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/ch12#ch12end08) from meaningful use measures in Stage 1 will lay the foundation for analytics expertise as these measures become more pervasive. As stated by Michael Davis, executive director of HIMSS Analytics, “Hospitals that survive the upcoming healthcare delivery transformation will be organizations that understand the need to use EMRs to collect, manage, share, and analyze data with the intent to continually improve their care delivery processes using best practices and

evidence-based medicine protocols.”9 (http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/ch12#ch12end09)

Anticipating the Future The development of new measures will likely require longitudinal data to measure across care settings and use data outside the traditional EHR. The National Quality Forum (NQF) has taken over guidance for developing new measures and retooling existing ones as eMeasures. As its work begins to stretch the boundaries of current analytics capabilities, healthcare will likely need tools that are neither widely used nor available today. Both the National Priorities Partnership and ONC’s Health IT Policy committee recommend that future stages of ARRA and other federal performance measurement efforts include the measurement of clinical outcomes, safety, and efficiency, necessitating expanded documentation with a clinical decision support system (CDSS). A HIMSS Analytics 2009 survey reports that the adoption rate of CDSS at the physician

documentation level is approximately 11%.10 (http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/ch12#ch12end10) In Stage 1, physicians must attest to the use of one clinical decision support rule per patient during the reporting period, in preparation for more extensive use. Additionally, as the complexity of measures increases, the need to aggregate disparate sources of data from across multiple care settings will accelerate the need for integrated clinical and financial business intelligence and data warehouse solutions. Gartner research published in July 2010 suggests that less than 5% of hospitals have deployed integrated clinical and financial business intelligence and data warehousing solutions, indicating a

need for rapid growth in this emerging market.11 (http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/ch12#ch12end11)

While the achievement of meaningful use goals is designed to produce tangible patient care results and significant financial incentives, measurement of IT adoption and related quality improvement represents only one aspect of a broader management imperative. The concept of “aspirational” metrics—that is, aspiring to a new standard—requires that organizations strive to not only meet current requirements but to “move the bar” toward organizational excellence. Sustainable success derives from a multifaceted strategy that requires constant care and feeding, and the management of an overwhelming amount of new data. While pursuing meaningful use incentives, organizations must continue to manage the overall business of healthcare, including

• Demonstrating the competency of the clinical staff

9/25/22, 4:40 PM Print

https://content.uagc.edu/print/McNeill.2947.17.1?sections=part03,ch11,ch12,ch13,ch14,ch15,ch16&content=all&clientToken=c8cf3d54-c96f-25b0-d… 10/27

• Ensuring that patient care occurs in a timely, efficient manner in the right care setting

• Recruiting and retaining qualified employees

• Modeling revenues and managing the costs of operations to remain financially viable beyond incentives

• Increasing market share

These business dimensions cannot be measured in silos. The correlation of cost, quality, care coordination, and efficiency is necessary to influence organizational decision making and to remove departmental and political barriers to drive transparency to key business processes.

Conclusion Meaningful use, if embraced as a foundational catalyst and not a regulatory burden, could well be the driver that promotes analytics maturity in healthcare. The meaningful use and IT adoption and reporting requirements form a strong foundation on which to build a new approach to managing the business of care. Building an IT infrastructure that encourages the use of technology and provides more time to care for patients is the first step. It is also a critical factor in the evolution of information management: the ability to transform data created as a by-product of patient care into actionable intelligence across the enterprise and distribute them to stakeholders when and where they need them to make decisions. This process requires aggregating data from multiple sources and applying healthcare logic and rules. Finally, managing behaviors by instilling the value of information in the people who manage care begins to create a culture of shared accountability for the organization’s

results.12 (http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/ch12#ch12end12)

Notes 1 (http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/ch12#ch12end01a) . 42 CFR Parts 412, 413, 422 et al. Medicare and

Medicaid Programs; Electronic Health Record Incentive Program; Final Rule 42 CFR Parts 412, et al. Medicare and Medicaid Programs; Electronic Health Record Incentive Program; Proposed Rule, Federal Register, Vol. 754, No. 1448, 1/137/28/2010, pp. 443131843- 201144588, http://edocket.access.gpo.gov/2010/pdf/2010-17207.pdf (http://edocket.access.gpo.gov/2010/pdf/2010-17207.pdf) .

45 CFR Part 170 Health Information Technology: Initial Set of Standards, Implementation Specifications, and Certification Criteria for Electronic Health Record Technology; Final Rule Federal Register, Vol. 75, No. 144, 7/28/2010, pp. 44589-44654, http://edocket.access.gpo.gov/2010/pdf/2010-17210.pdf (http://edocket.access.gpo.gov/2010/pdf/2010-17210.pdf) .

2 (http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/ch12#ch12end02a) . Details may be found at http://onc- chpl.force.com/ehrcert/CHPLHome (http://onc-chpl.force.com/ehrcert/CHPLHome) .

3 (http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/ch12#ch12end03a) . “Lessons from EHR Pros,” HealthData Management, http://www.healthdatamanagement.com/news/ehr-implementation-lessons-meaningful-use-41761-1.html (http://www.healthdatamanagement.com/news/ehr-implementation-lessons-meaningful-use-41761-1.html) .

4 (http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/ch12#ch12end04a) . Read more: EMR Daily News, http://emrdailynews.com/2011/02/24/new-himss-analytics-data-shows-44-of-hospitals-likely-to-be-ready-for-stage-1-of- meaningful-use/#ixzz1HoQwrlz7 (http://emrdailynews.com/2011/02/24/new-himss-analytics-data-shows-44-of-hospitals-likely-to-be-ready-

for-stage-1-of-meaningful-use/#ixzz1HoQwrlz7) .

5 (http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/ch12#ch12end05a) . Adapted from D. A. Marchand, W. J. Kettinger, and J. D. Rollins, “Information Orientation: The Link to Business Performance” (Oxford: Oxford University Press, 2000).

6 (http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/ch12#ch12end06a) . Hype Cycle for Healthcare Provider Applications and Systems, 2010, Thomas J. Handler, M. D., July 27, 2010 ID number G00205364.

7 (http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/ch12#ch12end07a) . “Future Directions for the National Healthcare Quality and Disparities Reports,” Institute of Medicine Report Brief, April 2010, www.iom.edu/ahrqhealthcarereports (http://www.iom.edu/ahrqhealthcarereports) .

8 (http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/ch12#ch12end08a) . Thomas Davenport et al., Competing on Analytics: The New Science of Winning (Harvard Business School, 2007).

9 (http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/ch12#ch12end09a) . Michael W. Davis, Executive Director, HIMSS Analytics, “The State of U.S. Hospitals Relative to Achieving Meaningful Use Measurement.”

10 (http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/ch12#ch12end10a) . Ibid.

11 (http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/ch12#ch12end11a) . Hype Cycle for Healthcare Provider Applications and Systems, 2010, Thomas J. Handler, M. D., July 27, 2010, ID number G00205364.

12 (http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/ch12#ch12end12a) . Adapted from D. A. Marchand, W. J. Kettinger, and J. D. Rollins, Information Orientation: The Link to Business Performance (Oxford: Oxford University Press, 2000).

9/25/22, 4:40 PM Print

https://content.uagc.edu/print/McNeill.2947.17.1?sections=part03,ch11,ch12,ch13,ch14,ch15,ch16&content=all&clientToken=c8cf3d54-c96f-25b0-db… 11/27

13 Advancing Health Provider Clinical Quality Analytics

Glenn Gutwillig and Dan Gaines

Current systems typically measure healthcare as a series of transactions rather than a process that may or may not be executed as designed. To reach the next level of quality in care delivery, practitioners require a comprehensive set of facts around healthcare delivery including compliance with every aspect of the protocol and the outcome(s) of care.

The public and private payer focus on managing population health is driving providers to transition from focusing on acute care delivery to an outcome based model. This focus is designed to reward delivery efficiency, lower cost, and accountability across the full continuum of care providers.

Increasingly educated and vocal consumers are demanding platforms that will enhance their participation in the healthcare continuum. They are demanding proof of quality care and positive outcomes in addition to reasonable access and cost. These metrics require that data typically imbedded in electronic medical records (EMRs) and across health information exchanges (HIEs) be available to the consumer.

Clinical Quality Analytics Background Over the past 25 years, provider organizations have attempted to adopt evidence-based clinical treatment guidelines as a cornerstone of their clinical quality programs. Their clinical quality measures are based on specific evidence-based practices, developed over time, which have been shown to provide the best care results to the most people. Example areas where clinical quality is typically measured and reported include heart attack, heart failure, stroke, pneumonia, and surgical care protocols and processes. Hospitals all across the United States measure and report their performance quality in these clinical areas. Today, data are generally updated and trended on a quarterly basis. Each measure represents what percentage of the time patients received the recommended care, so a higher score represents better performance. In most organizations, these clinical guidelines are implemented through education and, in some cases, EMR clinical workflows and order-sets.

To address the ongoing need to improve quality and meet regulatory benchmarks, most if not all hospital organizations have deployed a clinical quality organization. At their core, these organizations are chartered with the responsibility to measure the quality of care being delivered and work with the clinical leadership to identify improvement opportunities and drive process change. The assessment of adherence to defined clinical guidelines and processes has been at the heart of measuring the quality of the care process. However, compliance with the guidelines continues to be both challenging and problematic for most providers. For example, the New England Health Institute’s (NEHI) research points to several contributing factors to poor compliance, including counterincentives in the payment system, the lack of integrated data, physician culture because many doctors receive little feedback on adherence to evidence-based clinical practice guidelines, and the development of guidelines themselves. In particular, the research points to the lack of inclusion and transparency in current guideline development leading to a decreased level of trust by physicians in the clinical practices they are being asked to follow.

Moving forward, the ability to measure a health provider’s compliance to process as well as to analyze the relationship between process compliance and population clinical outcomes will be a fundamental requirement. In addition, evaluating outcomes and costs associated with care presents an additional opportunity to improve care delivery and providers’ management processes.

Collecting, analyzing, and reporting clinical quality data are often difficult and are viewed by many provider organizations as burdensome. In addition, the financial incentives from the payers to focus attention on clinical quality in the health provider community have historically been lacking. There have been incentive programs of various forms, but they have been, for the most part, inconsistent and of minimal incremental value to the providers.

Today, the drivers for improved quality analytics are intensifying, and the nature of analytics is evolving. Key drivers include

• Reimbursement model changes—Pay for Performance (P4P), Accountable Care Organizations (ACOs), and value-based purchasing will demand greater focus on quality, with revenue and profits at risk. Many health systems now share the view that the path to becoming a financially successful ACO must include constant monitoring of clinical performance, evaluation and improvement of clinical quality, reliability, and operational efficiency over the full range of providers from community-based to tertiary medical centers. It is becoming increasingly clear that payers of all types will demand greater focus on quality, with increasing amounts of revenue at risk for noncompliance.

• Government reporting requirements—Never events, meaningful use, Center for Medicare/Medicaid Services (CMS) Core, The Joint Commission (TJC), formerly the Joint Commission on Accreditation of Healthcare Organizations (JCAHO)—requirements are all evolving toward outcomes management.

• Evolving quality measurement programs—The Physician Quality Reporting Initiative (PQRI) approach is phasing out in favor of more stringent National Quality Forum (NQF) standards. Outcomes will become more important than process compliance.

• Evidence based care—A significant increase in clinical protocols will require an in-depth understanding of compliance and process management. In evidence-based medicine (EBM) the ability to analyze the impact of specific interventions is becoming a key issue.

Accenture’s Health Analytics team working with leading providers is pioneering the design, development, and deployment of next generation clinical quality analytics. These solutions support health providers’ quality improvement initiatives and will capture clinical treatment guidelines as process definitions, evaluate and analyze clinical encounters for their compliance to the appropriate guidelines, and assess the impact of noncompliance on clinical and financial outcomes. Identifying noncompliance issues will facilitate the understanding of the root cause

9/25/22, 4:40 PM Print

https://content.uagc.edu/print/McNeill.2947.17.1?sections=part03,ch11,ch12,ch13,ch14,ch15,ch16&content=all&clientToken=c8cf3d54-c96f-25b0-d… 12/27

of compliance issues allowing for more focused and efficient process improvement programs resulting in both improved outcomes and reduced costs. Finally, this approach creates the ability to track and analyze the results of improvement programs more efficiently and with greater accuracy than is typically possible today through either standard outcomes or key performance indicators (KPIs) monitoring.

Access to EMR data is an essential prerequisite, but not the complete source of data. The data used to measure compliance to protocols must be structured and augmented to act as a meaningful source for clinical quality information. Currently EMR systems are focused on patient centric “record keeping” and used to manage care at the individual level.

While timeline data is captured in EMR systems, the data are not organized to analyze the temporal (chronological) process of care. The next generation of clinical quality analytics will provide the ability to

• Focus on “events” at both the patient and specific target population levels.

• Analyze a treatment outcome from a quality perspective, either a relatively short-term event like sepsis, or long-term treatment of a chronic condition like diabetes.

• Compare actual events to expected or desired events, down to the patient level, for specific populations under care.

• Derive and predict the impact of compliance.

Next Generation Clinical Quality Analytics Solutions To dramatically improve the state of clinical quality analytics several key capabilities and technologies are integrated to create a comprehensive solution:

• The ability to model multistage complex clinical processes using a time based clinical process modeler for patients with a range of demographic and clinical factors

• The ability to compare enriched data sets of actual care events against the yardstick of the clinical protocol process maps

• The availability of text mining capabilities and a user interface for required manual chart reviews

• A compliance evaluation and analytics engine to assist with root cause analysis and provide predictive capabilities that can model the impact of performance improvement

• A clinical review and collaboration environment

• An analysis and visualization environment for clinical quality data

• A workbench model as shown in Figure 13.1 (http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/ch13#ch13fig01) to support numerous parallel quality improvement initiatives

Figure 13.1 Accenture’s Clinical Quality Workbench Analysis Architecture (Source: Accenture, © 2011)

9/25/22, 4:40 PM Print

https://content.uagc.edu/print/McNeill.2947.17.1?sections=part03,ch11,ch12,ch13,ch14,ch15,ch16&content=all&clientToken=c8cf3d54-c96f-25b0-d… 13/27

The Clinical Quality Analytics in Action—Case Study: Sepsis Overview To test the potential for this new set of quality metrics, several case studies have been initiated. Though not very common, sepsis is a serious medical issue in hospitals, arising from infections that overwhelm the bodies’ defenses and can result in death in as many as 40% of cases. Sepsis is characterized by a whole-body inflammatory state (called a systemic inflammatory response syndrome or SIRS) and the presence of a known or suspected infection. The body may develop this inflammatory response by the immune system to microbes in the blood, urine, lungs, skin, or other tissues. A lay term for sepsis is blood poisoning, more aptly applied to septicemia, as used in the example. Severe sepsis is the systemic inflammatory response with infection, plus the presence of organ dysfunction.

Sepsis was selected as an example to demonstrate the Clinical Quality Workbench for several reasons:

• Treatment of sepsis has been identified as a potential cost and outcome opportunity for leading health providers.

• A standing protocol to manage and treat the condition is available.

• Outcomes and progression can be clearly defined.

• Most of the data necessary to assess compliance should be available in the hospital’s EMR system.

To facilitate robust analytics capabilities we defined the existing clinical protocol for sepsis as a time-based process.

This process definition describes the clinical setting, protocol entry criteria, diagnostic steps, evaluation criteria, key decision points, treatment steps, and evaluation criteria as well as related time intervals, exit criteria, and the needed outcome measurements.

In addition, while much of the clinical protocol data is available in discreet data elements, such as lab, pharmacy, and vital signs, some critical data points can often only be found in the free text areas of the EMR, such as physician and nursing notes and radiology reports. Thus within the context of each clinical protocol, it is often highly desirable to capture valuable unstructured or semistructured data (related to notes, images, etc.) and combine it with the structured EMR data to perform a holistic and robust analysis.

Figure 13.2 (http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/ch13#ch13fig02) highlights a specific process component of the sepsis protocol and illustrates how an actual encounter is compared to the protocol. Each process component defines specific functions, tasks, and related measurement criteria including the definition of compliance for the step. For example, in Step B14 the mental status of the patient must be checked and documented every 2 hours. In this single process step, a high rate of noncompliance, (49%), was observed, which directly correlated to increased mortality rates.

Figure 13.2 Clinical protocol for sepsis (Source: Accenture, © 2011)

9/25/22, 4:40 PM Print

https://content.uagc.edu/print/McNeill.2947.17.1?sections=part03,ch11,ch12,ch13,ch14,ch15,ch16&content=all&clientToken=c8cf3d54-c96f-25b0-d… 14/27

The output of the clinical quality analysis correlated advanced sepsis and septic shock with low compliance scores for one unit in the study hospital (see Figure 13.3 (http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/ch13#ch13fig03) ).

The data supported further investigation of process compliance patterns for multiple delivery units within the hospital and root cause analysis. Figure 13.4 (http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/ch13#ch13fig04) depicts a more detailed analysis of compliance to protocol by unit 2.

Figure 13.3 Correlation between advanced sepsis/septic shock and low compliance scores (Source: Accenture Health Analytics Research, © 2011)

Figure 13.4 Compliance to protocol by unit 2 (Source: Accenture Health Analytics Research, © 2011)

Figure 13.5 (http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/ch13#ch13fig05) illustrates significant variations in compliance to protocols by shift staff and specifically variation to protocols around the time of shift change. The contributing factors were identified, and feedback and process improvement were developed to support corrective efforts.

9/25/22, 4:40 PM Print

https://content.uagc.edu/print/McNeill.2947.17.1?sections=part03,ch11,ch12,ch13,ch14,ch15,ch16&content=all&clientToken=c8cf3d54-c96f-25b0-d… 15/27

Figure 13.5 Compliance variations by staff by shift and by hour (Source: Accenture Health Analytics Research, © 2011)

The ability to pinpoint areas of noncompliance and to improve quality will drive significant clinical and financial impact for healthcare providers. For this one protocol:

• The impact of improving compliance on unit 2 to just the average of the hospital’s two other units:

• Would be expected to result in four fewer deaths from sepsis on this unit alone per year

• Would be predicted to result in 23% fewer patients progressing to severe sepsis or septic shock, saving the hospital $250k to $500k overall in this unit

• Lifting performance across units for this one area of sepsis could reduce cost by $1.5 to $2 million overall.

Extending this type of clinical quality analysis across the hundreds of protocols in place in a typical hospital can drive significant impact. For example, by extrapolating across a series of protocols for a medium to large hospital system, the improved cost performance opportunities are in the $50+ million per year range.

Conclusions The pressure to improve quality across the healthcare continuum has continued to increase and is being driven by patients as well as regulations. New technology has provided an opportunity to make significant improvements in care through understanding the detailed sources of variation and error in the care delivery system. Accenture is working with clients to bring this new technology to bear and to develop the models that show significant promise in helping to make significant improvements in the understanding of healthcare quality issues. These capabilities will enable providers to understand and achieve optimal outcomes, enhance their quality processes, and develop an organization- wide awareness of quality and its impact on the both the lives under their care and the financial bottom line.

9/25/22, 4:40 PM Print

https://content.uagc.edu/print/McNeill.2947.17.1?sections=part03,ch11,ch12,ch13,ch14,ch15,ch16&content=all&clientToken=c8cf3d54-c96f-25b0-d… 16/27

14 Improving Patient Safety Using Clinical Analytics

Dean F. Sittig and Stephan Kudyba

Clinical analytics, coupled with the data extracted from real-time, point-of-care electronic health record systems, has the potential to significantly increase patient safety through the use of “triggers” or computer-based algorithms that automatically identify patterns in data that suggest errors have occurred in healthcare-related activities. This chapter describes research currently being conducted to identify three different types of errors in 1) medical diagnosis, 2) medication administration, and 3) use of electronic health records (EHRs). Using these new analytic measurement systems, researchers can better evaluate the effect of the myriad health information technology interventions on patient safety.

Introduction

Patient safety refers to the need for those involved in the delivery of healthcare to avoid harming the patients they are treating.1

(http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/ch14#ch14end01) Beginning with the Institute of Medicine’s landmark 1999 report titled “To Err is Human,” there has been an increasing emphasis on all aspects of medical errors, where medical error is defined as “the failure of a planned

action to be completed as intended or the use of a wrong plan to achieve an aim.”2

(http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/ch14#ch14end02) Even with this increased emphasis, there has been no reduction in the number or severity of medical errors as reported by clinicians or as identified after the fact by clinicians trained to read through a patient’s

medical record looking for evidence of error.3 (http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/ch14#ch14end03) These errors can occur in almost any aspect of the clinical work process, including diagnosis (e.g., creating an incomplete or incorrect list of potential diagnoses, ordering the wrong diagnostic tests, or incorrectly interpreting the test results), treatment planning (e.g., selecting the wrong therapeutic procedure) or administration (e.g., performing the wrong therapeutic procedure or performing the correct procedure incorrectly), and in failure to follow up on known clinical problems (e.g., failure to perform a needle biopsy on a patient with an abnormal mammogram).

Before an individual or organization can begin fixing the problems that lead to these errors, they must be able to detect that an error has occurred and investigate its underlying causes. Without significant progress in enhancing the understanding of the causes of medical errors, and reducing their number and severity, patients will continue to suffer from unnecessary patient harm as a result of erroneous or delayed diagnosis, incorrect or unnecessary medication administration or therapeutic procedures, for example.

Recently patient safety researchers have demonstrated that simply relying on clinicians to self-report these errors overlooks the vast majority of occurrences. Toward this end, researchers have begun to develop “triggers” or automated algorithms to identify abnormal patterns in laboratory test results, clinical workflows, or patient encounters. These triggers are easily computed using existing clinical analytic techniques applied to large clinical and administrative data warehouses.

Background Before an organization can develop an automated error detection system based on automated triggers, several key clinical analytic infrastructure components are required:

• An advanced electronic health record system (EHR) that captures information on all patient visits, clinical problems, medications, laboratory test results, and therapeutic procedures using a controlled clinical vocabulary. Simply recording all this information in a free- text note precludes any automated analysis without significant natural language processing (NLP) to identify specific data items.

• An offline clinical and administrative data warehouse that re-indexes the patients’ data to allow rapid retrieval of longitudinal patient information (e.g., find all patients with a cholesterol level greater than 240 milligrams per deciliter of blood) rather than the patient- centered view (e.g., what is patient X’s cholesterol level) that is required by clinicians to manage an individual patient. This offline data warehouse should also incorporate and integrate data from multiple external data sources (e.g., billing data, pharmacy dispensing data, geographic location, etc.) to allow researchers to better identify and isolate the patient’s clinical context.

• A set of clinically-tested algorithms or “triggers,” implemented as queries against the data warehouse, that can identify specific patterns that have a high probability of association with a medical error (e.g., an order for an antidote for example, Naloxone) to a medication

overdose (e.g., morphine).4 (http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/ch14#ch14end04) These algorithms need to be carefully tested to ensure that they have suitable sensitivity (i.e., identify all the important cases) and specificity (i.e., do not identify too many incorrect cases).

• A dedicated team of clinicians responsible for investigating all the “incidents” identified by the triggers. This team of clinicians must be able to quickly ascertain whether the identified individuals truly had the suggested condition and follow up on these findings in an appropriate manner.

Once an organization has all of these technical and social systems in place, they can begin to identify, investigate, report, and manage at least a subset of their clinical errors. The following sections describe three specific examples that illustrate the potential use of clinical analytics to identify various types of clinical errors.

9/25/22, 4:40 PM Print

https://content.uagc.edu/print/McNeill.2947.17.1?sections=part03,ch11,ch12,ch13,ch14,ch15,ch16&content=all&clientToken=c8cf3d54-c96f-25b0-d… 17/27

Triggers for Diagnostic Errors This diagnostic process begins with the initial data gathering phase in which the patient’s signs and symptoms are collected via interview, observation, and physical examination. It continues with the hypothesis generation phase in which the clinician compares the patient’s signs and symptoms to the medical knowledge base. These hypotheses are then tested in the diagnostic test ordering phase in which the clinician attempts to either confirm or rule out specific diagnoses through the judicious selection of tests of various physiologic processes (e.g., testing the patient’s ability to form a clot [necessary before one can safely perform surgery] by ordering coagulation tests—Prothrombin Time, Partial Thromboplastin Time, and International Normalized Ratio). The results of these tests help the clinician make the final diagnosis, which forms the basis for the various treatment options she must consider.

Diagnostic errors (i.e., missed, delayed, or wrong diagnosis5 (http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/ch14#ch14end05) ) can occur in any clinical setting and are arguably the leading type of error in primary care. Ambulatory practice-related diagnostic errors account for up to

40% of malpractice claims and result in average claims of $300,000.6 (http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/ch14#ch14end06) There are multitudes of ways in which these errors can occur, but the end result of these errors is that the patient does not regain his health and often returns to the health system a second or third time seeking help.

In an attempt to identify potential diagnostic errors in the ambulatory setting, regardless of their cause or position in the diagnostic process (e.g., faulty data collection, inadequate medical knowledge, erroneous or lost test results, etc.), Singh et al. developed Structured Query Language (SQL)-based queries to detect the presence of one of two mutually exclusive events: trigger 1) a primary care visit (index visit) followed by a hospitalization in the next ten days; or trigger 2) an index visit followed by one or more primary care, urgent care, or emergency department

visits within ten days.7 (http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/ch14#ch14end07) Follow-up evaluation of these triggers resulted in a positive error identification rate of 16.1% and 9.4% for triggers 1 and 2, respectively, with an error rate of 4% in cases that met neither screening criterion. Further investigation into the causes of these errors showed that the most common errors were

• Failure or delay in asking the appropriate questions of the patient (e.g., asking a patient with suspected tuberculosis whether she has recently traveled to Asia or sub-Saharan Africa) or ordering the appropriate diagnostic tests (e.g., x-ray or blood analysis)

• Misinterpretation or suboptimal weighting of critical pieces of data from the history and physical examination

• Failure to recognize urgency of illness or its complications.

Similar triggers have been used for diagnostic errors that occur in the hospital or its Emergency Department (ED), for example, patients who

return to the ED within 72 hours of hospital discharge8 (http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/ch14#ch14end08) or any patient who spends greater than 6 hours in the ED, with similar accuracy in error identification.

Triggers for Medication Errors Medication administration-related errors are common in all clinical settings as well. Medication errors can occur due to incorrect medications, doses, routes, time of administration, and even administering medications to the wrong patient. In an attempt to develop a tool to help

organizations begin to identify these errors, the Institute for Healthcare Improvement (IHI) developed a global trigger tool (GTT).9

(http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/ch14#ch14end09) Briefly, the GTT provides a standard methodology for reviewing patient records for triggers, or indicators, of potential adverse events. Each of these triggers is then reviewed by a clinician for final determination of whether an error actually occurred.

Assuming that an organization is using a computerized order entry system, the automatic detection of specific adverse events related to medication administration can be relatively straightforward using the GTT. Example triggers from this tool include looking for abrupt, unanticipated medication stop orders, abnormal laboratory results (e.g., Partial Thromboplastin Time (PTT) greater than 100 seconds in patients taking heparin), or use of an antidote medication (e.g., Diphenhydramine [Benadryl] administration). One can reliably identify errors with a sensitivity approaching 95% and a specificity approaching 100%.

Several hospitals have now automated the entire GTT and on a daily basis scan their entire hospital looking for potential adverse events.10

(http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/ch14#ch14end10) , 11 (http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/ch14#ch14end11) Using a list of all potential medication errors, a pharmacist reviews the medical records of all patients to confirm the errors. By looking at such a long list of potential errors, an organization can effectively identify problematic areas, processes, procedures, or even clinicians and begin addressing them.

Triggers for Electronic Health Record (EHR)-Related Errors In November 2011 the Institute of Medicine released a report titled “Health IT and Patient Safety: Building Safer Systems for Better Care” in which they highlighted the possibility that the health information technology (HIT) (e.g., electronic health records or barcode medication administration systems) itself could be a source of potential patient harm. This new type of “HIT-related error occurs anytime the HIT system is unavailable for use, malfunctions during use, is used by someone incorrectly, or when HIT interacts with another system component incorrectly,

resulting in data being lost or incorrectly entered, displayed, or transmitted.”12

(http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/ch14#ch14end12) While HIT is often seen as a solution to potential patient safety issues, clearly, we need methods of measuring and monitoring the systems themselves to ensure that they are safe for use.

9/25/22, 4:40 PM Print

https://content.uagc.edu/print/McNeill.2947.17.1?sections=part03,ch11,ch12,ch13,ch14,ch15,ch16&content=all&clientToken=c8cf3d54-c96f-25b0-d… 18/27

Toward that end, in 2008, Koppel et al. developed an automated trigger to identify potential medication ordering errors.13

(http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/ch14#ch14end13) Briefly, their trigger looked for all medication orders that were canceled by the ordering physician within 2 hours of their entry. They hypothesized that these “rapid discontinuations” represented orders that were subsequently determined to be suboptimal, or in error, in some way. To evaluate their system they reviewed all orders meeting this criteria over a 24-day period (total = 398). Upon review, two-thirds of the orders discontinued within 45 minutes were deemed inappropriate in some way (e.g., often these orders were entered on the wrong patient).

In another example of an automated trigger used to identify errors related to use of the EHR itself, Wilcox et al. were able to estimate the prevalence (or rate) of wrong patient notes (i.e., clinical notes judged to pertain to a different patient than the one they were stored under) in the

electronic medical record.14 (http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/ch14#ch14end14) Using their EHR, they identified cases in which the demographic information (i.e., patient gender) contained in the free-text portion of the progress note, differed from that contained in the coded portion of the patient’s record. They did not find a large percentage of mismatched patient records (e.g., approximately 0.5%), but the trigger is still important. They also realize this estimate represents only half of the likely errors, since wrong patient errors are just as likely in patient notes in which the gender matches.

Conclusion By combining the clinical and administrative information contained in state-of-the-art electronic health records with the newfound clinical analytics capabilities of modern data warehouses, clinical informatics researchers are poised to revolutionize the study of patient safety. These new computing systems with their automated algorithms are capable of sifting through thousands of patient records to identify potential clinical errors and systematically measure patient safety in ways never before anticipated. Once these measurement systems are in place, informatics researchers can better evaluate the effect of the myriad clinical decision support interventions currently under development. Only after all these new capabilities are in place and functioning as anticipated, can we expect to see the transformative improvements in patient safety that every patient deserves.

Notes 1 (http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/ch14#ch14end01a) . Kilbridge P. M., Classen D. C. The informatics

opportunities at the intersection of patient safety and clinical informatics. J Am Med Inform Assoc. 2008 Jul-Aug;15(4):397-407.

2 (http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/ch14#ch14end02a) . Kohn L., Corrigan J., Donaldson M., eds. To Err Is Human: Building a Safer Health System. Committee on Quality of Healthcare in America, Institute of Medicine. National Academies Press; Washington, DC; 1999.

3 (http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/ch14#ch14end03a) . Jha A. K., Classen D. C. Getting moving on patient safety—harnessing electronic data for safer care. N Engl J Med. 2011 Nov 10;365(19):1756-8.

4 (http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/ch14#ch14end04a) . Kahn M. G., Ranade D. The impact of electronic medical records data sources on an adverse drug event quality measure. J Am Med Inform Assoc. 2010 Mar-Apr;17(2):185-91.

5 (http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/ch14#ch14end05a) . Newman-Toker D. E., Pronovost P. J. Diagnostic Errors —The Next Frontier for Patient Safety. JAMA. 2009;301:1060–1062.

6 (http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/ch14#ch14end06a) . Chandra A., Nundy S., Seabury S. A. The growth of physician medical malpractice payments: evidence from the National Practitioner Data Bank. Health Aff (Millwood) 2005;(Suppl Web Exclusives):W5-240–W5-249.

7 (http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/ch14#ch14end07a) . Singh H., Thomas E. J., Khan M. M., Petersen L. A. Identifying diagnostic errors in primary care using an electronic screening algorithm. Arch Intern Med. 2007 Feb 12;167(3):302-8.

8 (http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/ch14#ch14end08a) . Nuñez S., Hexdall A., Aguirre-Jaime A.. Unscheduled returns to the emergency department: an outcome of medical errors? Qual Saf Healthcare. 2006 Apr;15(2):102-8.

9 (http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/ch14#ch14end09a) . Classen D. C., Resar R., Griffin F., Federico F., Frankel T., Kimmel N., Whittington J. C., Frankel A., Seger A., James B. C. “Global trigger tool” shows that adverse events in hospitals may be ten times greater than previously measured. Health Aff (Millwood). 2011 Apr;30(4):581-9. Erratum in: Health Aff (Millwood). 2011 Jun;30(6):1217.

10 (http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/ch14#ch14end10a) . Classen D. C, Pestotnik S. L., Evans R. S., Burke J. P. Computerized surveillance of adverse drug events in hospital patients. JAMA. 1991 Nov 27;266(20):2847-51.

11 (http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/ch14#ch14end11a) . Szekendi M. K., Sullivan C., Bobb A., et al. Active surveillance using electronic triggers to detect adverse events in hospitalized patients. Qual Saf Healthcare 2006. 15184–190.

12 (http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/ch14#ch14end12a) . Sittig D. F., Singh H. Defining health information technology-related errors: new developments since to err is human. Arch Intern Med. 2011 Jul 25;171(14):1281-4.

13 (http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/ch14#ch14end13a) . Koppel R., Leonard C. E., Localio A. R., Cohen A., Auten R., Strom B. L. Identifying and quantifying medication errors: evaluation of rapidly discontinued medication orders submitted to a computerized physician order entry system. J Am Med Inform Assoc. 2008 Jul-Aug;15(4):461-5.

14 (http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/ch14#ch14end14a) . Wilcox A. B., Chen Y. H., Hripcsak G. Minimizing electronic health record patient-note mismatches. J Am Med Inform Assoc. 2011 Jul-Aug;18(4):511-4.

9/25/22, 4:40 PM Print

https://content.uagc.edu/print/McNeill.2947.17.1?sections=part03,ch11,ch12,ch13,ch14,ch15,ch16&content=all&clientToken=c8cf3d54-c96f-25b0-d… 19/27

15 Using Advanced Analytics to Take Action for Health Plan Members’ Health

Stephan Kudyba, Thad Perry, and John Azzolini

“Hot spotting” has practically become the battle cry for managed care organizations since an article of the same name was published in The New Yorker in January 2011 (Gawande, 2011 (http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/ch15#ref08) ). In this article a physician, Jeffrey Brenner, realized after attempting to save a shooting victim it was possible to study healthcare utilization patterns in much the same way assault and crime patterns are tracked. For example, Brenner discovered that across a six-year period, more than 900 people in Camden, New Jersey, accounted for more than $200 million dollars in healthcare expenses. One of these patients had 324 admissions in five years while another patient cost $3.5 million. Additionally, he found that 1% of the patients who used Camden’s healthcare facilities accounted for 30% of the total costs.

Called the “Top 1%” by the managed care industry, Brenner’s observation was important not because it was new information (e.g., Stanton & Rutherford, 2005 (http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/ch15#ref17) ), but because it clearly articulated this situation to the general public. Patients like these, “hot spots,” were the ones in need of better quality of care. From his experience as a physician, he knew that those patients with the highest costs are usually the ones who experience the worst care. Though hardly a new idea, in less than a year since this article was published, the concept of identifying and intervening with the most severe or clinically complex patients to lower future healthcare costs has seen resurgence in the managed care arena.

State Medicaid agencies have embraced the hot spotting concept, challenging their Medicaid Managed Care Plans (MCPs) to decrease overall healthcare expenditures by increasing the quality of care for their member populations, especially those who are currently a significant cost burden to the program. As early as 2005, the Ohio Commission to Reform Medicaid (OCRM) published a report on Ohio Medicaid making recommendations and outlining action steps focused on (a) long-term care, (b) care management, (c) pharmacy management, (d) member eligibility, (e) financial administration, and (f) overall program structure and management (Revisiting Medicaid Reform, 2009 (http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/ch15#ref15) ). Recognizing the difficulty of healthcare reform, this report, and subsequent study, underscored the complexity of the relationships between quality, delivery, management, administration, cost, and payment of healthcare.

Clearly, quality of care and cost of care are on a collision course. Is it possible to develop a system that establishes and supports our nation’s need for quality, affordable healthcare, or has decades of fee-for-service payment systems independent of quality of care and/or patient outcomes made it impossible to reform heath care delivery? Unfortunately, as long as healthcare facilities, providers, and suppliers fail to work together and independently treat patients as profit centers, there appears to be little hope for healthcare reform.

This is one reason that the new healthcare reform law is so controversial. It is widely known that health plan members who receive the right type of care at the appropriate time have better health outcomes, thereby decreasing their cost burden on an already overstressed healthcare economic system. Proponents and critics of healthcare reform do agree on one thing—there are ways to decrease overall healthcare costs through population-based care management programs. Brenner asserted in The New Yorker article, “For all the stupid, expensive, predictive- modeling software that the big vendors sell, you just ask the doctors, ‘Who are your most difficult patients?,’ and they can identify them.” Though this statement has simplistic appeal, it cannot exist as a management strategy applicable to broad populations. The process of designing and running a population-based program is not as trivial as this statement implies.

To be effective, a data-driven, procedural approach must be taken, which allows for the coordination and collaboration of all constituents in the healthcare delivery process. Moreover, robust data management, data mining, and analytic processes are essential for successful population- based programs because improvements in quality and cost of care can only occur when the right members receive the right services at the right time (Cousins et al., 2002 (http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/ch15#ref05) ; Perry et al., 2004 (http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/ch15#ref12) , 2007 (http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/ch15#ref13) ). For these data-driven methods to influence clinical and financial outcomes of our healthcare delivery system, the information derived from them must be actionable.

Clearly, care management programs already exist to address these issues. However, one must consider the following questions: If there are already programs, protocols, and processes in place, why are we not seeing significant improvements in quality and cost of care? Why are health plans, facilities, and provider practices still challenged with improving quality of care while reducing overall costs? Therefore, we are not presenting new concepts in care management, but rather a conceptual framework comprised of factors influencing one’s overall healthcare experience. This combination of factors allows for the optimization of care management interventions that maximize the ability to target and identify those members whose cost burden will continue to increase in the absence of care management support activities.

To develop actionable care management information, it is necessary to create a framework that takes into account data derived from numerous sources. Expert systems created to support clinical decision support systems (CDSS), are relatively common in facilities and provider offices. Many of these systems are designed to improve (a) patient safety, (b) quality of care, and (c) efficiency of delivery (Coiera, 2003 (http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/ch15#ref03) ; Raghupathi, 2007 (http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/ch15#ref04) ). These expert systems are extremely helpful in clinical settings because they provide actionable information. That is, the information derived from these systems is directly applicable to improving the treatment of specific patients. If expert systems are successfully implemented at a provider/facility level, then it is reasonable to expect that expert systems can also be implemented at a health plan/managed care level.

9/25/22, 4:40 PM Print

https://content.uagc.edu/print/McNeill.2947.17.1?sections=part03,ch11,ch12,ch13,ch14,ch15,ch16&content=all&clientToken=c8cf3d54-c96f-25b0-d… 20/27

Actionable Information—A Conceptual Framework We propose developing a care management expert system that combines information from three healthcare cost and utilization domains:

• Current and predicted healthcare costs

• Utilization impact

• Member engagement

Using current and predicted healthcare costs, it is possible to identify those health plan members who are currently highest cost and will remain high cost in the absence of healthcare interventions. Since it is known that current high cost does not necessarily predict future high cost (see Ridinger & Rice, 2000 (http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/ch15#ref16) ), predictive modeling is necessary to determine those individuals with the highest probability of either remaining or becoming high cost health plan members. Without these analytic methods, it would not be possible to direct care management interventions to the right risk groups to significantly influence the overall cost burden of these health plan members. To simplify the challenge, it is helpful to visualize the relationship between current and future healthcare costs using a resource allocation matrix (e.g., Donaldson et al., 2002 (http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/ch15#ref07) ).

As shown in Figure 15.1 (http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/ch15#ch15fig01) , the relationship between current and future cost burden is an important factor in determining where to allocate resources. Simple predictions (i.e., those that do not require involved predictive modeling techniques) allow for the classification of members into either “good use” or “poor use.” Likewise, complex predictions (i.e., those that require involved predictive modeling techniques) do the same. It is, obviously, more difficult to determine those low cost members who will become high cost as well as those high cost members who will become low cost. By identifying those future high cost members, the correct resources can be allocated to the appropriate interventions needed to help control healthcare costs. Since the challenge of this care management strategy is to control future healthcare costs, the first step in creating actionable information is to classify the member population.

Figure 15.1 Resource allocation matrix

Knowledge Discovery through Multivariate Analytics Simple database queries can be utilized to search vast data sources to identify current high cost patients in a patient population (e.g., those incurring extensive health treatments, as mentioned in the New Yorker article). This information can be beneficial since resources can be applied to those particular cases to address those factors contributing to continued poor health maintenance. One limitation of this approach, however, is that it requires individuals to become high service utilizers or high cost before resources can be applied to alleviate or mitigate the problem. In other words, the limitation of this more hind-sight or reactive analytic method is that the focus is placed on what has already happened, where little knowledge is generated as to why individuals become high cost.

More effective cost reduction and health enhancement policies can be achieved through proactive analytics that identify likely future “poor health” or high cost candidates (Kudyba, 2005 (http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/ch15#ref10) ). Advanced analytical methods applied to robust data resources enable decision makers to identify individuals at risk of requiring extensive health treatments or high cost candidates. Methodologies such as logistic regression help identify noteworthy patterns existing in corresponding data. These patterns can include variables involving patient demographic and behavioral information, symptomatic and diagnostic data, and treatment related data to name a few. The resulting models identify not only the segments of a patient population that are likely high cost candidates but also provide the possible factors that lead patients to be high cost...or the “why” behind the high cost results.

Advanced analytic methods that incorporate a multivariate approach and the utilization of mathematical and algorithmic processing of data in conjunction with statistical testing techniques are often referred to as “knowledge discovery techniques.” The “knowledge” refers to the identification of patterns or relationships between variables in a particular data set that explain “why” things happen, not simply “what” has happened (see Figure 15.2 (http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/ch15#ch15fig02) ). Perhaps the most important information to extract from data with regards to the complexity of individuals’ health status includes patterns in dietary behavior, physical attributes, treatment and medication practices, and so on that can provide insights as to what combination of behavioral and descriptive variables lead to a less healthy, higher cost individual. This information is truly actionable as it empowers healthcare providers to more

9/25/22, 4:40 PM Print

https://content.uagc.edu/print/McNeill.2947.17.1?sections=part03,ch11,ch12,ch13,ch14,ch15,ch16&content=all&clientToken=c8cf3d54-c96f-25b0-d… 21/27

accurately apply available resources to mitigate costs and maintain a healthier population by implementing preemptive or proactive treatment to high risk candidates, thus mitigating costs before they incur. The result is not only reduced cost for providers but a healthier population.

The next step in the conceptual model involves determining the opportunity to impact a member’s healthcare utilization level. This step is multifaceted and requires the integration of both a member’s health status as well as their use of healthcare services. Health status refers to the current standing of an individual’s clinical, physical, and mental health (Kudyba et al., 2008 (http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/ch15#ref09) ). Health status is commonly determined through general health assessments (GHAs), health risk assessments (HRAs), clinical severity estimators (e.g., ACGs—adjusted clinical groups), as well as other sources of member-specific healthcare information (e.g., provider records, treatment notes, etc.). This type of information allows care managers to better understand if they can impact these members’ utilization behaviors, given their current health statuses. If a member suffers from a disease or condition that cannot be impacted by care management interventions, then decisions must be made as to the level and type of services and support they receive.

Figure 15.2 Conceptual model for predicting patient costs

Similarly, a member’s utilization history must be analyzed to determine the opportunity to modify his behavior. For example, it has been well documented that care management interventions can influence health plan members’ emergency department and hospital admission rates (Brandon, 2011 (http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/ch15#ref02) ). With the current market focus on emergency department diversion (EDD), readmission rates, and prescription drug rates, it is becoming increasingly important to integrate these sources of information. The development of enterprise data warehouses (EDW) at health plans is a direct result of the need to collect and use these data sources to track their members’ utilization behaviors. As described earlier from the “Hot Spots” article, when a health plan member has 324 admissions in a five-year period, one has to ask if this level of utilization was necessary. Consequently, the second step of this process involves an assessment of health status and utilization history, which further filters the membership population by taking those members who are predicted to be future high cost and identifying those who can be impacted by care management interventions.

Understanding and measuring member engagement is the final step in this conceptual model. Engagement can mean many different things, but in this context it refers to a member’s willingness to transmit, receive, and act upon customized communications. Engagement is emotional in nature; it often describes one’s connection with healthcare providers and predicts how actively members participate in the management of their own healthcare (Arnold, 2007 (http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/ch15#ref01) ). In other words, engagement involves members’ management of their own healthcare services to meet their healthcare needs. When members are engaged, they are actively involved and focused on their healthcare behaviors, resulting in better choices, better provider communications, and, ultimately, better outcomes (Protheroe, 2008 (http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/ch15#ref14) ).

Patient adherence to treatment regimens is a good example of why engagement is important in this conceptual model. For example, glycemic control is essential to individuals suffering from either Type I or Type II diabetes. This condition requires a high level of self-management; in other words, highly engaged individuals with diabetes have much better outcomes than those who are not engaged. An engaged individual will most likely adhere to the treatments prescribed by his providers (Delamater, 2006 (http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/ch15#ref06) ; Milano, 2011 (http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/ch15#ref11) ). Therefore, when members are defined by their future high cost, impact of care management interventions, and engagement in healthcare, the resulting actionable information increases the opportunity to provide the right members, the right services, at the right time.

9/25/22, 4:40 PM Print

https://content.uagc.edu/print/McNeill.2947.17.1?sections=part03,ch11,ch12,ch13,ch14,ch15,ch16&content=all&clientToken=c8cf3d54-c96f-25b0-d… 22/27

Conclusions There is no mistaking that population-based care management programs are difficult to develop, implement, and manage. A few of the more pervasive issues are as follows:

• Healthcare data sources are seldom standardized, often full of errors, and difficult to aggregate.

• Privacy and security regulations add additional complexity to data management efforts.

• Administrative claims data are collected on multiple systems and have varying file formats and record layouts.

• Health plan membership constantly changes, with members entering and exiting the program based on employment and eligibility status.

• Provider enrollment is extremely variable based on participation status, physical location, and professional specialty.

• The prevalence of chronic conditions continues to rise at an alarming rate.

• Treatment and drug regimens continually evolve as well as practice guidelines, evidence-based medicine, and medical technology.

• Healthcare reform brings new concepts such as health homes, accountable care organizations, and health insurance exchanges.

The preceding list only scratches the surface. However, as one contemplates current healthcare delivery challenges, a common theme surfaces. All of these issues have a singular dependency—actionable information.

Figure 15.3 (http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/ch15#ch15fig03) represents an elementary formulation of the conceptual framework proposed in this chapter. The input to this model (multiple data sources) continues to be an important topic by itself. Data management, data aggregation, and data warehousing of the multitude of disparate healthcare data sources must occur before any other activity can take place. The scope and magnitude of this effort cannot be overemphasized. In order to effectively manage health plan members at a population level, administrative claims data sources (e.g., medical claims, pharmacy claims, durable medical equipment and supply claims, laboratory claims, behavioral health claims, dental claims, home health claims, skilled nursing and nursing home claims, etc.), health and risk assessment data sources (e.g., general health assessments, health risk assessments, disease specific surveys, etc.), risk and grouping data sources (e.g., episode treatment groups, diagnosis related groups, adjusted clinical groups, etc.), as well as many other healthcare data sources must be combined in an understandable, functional way. Health plans and business intelligence vendors are working on this challenge, with varying degrees of success as they work towards building enterprise data warehouses (EDW). This is the healthcare industry’s biggest challenge because without clean, robust data, it is not possible to implement data-driven care management programs, not to mention all of the new processes and programs created by the Affordable Care Act (ACA).

Figure 15.3 A simple formulation of the conceptual framework

Once data are entered into the system, the process of identifying the best candidates for care management programs begins. Following the proposed framework, each of the three categories acts as a population filter. Once the target population is identified (e.g., members with asthma), current and predicted cost burden is calculated for each individual member. The result of this analysis is a ranking of all members with asthma (continuing with this example) from highest to lowest predicted cost. The next step involves assessing the estimated utilization impact of the care management program’s interventions on these members with asthma. Those members who are determined to be most “impacted” by the care management program are ranked higher than those who are not. These two sources of information (predicted cost burden and utilization impact) are then combined to produce a new risk ranking; this ranking takes into account those members with the highest predicted cost burden who also have the highest probability of utilization impact by the care management program.

Combining these two sources of information gives care managers not only the ability to identify those members with the highest predicted costs but also those members they will have the best opportunity to impact. Consequently, a member with high predicted cost with low utilization impact would be ranked lower than a member with a moderately high predicted cost with a high utilization impact. Using this methodology, the entire population of members with asthma can be ranked from highest to lowest risk based on predicted costs and utilization impact. As described previously, this actionable information allows care managers to make appropriate resource allocation decisions. Since it would be a poor use of resources to concentrate efforts on high cost members with low to no probability of utilization impact, this information would be valuable for program management.

Nevertheless, this proposed framework has one final filter—member engagement. Once the member population is identified and risk ranked by predicted cost and utilization impact, member engagement is assessed. It does not matter how high one’s financial risk might be or how impacted she could be by care management interventions if the member is not engaged in the process. If a member is not emotionally invested

9/25/22, 4:40 PM Print

https://content.uagc.edu/print/McNeill.2947.17.1?sections=part03,ch11,ch12,ch13,ch14,ch15,ch16&content=all&clientToken=c8cf3d54-c96f-25b0-d… 23/27

in the program, there is little chance that they will benefit from the care management interventions. Without engagement, care management does not work. Consequently, at a population level, using engagement as the last filter will further refine the information derived from the model, thereby increasing the chances that care managers are expending efforts on those members who will be high future cost, have high utilization impact opportunities, and are engaged in the care management process. This framework, comprised of three information filters, will identify those members who, if left unattended, will present the highest cost and utilization risk and are able to engage successfully in the program. This information is essential at a population management level because limited care management resources are available and correct resource allocation is paramount to program success.

Clearly, population management programs rely on data-driven methods to target populations, identify high risk members, allocate limited resources, and assess program outcomes. To increase quality of care while decreasing healthcare costs, care management programs must focus their efforts on intervening with the right members, with the right services, at the right time. Once data are transformed into actionable information, program successes are possible, resulting in positive care management utilization and financial outcomes. Possessing data is not enough—actionable information must be created and used by health plans to meet current and future challenges. The conceptual model proposed in this chapter is one example of how the thoughtful consolidation of information can assist in the correct allocation of resources in a population-based care management program.

To be effective, a data-driven, procedural approach must be taken, which allows for the coordination and collaboration of all constituents in the healthcare delivery process.

References Arnold, S. (2007). Improving Quality Healthcare: The Role of Consumer Engagement. Robert Wood Johnson Foundation Issue Brief. October 2007.

Brandon, W. (2011). Reducing emergency department visits among high-using patients. Journal of Family Practice. FindArticles.com (http://FindArticles.com) . 15 Nov, 2011.

Coiera, E. (2003) The Guide to Health Informatics. 2nd Edition: Arnold, London.

Cousins, M., Shickle, L., and Bander, J. (2002). An Introduction to Predictive Modeling for Disease Management Risk Stratification. Disease Management 5: 157-167.

Delamater, A. (2006). Improving Patient Adherence. Clinical Diabetes. 24(2): 71-77

Donaldson, C., Currier, G., and Mitton, C. (2002). Cost Effectiveness Analysis in Healthcare: Contraindications. British Medical Journal. BMJ; 325:891.

Gawande, A. (January 24, 2011). The Hot Spotters: Can We Lower Medical Costs by Giving the Neediest Patients Better Care? The New Yorker.

Kudyba, S., Perry, T., and Rice, J. (2008). Informatics Application Challenges for Managed Care Organizations: The Three Faces of Population Segmentation and a Proposed Classification System. Int. J. of Healthcare Information Systems and Informatics. 3(2): 21-31.

Kudyba, S., Hamar, B., and Gandy, W. Enhancing Efficiency in the Healthcare Industry, Communications of the ACM, December 2005.

Milano, C. (2011). Can Self-Management Programs Ease Chronic Conditions? Managed Care January 2011. MediMedia USA.

Perry, T., Tucker, T. et al. (2004). The Application of Data Mining Techniques in Health Plan Population Management. Chapter VII. IT Solutions Series: Managing Data Mining. Idea Group Publishing.

Perry, T., Kudyba, S., and Lawrence, K. (2007). Identification and Prediction of Chronic Conditions for Health Plan Members using Data Mining Techniques. In Data Mining Methods and Applications. New York: Taylor & Francis: 175-182.

Protheroe, J., Roger, A., Kennedy, A., MacDonald, W., and Lee, Victoria (2008). Promoting Patient Engagement with Self-Management Support Information: A Qualitative Meta-Synthesis of Processes Influencing Uptake. Implementation Science. 3:44.

Revisiting Medicaid Reform: A project of The Center for Community Solutions and the Center for Health Outcomes, Policy, and Evaluation Studies (HOPES) of Ohio State University. (2009). A Status Report on Recommendations from the Ohio Commission to Reform Medicaid Four Years Later: January 2005-January 2009.

Raghupathi, W. (2007). Designing Clinical Decision Support Systems in Healthcare: A Systematic Approach. International Journal of Healthcare Information Systems and Informatics. 2(1): 44-53.

Ridinger, M. and Rice, J. (2000). Predictive Modeling Points Way to Future Risk Status. Health Management Technology. 21: 10-12.

Stanton, M., and Rutherford, M. (2005). The High Concentration of U.S. Healthcare Expenditures. Agency for Healthcare Research and Quality. Research in Action, Issue 19. AHRQ Pub. No. 06-0060.

9/25/22, 4:40 PM Print

https://content.uagc.edu/print/McNeill.2947.17.1?sections=part03,ch11,ch12,ch13,ch14,ch15,ch16&content=all&clientToken=c8cf3d54-c96f-25b0-d… 24/27

16 Measuring the Impact of Social Media in Healthcare

David Wiggin

The intersection of social media and the U.S. healthcare system is rapidly expanding. The goal of this chapter is to present a snapshot of current and emerging uses of social media and take a step toward the development of an analytical model to measure the impact of social media on U.S. healthcare. This becomes most interesting if social media can play a measurable role in improving the health of the population.

Why Measure at All? When you can measure what you are speaking about, and express it in numbers, you know something about it, when you cannot express it in numbers, your knowledge is of a meager and unsatisfactory kind; it may be the beginning of knowledge, but you have scarcely, in your thoughts advanced to the stage of science.—Lord Kelvin

Lord Kelvin was on to something here. If we can quantify the value of social media, we can make informed decisions about social media investments. Today, social media investments are being made in healthcare to increase market share in competitive markets, improve quality of care, reduce the cost of care, and reduce health risk. Ultimately, the desired outcome of all these investments is to improve the health of the population. While we start by measuring first order impacts like market share and cost, we should seek ways to measure impact on population health.

Working Definition of Social Media in Healthcare A paper written by the Computer Services Corporation (CSC) in March 2012 offered this definition: “Social media is the process of people using

online tools and platforms to share content and information through conversation and communication.”1

(http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/ch16#ch16end01)

Most of us probably think about Facebook and Twitter first when we hear the phrase social media. However when we consider the impact of social media on healthcare, we need to consider the powerful combination effect of these channels with

• Blogs—Most often by physicians, sometimes by patient advocates

• Affinity group sites—Condition-specific or role-specific

• Reference sites—Dot-coms, crowd sourced sites, patient experience rating sites

While at first blush, reference sites may seem to not belong in this conversation, they are in fact a vital part of the social media landscape. So much of what happens in social media includes a link to content from another site. Twitter is a hundred times more powerful and interesting because tweets are link-enabled to content. The key point from this definition is that social media analytics must take into account the constellation of connected content.

The Complexity of Social Media and Healthcare Before we launch into a conversation about how social media is being used in healthcare today, let’s first consider the relative complexity of the topic.

If social media is the connection point of two or more parties, then the number and complexity of relationships is important to understand. Let’s first consider the intersection of social media and consumer packaged goods (CPG). There we have businesses (Proctor & Gamble, Unilever, Colgate-Palmolive), and we have consumers (you and me). In this case, the product-to-consumer relationship may be well understood. There are many channels connecting CPG companies to consumers, each with a measurable impact—distribution channels, advertising channels, and so on.

Healthcare is more complicated because more than two primary parties are involved, thereby generating more relationships, each in turn subject to connection via a variety of channels. If that weren’t enough, there is a lot of variety among healthcare patients/consumers, including varying conditions and goals, the complications of multiple conditions, many available provider options, regional variations in care delivery, and uncertainty in outcomes.

Each of these factors increases the complexity of the relationships. Let’s take a look at these complexity multipliers. They include

• Five major roles—patients, physicians (and other outpatient care), hospitals, payers (employers, health plans), and health IT

• 500 categories of chronic and acute conditions (episodes of care) with more granular detail available via 68,000 diagnosis codes (ICD-10 CM)

• Variations in the practice of medicine regionally, and by individual physicians (700,000) and hospitals (5,700) within a region

• Variations in patient behavior based on socioeconomic, cultural, and regional influences (compliance with treatment, lifestyle choices, and avoidable chronic disease)

It’s one thing to measure the social media impact of notable Facebook campaigns like those of American Express, Red Bull, Lacoste, and Petco. However, when it comes to the significantly more complicated industry landscape in healthcare, we should expect the analytics at the

9/25/22, 4:40 PM Print

https://content.uagc.edu/print/McNeill.2947.17.1?sections=part03,ch11,ch12,ch13,ch14,ch15,ch16&content=all&clientToken=c8cf3d54-c96f-25b0-d… 25/27

intersection of social media and healthcare to also be more complex.

Who Is Involved in Each Category of Social Media Use Today? As we continue to understand the complexity of healthcare social media (HCSM) analytics, Figure 16.1 (http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/ch16#ch16fig01) takes a step toward segmenting that complexity among interactions. It summarizes the number of roles (or nodes) in the social network by purpose of the connection.

Figure 16.1 Nodes in the social network by connection for HCSM

The colors categorize our readiness to measure the social media interactions described by the interaction. The check marks identify which role (stakeholder) is involved in that connection.

What Analytics Are Enabled Today and in the Future? In Figure 16.1 (http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/ch16#ch16fig01) , the first three categories of social media use and the analysis of these areas have the benefit of being comparable to work that has already been done in other industries. Marketing effectiveness, research, sentiment analysis, and brand management are all well-established disciplines. For example, in an article titled “Brands Ignore Negative Social Buzz at Their Peril,” eMarketer.com, reported “that 46% of US internet users it surveyed had turned to companies’ social

media sites to vent their frustrations about poor experiences.”2 (http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/ch16#ch16end02) One healthcare industry analyst commented that the healthcare industry would do well to learn from the retail and banking industries when it comes

to the use and analytics of social media.3 (http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/ch16#ch16end03) For the purposes of this chapter, we focus on social media uses that are unique to healthcare.

The fourth category in Figure 16.1 (http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/ch16#ch16fig01) , Provider Collaboration/Education, carries big potential in encouraging innovation and change in the physician community. In this use of social media, a growing collection of healthcare providers have formed a broad network of specialists, whose most advanced practitioners are sharing their findings and teaching others.

Social media provides clinicians with a convenient way to follow the work of specialists, researchers, and other peers, with an investment of just a few minutes a day. Using Twitter, they follow individuals (@David_Wiggin) and topics (#HCSM), and receive the benefit of articles, blog posts, and other materials that are germane to their professional lives. This will supplement, if not replace, having to wade through journals and online magazines and maintaining multiple e-mail subscriptions and feeds. They can follow the work of key opinion leaders and others whose views they trust. The result is a social network of clinicians who are continuously learning and sharing, and who have entered into a form of informal professional accountability. While much of this information is available via traditional sources, many find the nature of the interaction to be more manageable and helpful due to the shorter time between question and answer.

9/25/22, 4:40 PM Print

https://content.uagc.edu/print/McNeill.2947.17.1?sections=part03,ch11,ch12,ch13,ch14,ch15,ch16&content=all&clientToken=c8cf3d54-c96f-25b0-d… 26/27

To give you an idea about how vigorous this community is, there’s a Web site called twitterdoctors.net. Every hour, it provides an updated list of the most influential physicians on Twitter. The number of followers ranges from thousands to millions, and the number of tweets is in the thousands to tens of thousands.

Measuring the impact of this use of social media will be tricky. Measuring influence, as this Web site does, is interesting, but it doesn’t explain how influence correlates to the delivery of high quality, coordinated care, or how costs have been reduced based on participation. So while counting tweets is easy and has some relevance, the more meaningful measures would be longer term improvements in the quality of care delivered and patient outcomes. Any analytics that demonstrates the improvement in the health of the population via the use of social media would be valuable.

The last four categories shown previously in Figure 16.1 (http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/ch16#ch16fig01) are all focused on patient health. Each can play a role in transforming the health crisis.

The fifth category, Patient Education, is interesting because social media is another channel for distribution of relevant healthcare content, usually by providers. Articles, video, and other content are pushed via a social media channel. For example, if a provider can send a YouTube video link to her patients, it has the potential to encourage appropriate patient lifestyle choices and leverage the physician-patient interaction at the office visit. In a sense, this is the social media upgrade to the printed brochure that you used to take home from the doctor’s office. In addition to the office visit conversation, physicians are directing patients to this education content via e-mail, Facebook messages, and by blogging.

The purpose of social media in this context is targeted communication with the goal of changing patient behavior. The following summary of population health analytics acknowledges the importance of Web-based and mobile solutions, which is a category aligned with social media:

Population health is holistic in that it seeks to reveal patterns and connections within and between multiple systems and to develop approaches that respond to the needs of populations. It defines wellness as more than the absence of illness and seeks to improve physical, emotional, and behavioral fitness. Population health tactics include rigorous analysis of outcomes. Understanding population- based patterns are critical antecedents to addressing population needs. That is, data informs the selection of effective population health management strategies to prevent or diminish illness in the future. Increasingly these efforts include the integration of face-to- face clinical care, telephonic support through coaching and care management, as well as web-based and mobile solutions, which

support greater health literacy and consumerism.”4 (http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/ch16#ch16end04)

The best way to measure the effectiveness of this use of social media is through the use of patient surveys. If the goal is to measure the beneficial impact of patient education, it will be important to hear from the patients. Only a percentage of patients who received brochures read them, and even a smaller percentage followed the guidance in the brochure, yielding a health benefit. The same is true with social media. Just because a link was sent or a video was watched doesn’t mean that education and lifestyle change took place.

The sixth category, Patient Affinity Groups, forms a leveraged crowd-sourcing opportunity for consumers to make informed choices about treatment and to learn from others who have been diagnosed with a condition. Further, when they achieve critical mass (enough participants) they collect self-reported patient data about what works and what doesn’t to form a growing self-reported comparative effectiveness study. This is in fact the business model for several Web companies today who manage these sites.

www.patientslikeme.com (http://www.patientslikeme.com) is a great example of a social network tool in this category. Here’s their story:

PatientsLikeMe was co-founded in 2004 by three MIT engineers: brothers Benjamin and James Heywood and longtime friend Jeff Cole. Five years earlier, their brother and friend Stephen Heywood was diagnosed with ALS (Lou Gehrig’s disease) at the age of 29. The Heywood family soon began searching the world over for ideas that would extend and improve Stephen’s life. Inspired by Stephen’s experiences, the co-founders and team conceptualized and built a health data-sharing platform that they believe can transform the way patients manage their own conditions, change the way industry conducts research and improve patient care.

So far, more than 150,000 people have subscribed and are able to learn about the experiences of others in managing symptoms, understanding the progression of disease, observing the effectiveness of various treatments, and finding moral support and encouragement from others who suffer from the same condition.

These sites also match patients with clinical trials, supporting both research and healing.

This is another case where it would be complicated to quantify and appropriately attribute the value of participating in this site across the whole participant pool. There will be many anecdotal stories that will continue to encourage participation, but isolating the benefit of this site to a population health measure will be difficult.

The seventh category, Patient Monitoring, takes advantage of the latest inventions and IT gizmos to help each of us stay on track in our lifestyle choices. The relatively new quantified-self movement grew out of the idea that if we could more easily track our behavior and even (optionally) join others in their quest to improved health, improved health would result.

Two great examples of this technology are the fitbit and Nike’s fuelband. Each records biometric information and transmits to your smartphone, providing you with data about exercise, calories burned, sleep, and so on. When combined with self-reported data like calorie consumption, individuals make a more informed choice about whether to snack at 2:00 p.m. and/or whether to have dessert that evening.

The potential for analytics on all this sensor data is great, and on an individual basis, this is included in the free smartphone app when you buy the $99 fitbit or similar device. Beyond the individual data, in the not too distant future, could we envision sharing some of these data with our

9/25/22, 4:40 PM Print

https://content.uagc.edu/print/McNeill.2947.17.1?sections=part03,ch11,ch12,ch13,ch14,ch15,ch16&content=all&clientToken=c8cf3d54-c96f-25b0-d… 27/27

accountable care organization (ACO) or health plan in exchange for a lower cost plan? From a patient/member perspective, this could be like the discount we receive for voluntarily filling out the Health Risk Assessment or placing a GPS equipped device in our automobiles to earn a lower safe driver rate from our auto insurance company.

This category of use makes the social media category because of the significant social dimension to the use of these technologies. Users are encouraged by the manufacturers and their peers to share their progress and compete with others, both of which fuel continued use and healthy lifestyle changes. This allows users to sign up for peer pressure and accountability that has a fun dimension.

The eighth category, Care Management, and its more focused predecessor, disease management, has long been recognized for its value in providing extra support to patients to ensure that they receive the care they need. Social media has made it possible for payer and provider organizations to interact with their members/patients on more channels, with more tailored communications.

Care management started with phone calls and letters in the mail. With broad adoption of home computers and cell phones, these were replaced with e-mail reminders and text messages. The Association of Managed Care just published a study touting the effectiveness of a text message

program to encourage patient compliance in medications and follow-up visits.5

(http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/ch16#ch16end05) Some providers have started to contact their patients via Facebook. Last

summer, the University of Iowa Children’s Hospital launched a program to improve medication adherence on Facebook.6

(http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/ch16#ch16end06)

Analytics on the cost/benefit of care management programs are mature and well understood today. The use of social media channels is a natural extension of earlier programs, and should plug into the existing analytics framework easily.

Conclusion The use of social media in the context of healthcare is growing and shows no signs of slowing down. It has been the topic of many conferences this year (Duke, NY Medical College, Mayo Clinic, etc.), and it is a promising collection of technologies that form a complicated, yet effective, channel that is accessed by all stakeholders.

However, the ability to attribute benefits to specific uses of social media varies from situation to situation. While we will always seek to do that, it’s important to realize that the value of social media may only be detected via patient survey (usage survey, satisfaction survey) or more importantly in measurements of the improvement of population health.

The best practice approach to analytics of this type continues to be collect detail data from relevant sources as it is generated, integrate the data for a comprehensive view of the patient, and put the data to use immediately via the appropriate channel(s)—social media in this case.

While it will be difficult to measure the long-term value of a social media investment with any precision, the short-term business value of these investments should be measured using traditional business metrics like market share (new customers, customer retention), sentiment analysis, and customer satisfaction.

Notes 1 (http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/ch16#ch16end01a) . “Should Healthcare Organizations Use Social Media? A

Global Update,” CSC, March 2012, p. 4. http://assets1.csc.com/health_services/downloads/CSC_Should_Healthcare_Organizations_Use_Social_Media_A_Global_Update.pdf (http://assets1.csc.com/health_services/downloads/CSC_Should_Healthcare_Organizations_Use_Social_Media_A_Global_Update.pdf) .

2 (http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/ch16#ch16end02a) . “Brands Ignore Negative Social Buzz at Their Peril,” eMarketer.com, July 16, 2012.

3 (http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/ch16#ch16end03a) . Janice Young, “The Healthcare Consumer We Hardly Know,” IDC Health Insights, April 4, 2012.

4 (http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/ch16#ch16end04a) . Raymond Fabius, MD, CPE, FACPE; Linda MacCracken, MBA; Jill Pritts, MS, “Vocabulary of Healthcare Reform,” January 2012, p. 39. Thomas Reuters. www.ahip.org/WP-VocabHCReform/ (http://www.ahip.org/WP-VocabHCReform/) .

5 (http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/ch16#ch16end05a) . Henry H. Fischer, MD; Susan L. Moore, MSPH; David Ginosar, MD; Arthur J. Davidson, MD, MSPH; Cecilia M. Rice-Peterson, RN, BSN; Michael J. Durfee, MSPH; Thomas D. MacKenzie, MD, MSPH; Raymond O. Estacio, MD; and Andrew W. Steele, MD, MPH, MSc, “Care by Cell Phone: Text Messaging for Chronic Disease Management,” Am J Manag Care, 2012;18(2):e42-e47.

6 (http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/ch16#ch16end06a) . Ken Terry, “Many Doctors Don’t Take Social Media Beyond Marketing,” Informationweek.com, April 10, 2012. http://www.informationweek.com/healthcare/patient/many-doctors- dont-take-social-media-beyo/232900043 (http://www.informationweek.com/healthcare/patient/many-doctors-dont-take-social-media-

beyo/232900043) .