3

Healthcare System Complexities, Impediments, and Failures

INTRODUCTION

The extent to which health care for Americans is timely, efficient, and appropriate for a given individual is determined by the characteristics of the delivery system. Moving to a learning healthcare system will require the identification of specific areas where system complexities slow or inhibit progress and the development of solutions geared toward overcoming impediments and failures.

Workshop discussions considered a number of process inefficiencies, structural barriers, and system failures that are significant impediments to quality and that preclude the delivery of highly effective, highly efficient, evidence-based health care. In the second workshop session, the focus turned to the areas of underperformance that may need the most attention and correction from an engineering perspective. Presenters in this session examined select obstacles inherent in multiple healthcare system components and certain flawed processes that particularly affect the generation and application of evidence. One goal of the session was to frame suggested ideas for how systems engineering might address some of health care’s most troublesome shortfalls.

This chapter begins with an overview of the healthcare culture. In his presentation William W. Stead, chief information officer of Vanderbilt University Medical Center, described the current healthcare environment as being characterized by competition, misaligned incentives, and inherent distrust among stakeholders. Throughout health care, Stead sees competing cultures at loggerheads—as exemplified by the tensions among consum-



The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement



Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.

OCR for page 117
3 Healthcare System Complexities, Impediments, and Failures INTRODUCTION The extent to which health care for Americans is timely, efficient, and appropriate for a given individual is determined by the characteristics of the delivery system. Moving to a learning healthcare system will require the identification of specific areas where system complexities slow or inhibit progress and the development of solutions geared toward overcoming im- pediments and failures. Workshop discussions considered a number of process inefficiencies, structural barriers, and system failures that are significant impediments to quality and that preclude the delivery of highly effective, highly efficient, evidence-based health care. In the second workshop session, the focus turned to the areas of underperformance that may need the most attention and correction from an engineering perspective. Presenters in this session examined select obstacles inherent in multiple healthcare system compo- nents and certain flawed processes that particularly affect the generation and application of evidence. One goal of the session was to frame suggested ideas for how systems engineering might address some of health care’s most troublesome shortfalls. This chapter begins with an overview of the healthcare culture. In his presentation William W. Stead, chief information officer of Vanderbilt University Medical Center, described the current healthcare environment as being characterized by competition, misaligned incentives, and inherent distrust among stakeholders. Throughout health care, Stead sees competing cultures at loggerheads—as exemplified by the tensions among consum- 11

OCR for page 117
118 ENGINEERING A LEARNING HEALTHCARE SYSTEM ers who want high service and low out-of-pocket costs, payers who want to select risk and limit cost, and purchasers who want more value at the lowest cost. Looking to a future that will be defined by individualized medicine, Stead suggested that tomorrow’s opportunities may not be fully realized without fundamental changes in the healthcare culture. Education for health professionals is only one area that needs reform. Another require- ment will be to move from the business of managing episodes of care to the business of caring for patients and populations. He added that similar fundamental reforms will need to be engineered into the business models of virtually every healthcare stakeholder—in payment mechanisms, and, notably, in the role of the individuals in managing their own care. Speaking from her perspective as a cardiologist and health policy ana- lyst, Rita F. Redberg, director of Women’s Cardiovascular Services at the University of California, San Francisco, noted that a marked proliferation in new diagnostic and treatment technologies has resulted in a precipitous increase in healthcare costs. Moreover, limited integration in the design of systems for health information technology (HIT) and technologies such as imaging systems has allowed their misuse and overuse, thus impeding their ability to improve healthcare quality. Redberg surveyed the current landscape of diagnostic and treatment technologies available for heart disease and offered suggestions for systemically evaluating and using these technologies in ways that improve care and reduce costs. She proposed that more systematic data collection and the development of more prospective registries would lead to better-informed decisions in health care. Addressing a concern that was raised throughout the workshop about the need for more robust data collection and mining capacities, Michael D. Chase, associate medical director of quality, Kaiser Permanente Colorado, asserted that the U.S. healthcare system has not fully leveraged clinical data to improve health outcomes. Impediments to full use of the data include limited data access, a problem that is exacerbated by inadequate adoption of electronic health records (EHRs) and lack of data standards. As health care has become more complex, the lag in the sophistication of data applications in evidence generation has become more acute. Engineer- ing principles, Chase suggested, could help those in charge of health care manage various complex processes and increase the use of data for clinical decision support. Chase offered examples and suggestions concerning how key delivery systems could be better integrated into healthcare systems in order to address critical areas in health care. For example, Chase proposed a patient-centered, population health–based view grounded in the principle of getting the right information to the right member of the healthcare team—including the patient—at the right time during the workflow or decision-making process. Chase presented a model that takes a broad look at decision support opportunities across a continuum of patient needs,

OCR for page 117
119 HEALTHCARE SYSTEM COMPLEXITIES, IMPEDIMENTS, AND FAILURES available healthcare professionals, tools and systems, and an extended time line for patient care. Amy L. Deutschendorf, senior director for clinical resource manage- ment at Johns Hopkins Hospital and Health System and principal of Clini- cal Resource Consultants, also observed that there has been an escalation in system and patient complexities throughout the current healthcare environ- ment. The crush of information, a plethora of new technologies, increased regulatory oversight, an aging population, and heightened consumer aware- ness and expectations have all contributed to the disorganization, frag- mentation, and discontinuity of patient care. Consequently, she argued, effective care coordination and linkage have become even more important. Deutschendorf spoke of the need for processes that ensure patient-centered alignment of care. One application is a care delivery process with communi- cation models and systems that can ensure the accurate and timely transfer of patient information throughout the healthcare continuum. Deutschendorf suggested a number of other changes, including more clarification, defini- tion, and distinctions between acute patient care and ambulatory care; bet- ter management of consumer expectations; and increased communication and collaboration between caregiving team members. Because models of care need to be based more firmly on evidence, she proposed that rigorous research be conducted to determine which care delivery models can yield appropriate safety outcomes and the highest possible quality outcomes. Speaking from his perspective as chief executive officer (CEO) of the University of Pennsylvania Health System (UPHS), Ralph W. Muller dis- cussed areas of successful transformation in administration and business systems at his institution. He highlighted projects on patient registration, billing, and revenue cycle management, and he discussed how each was transformed in order to be more effective. He also described a project that examined how UPHS inpatient and outpatient operations were improved through a combination of systems analysis, reporting systems, incentive alignment, and continuous change management. In discussing lessons learned in several areas of day-to-day practice—as well as from significant, documented results—Muller illustrated how engineering-specific interven- tions can change systems of care. In recounting examples of reform at UPHS, Muller also highlighted elements of a methodology for conceptual- izing change in the face of entrenched health cultures. He offered specific lessons learned about using data and analysis to identify opportunities and motivate change, redesign workflows and restructure roles, integrate infor- mation technology, establish goals and monitor performance, and create meaningful incentives. The final speaker in the second session, Eugene C. Nelson of the Dartmouth–Hitchcock Medical Center, said that we will need a healthcare system information environment that provides critical knowledge that can

OCR for page 117
120 ENGINEERING A LEARNING HEALTHCARE SYSTEM be used to effectively manage individuals over time, evaluate and improve the quality and value of clinical practice, and facilitate basic translational and outcomes research. Nelson described a successful transformative activ- ity at the Dartmouth–Hitchcock Spine Center that designed, tested, and re- fined patient-centered “feed-forward” and “feedback” data systems, which are built into the flow of healthcare delivery in order to support patient care and generate information and knowledge concerning entire patient populations. Nelson detailed the issues and concerns that motivated the project, discussed the challenges of designing the systems, and described their positive impacts on system effectiveness and patient satisfaction. He also outlined a promising approach for creating sustainable feed-forward data systems based on the formation of “collaboratories,” or professionally organized networks for advancing health care and healthcare research. HEALTHCARE CULTURE IN THE UNITED STATES William W. Stead, M.D., Vanderbilt Uniersity Medical Center This paper begins with three observations about the culture of health care in the United States. First, that culture is centered on individual expert health professionals; their behaviors reflect the way they are selected, the way they are educated, and what it takes to survive in their work environ- ment. These cultural roots of the health professions must be addressed if change in health care is to be realized. Second, the culture of health care in this country is one of a clash among competing forces. Stakeholders work against each other to obtain advantage for themselves at the expense of others. If we are to achieve meaningful improvement, this competitive clash needs to be transformed into a competition to work together to achieve the right results for the patient. Third, today’s health care faces discontinuous, disruptive change. The way health professionals make decisions will not scale up to handle the data load that is resulting from biological discover- ies in genomics, proteomics, and other areas. This last observation is good news. As the health professions and other stakeholders realize that they cannot escape disruptive change, we will have a once-in-a-century chance to test better approaches to health care. Building on these observations, this paper contrasts the current healthcare culture with a future culture in which care is delivered through systems approaches. The Culture of the Health Professions The culture of the health professions is rooted in their education. In the first phase of that education, the scientific basis of health and disease and the scientific method are taught. The goal is for each professional to have

OCR for page 117
121 HEALTHCARE SYSTEM COMPLEXITIES, IMPEDIMENTS, AND FAILURES a current fact base and to know the method by which facts are discovered. This phase of education is preparation to act on what is known, interpret new literature, and learn from practice. By way of analogy, at the end of this phase, students have learned how the car works and how it is built, but they have no idea how to plot a path from point A to point B. In the second phase of education, students learn practice through an apprentice- ship model in which they are mentored by a variety of individual experts. To continue the analogy, in this phase students learn the many ways to use the car to get from point A to point B and which ways work best. The third phase of education extends throughout the career as learning continues through practice and reading. If something unusual is seen in a patient or something new is tried on the chance that it might work, case reports are written to share observations. When the effects of alternative approaches are sought, a trial is conducted and the results written up. However, learn- ing remains individual. Each health professional seeks to be the best expert at caring for the cases he or she sees. The culture of the health professions is influenced by the way decisions are made. The reasoning of health professionals, because they are experts, takes place through the recognition of patterns. A person with fever, cough, infiltrate on a chest X-ray, and an elevated white count is suspected of hav- ing pneumonia, while a low white count causes concern that the immune system is overwhelmed. These conclusions are based on the entire picture, in much the same way that a constellation in the night sky is recognized. There is no systematic processing of data and calculation of combinatorial probabilities as is done by a novice in a learning situation. In addition, the data used to make decisions are imprecise. Many measurements used in clinical practice are correlative measures, not direct measurements of the substance itself. For example, nephrologists used to measure serum creati- nine, an indicator of renal function, by the light absorption of a compound formed by the adduct formation between creatinine and the picrate ion. Other compounds were absorbed at the measured frequency, causing falsely elevated measures. At a time when the sensitivity of the test was ±0.3, the threshold for treating transplant patients for rejection was a change of 0.3. In other words, physicians erred on the side of treatment with a toxic drug because treatment had to be started early to save the transplant. That kind of reasoning was used regularly, in the face of uncertainty, in life and death situations, under an oath that says “do no harm.” The culture of the health professions has also been shaped by the expo- nential increase in biomedical knowledge and technology. This overload is handled through specialization and subspecialization. In the process, some are learning more and more about less and less, while the rest are learning less and less about more and more. The workflow requires large amounts of multitasking, is interruption driven, and is nontransparent. There is no

OCR for page 117
122 ENGINEERING A LEARNING HEALTHCARE SYSTEM chance to sit and reflect. Compensation models reward piecework, proce- dures, and technology. Health professionals do their best to deliver excep- tional care despite the “system.” Time is the most limited resource. The combination of these internal roots and external pressures has led the culture of the health professions to become one in which circumstances that conflict with quality health care are accepted. Variability in practice is accepted as well. The best experts are sought out and expected to disagree. What other industry would report success if there were a shift in perfor- mance on a recommended practice from 60 to 80 percent of cases? If 5 practices need to be followed for each patient with a condition, and each is performed correctly 80 percent of the time, the probability that all 5 will be done correctly for a given patient is just 33 percent. The health profes- sions’ culture accepts process improvement targets that are far lower than necessary to have the desired effect on clinical outcomes. Autonomy is a goal of training. Challenges from those lower in the hierarchy are not acceptable. The conditions under which health profes- sionals function lead to increased self-confidence and cynicism (Gray et al., 1996). The fragmentation in care results in less of a sense of responsibility. Although everyone knows the healthcare system is broken, each individual believes his or her own practice is quite good. Data showing the variability in practice are met with surprise. By and large, health professionals are passionate about doing the right thing and are attempting to provide care for patients despite the system. Most of the time, they do a good job. The trouble is that most of the time is insufficient to avoid the quality problems that are ubiquitous in health care. The Clash Among Competing Forces The culture of the health professions is just one of many cultural chal- lenges to achieving better health care. The healthcare system in the United States is a clash among competing forces; it is not a system. Health profes- sionals, for example, focus on payment for services and autonomy. Care facilities seek high-margin services and low supply costs. Suppliers focus on intellectual property protection and volume. Meanwhile, consumers seek accessible services and low out-of-pocket costs. Payers pursue the right to select risk and limit cost. Purchasers want more value at the lowest cost. As Porter and Teisberg (2006) point out, the different stakeholders compete in a zero-sum game. The only way a payer can reduce costs for a purchaser, such as an employer, is to negotiate with the provider to take less or force the consumer to receive less. Because employers are working out- side of the direct care process instead of improving that process, they add administrative overhead. As the other stakeholders respond, the increase in

OCR for page 117
12 HEALTHCARE SYSTEM COMPLEXITIES, IMPEDIMENTS, AND FAILURES overhead is compounded, and the system becomes more expensive and less workable for the patient. This clash among stakeholders raises several cultural barriers to qual- ity health care. Incentives are not aligned. Providers are paid more if they overuse resources and if they provide poor care leading to rework. They are paid less if they provide such good care that other care is not necessary. They are paid more for technical and episodic tasks and little for cogni- tive, coordinative work. Healthcare CEOs have limited power given the autonomy of health professionals and the competition among hospitals for physicians. The stakeholders distrust each other. Although individuals trust their own physicians, they do not trust the “system” (Norris, 2007). They are the ball in the healthcare ping-pong match. They are forced to change health plans regularly as employers and government seek to control costs. A Medicare beneficiary sees a median of two primary care providers and five specialists per year, and Medicare beneficiaries with multiple chronic diseases see up to 16 health professionals (Pham, 2007). The culture of health care accepts waste. In his keynote address, Brent C. James outlined the data. Administrative overhead in U.S. health care may be as high as 40 percent. Thirty percent of the care provided may be unnecessary; as much as 70 percent may be preventable. Given the rapidly escalating cost of health care, tension exists over the cost of new technology, which has accounted for half of that increase in recent decades. Can we afford ever better technology? Does the increased cost of health care hurt the economic competitiveness of the country by increasing the cost of everything we do? Finally, the culture accepts poor outcomes on a population basis. In the United States, 109 deaths per 100,000 patients each year are attributable to health care, as compared with 65 in France (Nolte and McKee, 2008). Yet France’s per capita healthcare spending is about half that of the United States. Toward a New Healthcare Culture Even if today’s health care provided acceptable quality and access at an affordable cost, the healthcare culture would face disruptive, discontinuous change because of the inevitable demise of expert-based practice (IOM, 2009a). Cognitive research shows that a human can handle from five to nine facts in a single decision (Miller, 1956). Even with today’s clinical de- scriptions of phenotype, the number of facts bearing on a decision already can exceed this capacity, contributing to the overuse, underuse, and misuse of medical care. The additional data from structural genetics will probably push us into the range of ten facts per decision. Full data on a person’s

OCR for page 117
12 ENGINEERING A LEARNING HEALTHCARE SYSTEM functional expression may create a ten-fold increase in the facts per deci- sion, and data on proteins may add a second ten-fold increase. Imagine a primary care provider trying to cope with such a massive amount of data in a 15-minute encounter. Clearly a new paradigm for clinical decision making will be necessary. This inescapable change will create a once-in-a-century chance to rethink roles—and therefore culture—in health care. Table 3-1 contrasts the current culture with a possible future culture in which systems approaches to health and health care are used to deliver the desired results every time. In the current culture of a clash among forces, people attempt to fix the nonsystem by layering fix on top of fix from the outside. Each fix adds complexity and cost without changing the funda- mentals of care delivery. The goal should be a future culture in which the system is continuously refined from the inside out. In this culture, people are recruited and educated to know their limits, to trust the system and their teammates, and to expect perfect collective performance or correction with each failure. Care is coordinated around populations, and the care deliv- ered is right for the individual through systematic use of evidence (IOM, 2009b). Each individual is a data point in a population database. Providers are taught to practice in multidisciplinary, high-performance teams, using simulation to perfect their skills and outcomes to guide course corrections (IOM, 2007). Coordinated care is paid for and, on the basis of the value, delivered. In the process of shifting toward this vision or other possible futures, health professionals must strive to preserve the best of the current culture. Most people engaged in health care are passionate about what they are do- TABLE 3-1 Comparison of Current and Possible Future Healthcare Cultures Current Culture Future Culture • Layer fix on fix from outside • Improve from the inside out • Trust oneself; provide care despite the • Know one’s limits; trust the system and system one’s team • Care safe for the masses • Right care for the individual • Manage episodes of care • Care for populations and the patient as a whole • Expert-mediated use of evidence • Systematic use of evidence • Each patient is an experiment with • Each patient is a data point in a n=1 population • Learn in disciplinary silos • Learn in teams • Learn by applying science through • Learn from simulation and outcomes practice • Pay for piece work and process steps • Pay for coordination and outcomes

OCR for page 117
12 HEALTHCARE SYSTEM COMPLEXITIES, IMPEDIMENTS, AND FAILURES ing and about being in health care. Every day, in every hospital or clinic, there are people who go far out of their way to help their patients, despite the ecology in which they work. That passion must be preserved. At the same time, changes must be made to roles, education, decision-making pro- cesses, payment structures, and the way success is measured—in short, to the professional and business models of every stakeholder in the system. DIAGNOSTIC AND TREATMENT TECHNOLOGIES Rita F. Redberg, M.D., M.Sc., Uniersity of California, San Francisco A marked proliferation of new diagnostic and treatment technologies has resulted in a precipitous increase in the costs of health care. Moreover, despite the potential of these technologies to improve the quality of health care, the limited integration in system design for such technologies as HIT; laboratory, radiology, and imaging systems; and monitoring and surgical equipment has allowed their misuse and overuse. This paper surveys the current landscape of diagnostic and treatment technologies available for treatment of heart disease and examines how they might be evaluated and employed more systematically to improve care and reduce costs. In the late 1970s, John Eisenberg and Sankey Williams at the University of Pennsylvania were studying the behavior of the house staff with the goal of changing their routine daily lab test ordering for inpatients. However, Eisenberg and Williams’s daily reminders to the house staff to order only those tests that would affect patient management were not successful in re- ducing the number of daily lab tests ordered. It was difficult to be criticized for ordering too many tests as one could also be criticized for omitting a potentially useful test. All of the incentives in medical training lean toward ordering more tests, and how the additional information improves patient care receives little consideration. This philosophy is ingrained in the culture and reinforced by patient demands and the public’s perception that more care means better care. At the time of the study, healthcare expenditures were on the order of 8 percent of the U.S. gross domestic product (GDP), and everyone expected that if healthcare expenditures reached 10 percent of GDP, things were going to change. Yet today, 30 years later, healthcare expenditures are at about 17 percent of GDP, the Medicare Trustees Report predicts that Medi- care will be insolvent by 2012, and people are still speculating about when things are going to change. At least there is now some cause for optimism that some meaningful changes will take place that will lead to healthcare resources being spent more wisely. This paper examines what factors might drive such changes. The focus is on four of the main drivers of healthcare

OCR for page 117
12 ENGINEERING A LEARNING HEALTHCARE SYSTEM costs: demographics, limited quality measures, the third-party payment system, and technology growth. In terms of demographics, as we live longer we become victims of our success. The population includes more older people, who, on average, make more intense use of healthcare resources than do younger people. At the same time, quality measures are limited, and it is quite challenging to measure and reward good-quality care. The result is a massive healthcare system in which some of the care is of good quality and some of bad qual- ity. Additionally, the third-party payment system insulates some of the main drivers of healthcare costs (patients and physicians) from the actual cost of care. When one enters a store to make a purchase, the cost is clearly marked, and one can judge the value of the item relative to one’s budget. In health care, by contrast, the cost to the consumer is generally unknown, and out-of-pocket costs are not related to the actual cost of care and often not related to the patient’s own consumption of care. Of course, health care is a different kind of commodity from such purchases as appliances. However, a system in which copayments are the same for a very expensive and a very inexpensive test encourages increased consumption of health care without consideration of value. Generally, patients who receive a great deal of health care pay no more than those who receive only a little. A similar situation exists at the physician level. When our hospital’s house staff is asked about the prices of the tests they order in the context of a discussion about why they are ordering a test and how the patient is going to benefit from its results, physicians rarely know what the tests cost. In an academic medical center, the costs of testing and new technology are invis- ible because doctors are removed from the payment system and insulated from the cost of health care. Similarly, house officers are often shocked to learn of the difference in cost between the latest fourth-generation antibiotic and older generics. Of all the factors that drive up healthcare costs, however, the growth of technology can be singled out as most significant. Technology, of course, has many benefits. Numerous examples exist of advances in technology that have led to great improvements in health care. However, before a new technology is embraced, a technology assessment should be performed to determine whether it will yield actual patient benefits that outweigh any possible risks. This point is best illustrated by randomized controlled trials. The current healthcare system does not emphasize the need for evidence of benefit before widespread diffusion of new technology. Today we are seeing a rapid proliferation of technologies for both diagnosis and treatment. A major example is imaging, whose rates have increased dramatically in the past few decades. For example, cardiac im- aging used by cardiologists has increased by 24 percent per year over the past decade. Looking just at Medicare data from 1999 to 2003, cardiac

OCR for page 117
12 HEALTHCARE SYSTEM COMPLEXITIES, IMPEDIMENTS, AND FAILURES imaging increased 45 percent. Computed tomography (CT) scans represent the largest part of the cardiac imaging increase; CT scans of various body parts, excluding the head, have increased by 85 percent (MedPAC, 2007). In 2005, the estimated cost for all imaging was $100 billion (Farnsworth, 2005). It is fair to say that the benefit to patients of this increase in imaging remains unclear. There have been no tremendous declines in mortality or improvements in health outcomes that are clearly related to the increase in imaging. So what is driving the increased use of imaging? Certainly, the technology has gotten better. Pictures are much clearer, for instance. And the technology has also become easier to use. Furthermore, imaging-related entrepreneurial activity, such as freestanding CT centers, has grown, and once one has made a capital investment in a very expensive CT scanner, the incentive to use it is great. Defensive medicine, such as ordering a specific test because of concern about being sued, is always mentioned as a driver of healthcare costs in relation to technology advances. Patient demand for the use of new technologies has also increased. Patients read about these ad- vances on the Internet, hear about them in the media, are bombarded with related direct-to-consumer advertising, and request use of the technologies from their doctors. Pictures are very powerful, and people are driven by images they see in the media. A recent collection of media clips, for example, showed a cover story in Time magazine about a CT angiogram, with the headline “How to stop a heart attack before it happens.”1 Yet how these tests could prevent a heart attack is unclear. Tests appear to have become confused with pre- vention, but the link between the two remains undetermined. Most preven- tion is based on lifestyle changes—such as better diet, increased physical activity, and smoking cessation—that individuals can make to reduce their risk of disease. If people eat a heart-healthy diet, exercise regularly, and do not smoke, they can reduce their chance of having a heart attack by 50 percent. They can also get a CT scan, but doing so is not going to change their chance of having a heart attack. It is possible, of course, that taking the test might make a person more likely to eat a healthy diet, exercise, and not smoke, but there are no data indicating this is the case. Still, patients appear to hear the message that getting such tests can prevent a heart at- tack. When people say they are doing something for prevention, they are usually talking about getting some kind of test. Medicare data show a tremendous increase in the use of all cardiac imaging modalities. CT has seen the biggest increase, followed by magnetic resonance imaging and then positron emission tomography. Looking at these data, one can certainly understand why the Medicare Payment Advi- 1 Time Magazine, September 5, 2005.

OCR for page 117
10 ENGINEERING A LEARNING HEALTHCARE SYSTEM same page. The patient and practitioners can view the patient’s clinical and functioning status and outcomes the patient hopes to experience. This in- formation is used to promote shared and informed decision making, which leads to a plan of care for the patient. The one-page summary includes such essential information as patient history, symptoms, the patient’s perceived options for treatment and desired health benefits, and clinical and function- ing status. This information is updated over time and is available for each visit, making it possible to compare visits over time. Figures 3-7 through 3-8 illustrate the process and the one-page summary report. The Dartmouth Spine Center feed-forward information system has been running and evolving for more than a decade. With research grant support from the NIH, a similar data system was exported to 13 other medical cen- ters in 11 states across the country to gather data for randomized controlled trials and for observational cohorts concerning back surgery; the data have resulted in numerous articles in leading clinical journals (Weinstein et al., 2007a). In addition, the feed-forward system has been adapted for several other clinical programs at Dartmouth–Hitchcock Medical Center, includ- ing breast cancer, general internal medicine, plastic surgery, bone marrow transplant, and cardiovascular care. The Spine Center case provides a proof of principle for the patient- Shared Feed Forward Decision Sub-Acute Making Care Management Functional Interdisciplinar y Restoration Patient Program Assessment Enrollment Orient ation Assignment Preventative Care People with People with Management healthcare needs healthcare needs met Feedback Disenrollment Functional Functional Health Status Health Status National Biological Spine Expectations Satisfaction National Status Clinical Network Against Spine Biological Survey Network Need Costs Status Database Costs FIGURE 3-7 Spine Center process for a-7.eps F3 feed-forward and feedback information system. SOURCE: Eugene C. Nelson and Trustees of Dartmouth College.

OCR for page 117
11 HEALTHCARE SYSTEM COMPLEXITIES, IMPEDIMENTS, AND FAILURES centric, feed-forward collaborative idea. The data system supports indi- vidualized, patient-centered care. Clinicians are now able to inform patients about their chances of success and the likelihood of complications for nonoperative vs. surgical treatment options based on research on people like them. Data are used for program evaluation and improvement as well as for comparative benchmarking. The data system contributes to the infrastructure for interdisciplinary research programs—from bench to bedside to outcomes experienced by patients. It is being used for retro- spective and prospective research. Quality and cost data are published on the Dartmouth–Hitchcock website (DHMC, 2008) for transparent public reporting on important populations of patients. This initiative has helped the organization become an accountable healthcare system (Nelson et al., 2005). One interesting footnote to the Spine Center case study is that Terry Adams, the Dartmouth business school professor mentioned earlier, had the experience of going to the Spine Center soon after it opened its doors. He did not know that the Spine Center had been designed based on his own research concerning how the world’s best-in-class service organizations worked to bring quality and value to customers at the point of service, but he was moved to write a letter to the local newspaper about the wonderful care he had just received from the center. He praised the center for using innovative information technology to focus on the patient’s individual and unique health state, to elicit the patient’s expectations for care outcomes and explore all treatment options, to help patients make wise treatment decisions based on medical evidence and personal preferences, and to work smoothly with a full interdisciplinary team without having to go from clinic to clinic and experience frustrating waits and delays. Discussion: A Solution, Limitations, and Conclusions This final section of the paper proposes a solution to the challenge cited at the beginning of the paper, describes some of the limitations associated with this solution, and offers concluding remarks. The Challenge and a Solution If the aim is to build an information environment capable of generating clinical information and knowledge that can promote a learning healthcare system, we believe an essential part of the solution—although clearly not the full solution—is to intentionally develop what we call “patient-centric, professionally organized, feed-forward collaboratories.” A few brief de- scriptions of the key terms in this phrase follow:

OCR for page 117
12 ENGINEERING A LEARNING HEALTHCARE SYSTEM Functional Status Clinical Status Patient Perceived Outcomes Benefits History & Symptoms 3-12 a, b Mostly bitmapped

OCR for page 117
1 HEALTHCARE SYSTEM COMPLEXITIES, IMPEDIMENTS, AND FAILURES FIGURE 3-8 Patient summary report with longitudinal data: Dartmouth Spine Center. NOTE: MCS = mental component Figure 3-12 C scale, ODI = oswestry disability index, PCS = physical component scale. SOURCE: James N. Weinstein and Trustees of Dartmouth College and Dynamic Clinical Systems, Inc. • atient-centric—The individual patient’s health status, health risks, P decisions based on preferences and values, perceptions of good care and good outcomes, and costs of care are at the forefront of all that is done (IOM, 2001). • rofessionally organized—The healthcare professionals who serve P patients are expected to be responsible for the design of patient- centric delivery systems and the supporting information systems that enable them to partner with patients in delivering patient- centric care. • eed forward—Keeping patients and their data together over time F requires a well-designed information system that enables key infor- mation and data to move with the patient through the healthcare

OCR for page 117
1 ENGINEERING A LEARNING HEALTHCARE SYSTEM system over time to promote quality, safety, efficiency, and the best and safest match of services to patient health needs at any point in time and at any place in the system. • ollaboratories—The term denotes a method for organizing virtual C organizations in a complex world that combines the idea of col- laboration across physically distinct settings and the idea of a sci- entific laboratory. The purpose is to form a community of practice that can build shared information repositories for use in advancing science and improving practice (Schneiderman, 2008). What we are proposing, therefore, is to thoughtfully design and test innovative collaboratories that have all of the key features embedded in the Spine Center case. Some of the key characteristics of healthcare col- laboratories would be • atient-centric and focused on relevant dimensions of health out- p comes for any given population of patients; • rofessionally organized to fit into the flow of health care for p the purpose of improving care while contributing to research and education; • ased on feed-forward methods to follow patients over time as b their healthcare experience evolves and to better match patients’ changing health status with an evidence-based preference-sensitive plan of care; and • ependent on feedback methods to track health risks, health status, d diagnoses, and treatments associated with health outcomes and costs and to analyze results at multiple levels of the system (patient, micro, meso, macro, community, and region). This type of population-specific, feed-forward collaboratory could ad- vance goals on three major fronts: • ealth care—Provide better care for patients by matching wants, H needs, and health status with desired, effective, and efficient treatments. • ealth research—Provide data for observational and prospective H research on the causes of disease and disability and on the effective- ness of alternative methods for treating disease and disability. • ealth professional education—Create better learning environ- H ments that are information rich, patient focused, outcomes driven, and engaged in advancing healthcare science as part of regular work.

OCR for page 117
1 HEALTHCARE SYSTEM COMPLEXITIES, IMPEDIMENTS, AND FAILURES The idea of patient-centric, feed-forward collaboratories is innovative, but not new. The best examples we know of today are the Dartmouth Spine Center and the National Spine Network (Weinstein et al., 2000), as well as the Karolinska Institute and the Swedish Rheumatoid Arthritis Registry. However, there are other research networks and communities of practice that have some collaboratory features, including the Cystic Fibrosis Foun- dation and cystic fibrosis centers in the United States; the Vermont Oxford Project and neonatal intensive care units in North America and Europe; the Autism Program at Geisinger Health System; the Northern New England Cardiovascular Group and cardiovascular programs in Maine, New Hamp- shire, and Vermont; and the Clinical Program Model at Intermountain Health Care (James and Lazar, 2007). Limitations Any effort to work with professional organizations and health systems to develop and evaluate feed-forward collaboratories will have to recognize the current reality and some of the challenges and limitations this reality imposes. A few of these are listed below: • ision—Only a few models of collaboratories in health care are V available, and these are not well known. • ewards—Limited incentives and resources exist to establish col- R laboratories (at least in a non−Clinical Translational Science Award [CTSA] world). • ealth Insurance Portability and Accountability Act and security— H Following patients over time and across settings requires careful attention to privacy and security issues. • easurement—Only a limited number of patient-based “gold M standard” metrics exist for gathering both generic and condition- specific information. • tandardization—Resistance exists among many clinicians to using S standard, fixed-field data entry, and there are concerns about wast- ing time and doing work that is not value added. • atient role—It is a new role for the patient to act as a primary P reporter of key information using standard approaches. Exercis- ing this role will require changes in patients’ expectations and an understanding that their information-providing task is essential for their own care as well as for improving care and advancing science. These challenges suggest the need to develop demonstration programs to evaluate, validate, and refine the feed-forward collaboratory approach.

OCR for page 117
1 ENGINEERING A LEARNING HEALTHCARE SYSTEM Conclusion The time may be right for testing the patient-centric, feed-forward col- laborative model. Powerful forces at work are creating a climate favorable to the development of collaboratories. These forces include communities of professional practice combining patient care and health research, the funding of research by the NIH through the new CTSA approach, the formation of regional health information organizations across the country, the emergence of new scientific paradigms that recognize complexity and the value of multiple research methods, and demands for better quality and value that are measured and transparent. An excellent example of these forces coming together can be seen in the new National Quality Forum (NQF) framework that is being considered for measuring the outcomes and efficiency of episodes of care. The NQF approach is illustrated in Figure 3-9. It calls for the collection of feed-forward, patient-centric data on populations of at-risk individuals residing in different regions of the country. Then, after the onset of an illness episode, it calls for following people over time to measure critical information, including patient factors for risk adjustment, informed decisions guided by patient preferences, treat- ment processes, symptoms, physical function, and emotional status. Finally, FIGURE 3-9 Generic episodes of care. SOURCE: Reprinted with permission from the National Quality Forum (NQF, 2009).

OCR for page 117
1 HEALTHCARE SYSTEM COMPLEXITIES, IMPEDIMENTS, AND FAILURES at the end of the illness episode, it calls for completing the assessment by measuring mortality, functional status, and costs of care. The following statement by Fisher (2008) summarizes the value of de- signing patient-centric, feed-forward healthcare collaboratories: The same underlying information system is required to improve the evi- dence base for both biotechnology and care delivery. We need to know: • Patient attributes and risks (including biologic markers). • Specific, targeted biologic interventions performed. • Attributes of the system—delivery methods—where care is provided. • Health outcomes and costs. We could then have a truly learning healthcare system: • omparatie effectieness research: Compare biologic interventions, C controlling for patient and system attributes. • omparatie performance assessment: Compare systems and care C delivery methods, controlling for patient and treatment attributes. The bold aim is to achieve better patient and population health and better healthcare outcomes by applying research and education. Accom- plishing this aim will require that our health system become composed of learning healthcare systems. We conclude with four key points. First, the IOM definition of health stresses the functioning and well-being of the indi- vidual and requires patient-reported information to measure health status. Second, patient-centric health risks, health status, and health outcomes are an essential component of any comprehensive approach for improv- ing health care and studying health outcomes. Third, it will be essential to design feed-forward information systems to accomplish the tripartite aim of improving healthcare outcomes, advancing biomedical research, and enhancing health professional learning. Fourth, we believe that developing and testing patient-centric, professionally organized collaboratories can help the nation achieve this bold aim. REFERENCES Ash, J. S., M. Berg, and E. Coiera. 2004. Some unintended consequences of information technology in health care: The nature of patient care information system-related errors. Journal of the American Medical Information Association 11(2):104–112. Bates, D. W. 2005. Physicians and ambulatory electronic health records. Health Affairs 24(5): 1180–1189. Brenner, D. J., and E. J. Hall. 2007. Computed tomography—an increasing source of radiation exposure. New England Journal of Medicine 357(22):2277–2284. DHMC (Dartmouth Hitchcock Medical Center). 2008. Quality reports. http://www.dhmc. org/qualityreports (accessed June 6, 2008). Farnsworth, C. 2005. Testimony before the House Ways and Means Subcommittee on Health. U.S. Congress, House of Representatives. March 17.

OCR for page 117
18 ENGINEERING A LEARNING HEALTHCARE SYSTEM Fisher, E. S. 2008. Learning to deliver better health care: Rigorous study of the most effective ways to deliver care as well as what care works best can result in not only better treat- ment but also significant cost savings. Issues in Science and Technology 24(3):58–62. Spring 2008. Gray, R. M., W. R. Newman, and A. M. Reinhardt. 1996. The effect of medical specialization on physicians’ attitudes. Journal of Health and Human Behaior 7(2):128–132. IOM (Institute of Medicine). 1997. Improing health in the community: A role for perfor- mance monitoring. Washington, DC: National Academy Press. _____. 2000. To err is human: Building a safer health system. Washington, DC: National Academy Press. _____. 2001. Crossing the quality chasm. Washington, DC: National Academy Press. _____. 2007. Clinicians and the electronic health record as a learning tool. In The learning healthcare system (pp. 268-274). Washington, DC: The National Academies Press. _____. 2009a. Evidence-based medicine and the changing nature of health care. In Leadership commitments to improe alue in health care: Finding common ground. Washington, DC: The National Academies Press. _____. 2009b. Practical frontline challenges to moving beyond the expert-based practice. In Leadership commitments to improe alue in health care: Finding common ground. Washington, DC: The National Academies Press. James, B., and J. Lazar. 2007. Sustaining and extending clinical improvements: A health system’s use of clinical programs to build quality infrastructure. In Practice-based learn- ing and improement: A clinical improement action guide, 2nd ed. Joint Commission Resources. Karsh, B. T. 2004. Beyond usability: Designing effective technology implementation systems to promote patient safety. Quality and Safety in Health Care 13(5):388–394. Kawamoto, K., C. A. Houlihan, E. A. Balas, and D. F. Lobach. 2005. Improving clinical practice using clinical decision support systems: A systematic review of trials to identify features critical to success. British Medical Journal 330(7494):765. Lin, G. A., R. A. Dudley, and R. F. Redberg. 2007. Cardiologists’ use of percutaneous coronary interventions for stable coronary artery disease. Archies of Internal Medicine 167(15):1604–1609. _____. 2008. Why physicians favor use of percutaneous coronary intervention to medical therapy: A focus group study. Journal of General Internal Medicine 23(9):1458–1463. MedPAC (Medicare Payment Advisory Commission). 2007. MedPAC 200 report to Con- gress. Washington, DC: MedPAC Miller, G. A. 1956. The magical number seven plus or minus two: Some limits on our capacity for processing information. Psychological Reiew 63(2):81–97. National Priorities Partnership. 2008. National priorities and goals: Aligning our efforts to transform America’s healthcare. Washington, DC: National Quality Forum. NQF (National Quality Forum). 2009. Measurement framework: ealuating efficiency across patient-focused episodes of care. Washington, DC: National Quality Forum. Nelson, E. C., P. B. Batalden, S. K. Plume, and J. J. Mohr. 1996. Improving health care, part 1: The clinical value compass. Joint Commission Journal on Quality Improement 22(8):243–258. Nelson, E. C., K. Homa, M. Mastenduno, E. S. Fisher, P. B. Batalden, E. F. Malcolm, T. C. Foster, D. S. Likosky, J. A. Guth, and P. B. Gardent. 2005. Publicly reporting compre- hensive quality and cost data: A health care system’s transparency initiative. Joint Com- mission Journal of Quality Improement 31(10):573–584. Nelson, E. C., P. B. Batalden, and M. Godfrey. 2007a. Quality by design: A clinical microsys- tems approach. San Francisco, CA: Jossey–Bass.

OCR for page 117
19 HEALTHCARE SYSTEM COMPLEXITIES, IMPEDIMENTS, AND FAILURES Nelson, E. C., P. B. Batalden, and J. Lazar. 2007b. Practice-based learning and improement: A clinical improement action guide, 2nd ed. Oakbrook Terrance: Joint Commission Resources, Inc. Nolte, E., and C. M. McKee. 2008. Measuring the health of nations: Updating an earlier analysis. Health Affairs (Millwood) 27(1):58–71. Norris, P. 2007. Skeptical patients: Performance, social capital, and culture. In D. A. Shore (ed.), The trust crisis in healthcare: Causes, consequences, and cures (pp. 32-46). New York: Oxford University Press. Peterson, E. D. 2008. Is information the answer for hypertension control? Archies of Internal Medicine 168(3):259–260. Pham, H. H. 2007. Care patterns in Medicare and their implications for pay for performance. New England Journal of Medicine 356:1130–1139. Porter, M., and E. O. Teisberg. 2006. Redefining health care: Creating alue-based competition on results. Boston, MA: Harvard Business School Press. Pronovost, P. J., C. A. Goeschel, and R. M. Wachter. 2008. The wisdom and justice of not paying for “preventable complications.” JAMA 299(18):2197–2199. Redberg, R. F. 2007. Evidence, appropriateness, and technology assessment in cardiology: A case study of computed tomography. Health Affairs (Millwood) 26(1):86–95. Rotschild, M., N. Elias, D. Berkowitz, S. Pollak, M. Shinawi, R. Beck, and L. Bentur. 2005. Autoantibodies against bactericidal/permeability-increasing protein (bpi-anca) in cys- tic fibrosis patients treated with azithromycin. Clinical and Experimental Medicine 5(2):80–85. Schneiderman, B. 2008. Science 2.0. Science 319:1349–1350. Shih, A., J. Quinley, T. K. Lee, and C. R. Messina. 2002. Assessing pneumococcal revac- cination safety among New York state Medicare beneficiaries. Public Health Reports 117(2):164–173. Silber, J. H., S. V. Williams, H. Krakauer, and J. S. Schwartz. 1992. Hospital and patient characteristics associated with death after surgery. A study of adverse occurrence and failure to rescue. Medical Care 30(7):615–629. Warburton, R. N. 2005. Preliminary outcomes and cost–benefit analysis of a community hospital emergency department screening and referral program for patients aged 75 or more. International Journal of Health Care Quality Assurance Incorporating Leadership in Health Serices 18(6–7):474–484. Weinstein, J. N., P. W. Brown, B. Hanscom, T. Walsh, and E. C. Nelson. 2000. Designing an ambulatory clinical practice for outcomes improvement: From vision to reality. Quality Management in Health Care 8(2):1–20. Weinstein, J. N., T. D. Tosteson, J. D. Lurie, A. N. Tosteson, B. Hanscom, J. S. Skinner, W. A. Abdu, A. S. Hilibrand, S. D. Boden, and R. A. Deyo. 2006. Surgical vs. nonopera- tive treatment for lumbar disk herniation: The Spine Patient Outcomes Research Trial (SPORT): A randomized trial. JAMA 296(20):2441–2450. Weinstein, J. N., T. D. Tosteson, J. D. Lurie, A. N. A. Tosteson, E. Blood, B. Hanscom, H. Herkowitz, F. Cammisa, T. Albert, S. D. Boden, A. Hilibrand, H. Goldberg, S. Berven, H. An, and SPORT Investigators. 2007a. Surgical vs. nonsurgical treatment for lumbar 2007a. degenerative spondylolisthesis. New England Journal of Medicine 356:2257–2270. Weinstein, J. N., K. Clay, and T. S. Morgan. 2007b. Informed patient choice: Patient-centered valuing of surgical risks and benefits. Health Affairs 26(3):726–730.

OCR for page 117