NCBI Bookshelf. A service of the National Library of Medicine, National Institutes of Health.

Institute of Medicine (US). Evidence-Based Medicine and the Changing Nature of Healthcare: 2007 IOM Annual Meeting Summary. Washington (DC): National Academies Press (US); 2008.

Cover of Evidence-Based Medicine and the Changing Nature of Healthcare

Evidence-Based Medicine and the Changing Nature of Healthcare: 2007 IOM Annual Meeting Summary.

Show details

7Policy Changes to Improve the Value We Need from Health Care


Almost every aspect of introducing more evidence into the health-care system—from establishing electronic medical records (EMRs) to measuring quality care to revamping the reimbursement system—has complicated political and regulatory implications. Experts agree that the way these changes are presented to policy makers and the public will be critical for their acceptance. Moreover, a consensus on how to measure the value of healthcare services is needed to align evidence applications with overall healthcare goals.

In her paper, Donna E. Shalala draws on her experience as head of the Department of Health and Human Services in the Clinton administration in the 1990s to identify the political and regulatory challenges encountered in introducing evidence-based medicine (EBM) to the healthcare system. The way the political system is organized powerfully affects the quality of health care and the way it is delivered. Even when the evidence base is strong, it can be very difficult to make changes within the governmental system. It might not be realistic to expect Medicare or Medicaid to lead evidence-based efforts, particularly in terms of reforming the payment system. Drawing upon lessons learned in efforts to make childhood immunization universal in the 1990s, Shalala notes that successfully driving evidence through the healthcare system requires a very specific goal, consensus on the problem, and a public-private partnership on how to provide the solution.

Michael E. Porter’s paper argues that fundamental to real reform is the improved ability to define and measure value in health care. Within the field, however, there is a lack of consensus on a clear and shared goal—a confusion that is also reflected in many aspects of healthcare delivery, particularly in reimbursement policies. Delivering care of value is a goal that can unite the interests of all stakeholders; and defining and focusing on value is an important opportunuity to catalyze systemwide improvement in health care.

Porter defines value in healthcare delivery as patient outcomes achieved relative to the total cost of attaining those outcomes. The implications of measuring value for current approaches to process guidelines, outcomes measurement, and quality and safety improvement are discussed, along with some analytic challenges for the field. Consideration is needed on how to best measure health outcomes for multiple medical conditions and account for care received from many different specialists, and the long-term nature of the care. And an important first step is a more rigorous approach to defining and describing outcomes. In addition, the healthcare reimbursement model should be revamped to bundle reimbursements across medical conditions and cycles of care. Porter concludes that all stakeholders can contribute towards improvement in healthcare delivery by organizing their efforts around value as a central goal in health care.


Donna E. Shalala, University of Miami

Any discussion of incorporating EBM into the current healthcare system should include an examination of the role of government and regulatory agencies. This paper focuses on the political realities that must be faced in efforts to improve health care in this country.

The first issue to consider is whether discussions of cost savings and effectiveness should be separated. As government agencies try to institute reforms, one thing that has caused trouble in attempts to translate evidence into first-rate care is the suspicion among healthcare professionals that EBM is simply a budget cutting exercise. Part of the quality movement in medicine has tried to separate evidence-based care from cost cutting, implying that there is a kind of purity in doing the right thing and in making sure that available evidence is translated to the bedside.

As methods to incorporate better evidence into health care are considered, it must be remembered that the power of words cannot be underestimated. Those with knowledge of the history of attempts to get universal coverage—both from the debates over Social Security in the 1930s and of the Clinton Health Care Plan in the 1990s with the accusations of socialized medicine—know the power of words to transfer reform into something completely different. Whether reform efforts are labeled as socialized medicine or viewed as an attempt to force a cookie-cutter approach that takes autonomy away from professionals, these are powerful words and images that have been used to derail attempts to introduce EBM and universal coverage in the United States.

The way the political system is organized powerfully affects the quality of health care and the way that it is delivered. Those who have suggested that universal healthcare coverage is needed before evidence can be incorporated into the system should note that even where we currently have a single payer or a large purchaser, this has still been extremely difficult. It seems logical that very large purchasers of health care such as Medicare, the Veterans Administration, or the military healthcare system should be able to put requirements in place that allow us to take EBM and transform the way care is delivered. However, even when the evidence base is very strong, it can be difficult to make these kinds of changes within the governmental system. Every step toward implementing change has a constituency and, usually, members of Congress who are particularly interested in those constituencies.

Therefore, another question that needs to be considered is whether we can expect Medicare to lead evidence-based efforts, particularly in terms of reforming the payment system. There already have been many attempts to bring EBM into the Medicaid and Medicare systems. In many cases, EBM was successfully integrated with policy outcomes, and the enormous purchasing power of the Medicare system was used to make changes. However, it is important to caution that with big changes, such as introducing large-scale demonstration projects, there can be a great deal of resistance from Congress. This resistance was sometimes fueled by individual professional groups and healthcare companies that were trying to protect not only their incomes, but also their way of doing things. While there is strong agreement that there should be a link between incorporating EBM into health care and being reimbursed for providing quality health services, there is no consensus in this country on how to do that.

With these points in mind, an examination of the ways in which EBM has successfully been incorporated into the system can help identify what political conditions are needed to make it happen. The last century of public health provides many strong examples of evidence being used to transform people’s lives. They include such basic concepts as cleaner water, reliable sewage systems, and cleaner air. More recently, in the area of tobacco control and smoking cessation, the clear evidence that tobacco causes diseases exerted the necessary political pressure that changed behaviors and achieved results.

Another example of the overwhelming evidence that galvanized change is childhood immunization. As late as the 1990s, during the Clinton administration, half of the children in this country were not being immunized at the right time. Because of the overwhelming evidence that immunization dramatically improves public health, one of President Clinton’s main health priorities was to ensure that all children were receiving immunizations by age 3.

The fundamental problem in achieving universal immunization in this country has always been that a universal healthcare program is not in place. In fact, many thought it would be impossible without this. However, veterinarians in this country had already devised a tracking system and a method for immunizing every cow, sheep, dog, and kitten. If veterinarians could achieve this without universal health care, it could certainly be done for the children as well.

It took strong evidence and a powerful political and organizational movement to get 90 percent of the children in this country immunized. That meant winning the support of the states, professional groups, pediatricians, and health maintenance organizations; obtaining funding from Congress so the shots could be provided in doctors’ offices, clinics, hospitals, and pharmacies at almost no charge; and developing an immunization tracking system.

This example illustrates that to have success in taking evidence and driving it through the healthcare system, we need a very specific goal, consensus on the problem, and a public-private partnership on how to provide the solution.

In the case of immunization, a sustainable system was created that remains in place to this day. However, it must be remembered that it was not very controversial; neither political party would lose contributions over offering immunizations. In addition, the government built in reimbursements and other incentives to ensure cooperation, such as denying enrollment in Head Start or child care centers without immunizations or withholding benefits from the Special Supplemental Program for Women, Infants, and Children.

Unfortunately, unlike immunization, clear answers do not exist for many of the problems currently at hand. In some cases, we have enough evidence to move forward, but the political climate surrounding these issues requires careful consideration of the consequences. It is safe to say that almost all healthcare professionals work toward what is best for their patients, but the interface between their livelihood, perceived underpayments, and the demands of EBM—including a patient’s ability to shoot the messenger based on the research and to raise questions about the message—leads one to believe that this is far more complex than initially thought.

In this country we have a fragmented health system and fragmented research institutions. Many federal agencies are gathering and reviewing evidence, from the Agency for Healthcare Research and Quality (AHRQ) to the Centers for Disease Control and Prevention to the Department of Defense (DOD) to the Veterans Administration (VA). In addition, the private sector is also doing this type of research, including all of the specialty boards and professional organizations, as well as insurance companies such as Aetna, WellPoint, United Healthcare, Humana, and Cigna. In fact, the private sector has more up-to-date usage information than many federal agencies. Many have suggested that the United States follow the lead of England and centralize this research and recommendations.

Before such a decision is made, the following questions should be asked: Should a system exist where the messenger is going to take all the shots? Can congressional protection be put in place for that messenger? When AHRQ’s predecessor, the Agency for Health Care Policy and Research, tried to change practices in back treatment based on research, the back surgeons took it apart and the agency was almost disbanded. There are many instances where research almost brought an agency to its knees because a variety of interests weighed in on that particular piece of research.

Many agree that the best option would be to tithe every health plan in this country for this research in order to build a base of support. That funding would not necessarily be funneled to only one agency, but a base for research is crucial, given the billions of dollars that we spend on health care. Furthermore, it is crucial to politically protect the messengers, those who do the research and present it in a way in which it can be translated and used for patients.

Finally, is there a political will to make these kinds of changes? Every Republican and Democratic candidate for president speaks of the importance of information technology (IT). Why are EMRs not yet standard practice? Why is agreement lacking on what goes in that medical record? The DOD and the VA have required 6 months time to begin to come to agreement about what should be included in a medical record and about the interoperability of the system. While it is not an easy task, there certainly exists political consensus in this country on IT, as well as among presidential candidates on creating an agency for evidence-based research and the possibility of greater investments in research.

However, many raise the issue of political conditions for fundamental healthcare reform. In 1991 and 1992, when the Clinton administration was introducing its healthcare reform package, healthcare premiums were increasing and spending on drugs was going up faster (Figure 7-1). The growth in uninsured Americans is much more significant now than it was then. The increase in Medicaid spending is significantly less now. Other factors such as the rates of unemployment, inflation, and the federal budget deficit also are lower now than in the early 1990s.

FIGURE 7-1. The health reform context, then and now.


The health reform context, then and now.

Even so, Americans have not changed their minds over these last decades about the health system’s need for reform. While there is agreement on the problem, Americans are not sure government ought to fix it.

The Clinton administration of the 1990s alarmed many with a bold universal healthcare plan. People are not convinced that the president of the United States should be in charge of health care, with issues such as international security and the economy at the forefront of his or her mind.

Yet it helps to illuminate a principle of public policy. In our previous grand steps to introduce large-scale government solutions—in the 1930s with the creation of Social Security and in 1965 with the development of Medicare and Medicaid—there was agreement about the problem. During those periods there was also agreement about the solution; there was no serious private-sector or state alternative. There was debate about the delivery system, which is why Medicare is delivered by private insurers, but there was an agreement that the population to be served—the old and the sick—could not be served equitably by the private sector or other levels of government. Similarly, a private-sector alternative to Social Security did not exist either.

To make a grand social policy change such as providing universal health care, there must be consensus on both the problem and the solution. The mistake we made was that we assumed both existed, but alas, in actuality, they did not. We had agreement only that there was a problem. This is similar to the challenge that faces the country today.

In the end, to bring research to bear on the way in which medicine is practiced in this country and the way in which we organize health care, the most self-interested people in our country must provide the leadership. Even though a number of presidential candidates intend to take on the issue, switching to an evidence-based system will require healthcare professionals’ making the case as shrewdly as they possibly can.


Michael E. Porter, Harvard University

In any field, defining and measuring value is fundamental to progress. Improved performance and accountability are possible only when there is a clear and shared understanding of the goal, or “objective function,” to use an economist’s term. In most industries, participants contract freely and prices are set in open, competitive markets. The objective function is clear—making a profit. In healthcare delivery, however, the objective function is not so clear. Participant choices are constrained, prices are set or heavily influenced by government, and, in many cases, a public service mission and non-profit entities coexist with the profit motive.

Moreover, the profit motive is itself compromised in healthcare delivery by the current U.S. reimbursement structure in health care, as well as reimbursement practices in many other national systems. Prevailing reimbursement creates a disconnect between profits for system actors and value for patients because providers are rewarded for services, not patient results. System actors are also sometimes rewarded for shifting costs to others.2 Such a divergence between profits and patient value makes defining, measuring, and reporting value even more essential to true improvements in care delivery.

Not surprisingly, there is confusion and lack of consensus about the appropriate objective function in health care. Many in the field talk about access as the objective. Others point to equity, community service, or better population health. Still others define the goal as patient satisfaction, quality, safety, or care that is consistent with medical evidence. Nearly all actors identify cost containment and achieving an operating surplus as among their principal aims. The Institute of Medicine’s own definition of goals for the healthcare delivery system includes no less than six elements: safety, effectiveness, patient centeredness, timeliness, efficiency, and equity (IOM, 2001). In addition to the existence of multiple goals, different actors in the system define the objective function differently, a sign that the participants may be working at cross-purposes.

I believe that many of the difficulties in improving healthcare delivery stem from confusion and disagreement about defining, measuring, and rewarding value. The primary objective for healthcare delivery should be value for patients, measured by patient health outcomes per dollar expended to achieve those outcomes. Value is the only goal that unites the interests of all the parties in the healthcare system. Improving value is also fundamental to achieving all the other goals, such as expanding access and improving equity.

Defining Value

Value in any field must be defined around the customer, not the supplier. Value must also be measured by outputs, not inputs. Hence it is patient health results that matter, not the volume of services delivered. But all outcomes are achieved at some cost. Therefore, the proper objective is the value of health care delivery, or the patient health outcomes achieved relative to the total cost (inputs) of attaining those outcomes. Efficiency, as well as other objectives such as safety, are subsumed by value.

Health outcomes refer to the set of objective outcomes, not just patient perceptions of outcomes which can be biased toward the service experience. There is not just one outcome of care for any health condition, but multiple outcomes that jointly constitute value. Patient circumstances and preferences will affect the weighting of these outcomes to some degree, a subject discussed later.

The costs of achieving outcomes refers to the total costs involved in care, not just the costs borne by any one actor or for any particular treatment or episode.3 Hence, shifting costs across parties by, for example, raising patient co-payments for prescription drugs, does not add value but simply changes who pays. The mismeasurement of costs works against true value improvement, and is endemic in healthcare delivery in every country, especially in the United States, because of the way that services are organized and paid for.

Value for patients improves when equivalent outcomes are achieved at a lower cost, or better outcomes are achieved at comparable (or lower) cost. Yet outcomes and costs are not independent. A powerful lever to reduce costs is to improve outcomes, such as through early detection that limits the complexity of care, less invasive treatment, faster recovery, or less need for subsequent care. The power of quality improvement to drive down costs is greater in health care than any other industry I have encountered, because of the basic truth that better health is inherently less expensive than poor health.4 That so many actors in health care treat cost containment as the goal, instead of patient value, has severely handicapped the rate of value improvement.

Access to health care is a basic goal of any healthcare system, but access per se does not constitute value. Access provides the opportunity for value to be created by the delivery system, but is not in and of itself the goal. If outcomes were universally measured, it would quickly become clear that the value of care is highly variable, even for patients with access. Improving value holds the key to expanding access to care in a way that is affordable.

Equity in health care for all individuals and groups is another desirable goal, but again equity itself is not value. Equitable care that is poorly delivered leads to a system in which everyone has equal access to suboptimal outcomes. Discussions of equity also tend to focus on inputs, not outputs. The best way to improve the equity of care, and perhaps the only way, is to measure value, ensure transparency of value, and reward value. Only in this way will the value delivered for every patient count, including individuals who are currently poorly served.

Measuring Value

Value in healthcare delivery is largely unmeasured, a striking fact about healthcare delivery not only in the United States but around the world. Failure to measure value is the most serious self-inflicted wound of the medical profession and the broader provider community, because it has slowed innovation and brought about micromanagement of physician practice.

Measuring value depends first and foremost on properly measuring health outcomes. The chain of causality that determines outcomes is illustrated in Figure 7-2, which embodies Donabedian’s important distinction between process and outcome (Donabedian, 1966). Patients have some initial or preexisting conditions. Services are delivered through processes of care delivery that reflect medical knowledge and are affected by patient initial conditions. The care delivery process should strongly influence the outcomes achieved.

FIGURE 7-2. Measuring value in health care.


Measuring value in health care. SOURCE: Michael Porter, 2007.

Some of the current challenges in measuring value are highlighted by Figure 7-2. First, there is a great deal of confusion about the distinction between processes and outcomes. There has been a proliferation of protocols, evidence-based guidelines, and practice standards that are used to measure “quality” and serve as the basis of pay-for-performance initiatives (Fonarow et al., 2007). Many participants in the healthcare system, and most quality measurement systems in health care, confound processes and outcomes or treat processes and structures as if they were outcomes.

While structural factors, protocols, guidelines, and practice standards are partial predictors of outcomes, they are not outcomes themselves (Brook et al., 2000). Adherence to these types of measures is an imperfect indicator of outcomes.

Process guidelines are invariably incomplete and omit important influences on the value of care (Krumholz et al., 2007). Practice standards often fail to adapt care sufficiently to individual patient circumstances—standardized processes do not guarantee standardized outcomes. Experience also shows that providers following identical guidelines achieve different results. Process guidelines also fail to cover the full cycle of care that actually determines value. Thus, process measurement alone will not assure that results will improve for all patients.

Moreover, process guidelines can slow innovation, because agreeing on guidelines is inevitably slow and invariably political. Medicine is constantly being refined, and guidelines can lag best practice or, conversely, lead to undue attention to processes that have yet to be definitively proven with a sufficient body of evidence. For example, best practice in treatment of post-menopausal women with estrogen has changed several times in the last decade alone, as new evidence has become available about the risks and benefits of the treatment for particular patient subpopulations.

Process control alone, then, is a risky and ultimately flawed approach to improving outcomes and increasing patient value. In any complex system, attempting to control behavior without measuring results will tend to limit progress to incremental improvement. Without a feedback loop involving the actual outcomes achieved, providers are denied the information they need to learn and to improve their care delivery methods. Process control is a tempting shortcut because processes are easier to measure (and less controversial) than outcomes, but there is no substitute for measuring both (Birkmeyer et al., 2006).

Another important distinction is that between health indicators and levels outcomes as shown in Figure 7-2. Indicators, such as hemoglobin A1c used in diabetes care as biological markers of blood sugar control, should be highly correlated with actual outcomes such as acute episodes and complications (de Lissovoy et al., 2000). However, such biological indicators are still predictors of results, not results themselves. To improve value in healthcare delivery, it will also be necessary to measure true outcomes and not rely solely or even predominantly on such indicators.

Figure 7-2 also includes patient compliance as an essential factor contributing to health outcomes. There is compelling evidence that patient compliance with recommended preventive measures, preparations for treatment (e.g., weight control, muscle strengthening), and treatments themselves has a major influence on outcomes. Yet there is a glaring absence of systematic measurement of patient compliance, a major gap in measurement. Focusing on adherence to provider practice guidelines without measuring compliance merely obscures the link between process and outcomes. Failing to measure compliance also absolves providers (and health plans) of responsibility to treat compliance as integral to care delivery.

There has been growing attention to patient satisfaction in health care, but sometimes in a way that obscures true value measurement. Figure 7-2 separates two roles of patient satisfaction in measurement: patient satisfaction with the process of care (including hospitality, amenities, etc.) and patient reporting of the results of care. There has been a tendency to rely too heavily on patient surveys in quality improvement programs, and surveys have focused mostly on the service experience. These surveys rarely cover patient compliance, a major gap. Many surveys also fail to address what is most important for value measurement, the actual health outcomes as perceived by the patient. Generic patient satisfaction surveys also fail to capture the specific aspects of health status relevant to each patient’s particular medical conditions, which are usually the most important outcomes.

While the service experience can be important to good outcomes, it is the outcomes themselves that constitute value. In the absence of true results measurement, patients will tend to default to friendliness, convenience, and amenities as proxies for excellence in healthcare delivery. Providers cannot rely too heavily on service satisfaction surveys as measures of outcomes, or the value delivered.

Value and the Concept of Quality in Health Care

An important corollary to defining the value proposition in health care is the definition of quality. In health care, the whole notion of quality has become a source of confusion and sometimes a distraction from genuine value improvement. Quality ought to refer to patient outcomes. Quality relative to cost determines value in health care, as it does in any field.

In health care, however, most quality initiatives are focused on processes of care and compliance with evidence-based guidelines. For example, of the 71 Healthcare Effectiveness Data and Information Set measures, the most widely used quality measurement system, only six are outcomes or health indicators and the balance are process measures. Of the comprehensive collection of quality measures found in the National Quality Measures Clearinghouse, the overwhelming majority are not outcomes (AHRQ, 1999). The quality movement in health care is on a dangerous path by trying to measure and control physician practice directly, rather than measuring outcomes. While outcome measurement is difficult, process measurement is not a substitute.

There has also been a tendency to equate safety and quality. The proliferation of safety initiatives is laudable, and has produced genuine improvements for patients. However, safety is just one aspect of quality and not necessarily the most important aspect. To say it another way, doing no harm is important, but improving the degree of recovery or the sustainability of recovery are just as important, if not more so. As I will discuss below, too much focus on safety instead of overall outcomes and value may lead to incremental process improvements affecting safety, rather than rethinking the overall delivery of care to improve total outcomes including safety.

The Unit of Value Creation

To understand value in any field, the unit for which value is measured should conform to the unit in which value is actually created. The unit of value creation should define organizational boundaries in care delivery, which is a central tenet of organizational theory. In health care, however, both measurement and organizational structure are misaligned with value creation. In fact, one of the principal reasons why value is mismeasured in health care, or not measured at all, stems from faulty organizational structures for healthcare delivery.

Patient value is created by the integrated care of a patient’s medical condition over the full cycle of care, rather than by a single specialist or by a discrete intervention (Porter and Teisberg, 2006). A medical condition is an interrelated set of patient medical circumstances best addressed in an integrated way. Care for a medical condition, such as breast cancer, diabetes, inflammatory bowel disease, asthma, or congestive heart failure, will normally require the involvement of multiple specialties. The definition of a medical condition includes the most common co-occurrences, or diseases that occur together. Caring for the medical condition of diabetes, for example, needs to integrate the care for hypertension and vascular disease. Value for the patient is created by the combined efforts to care for the patient’s medical condition, not by any one specialty.5 True accountability for value is inherently shared among the providers involved.

The unit of value creation in health care delivery—care for a medical condition encompassing the cycle of care—collides with the way delivery is currently organized in the United States and in virtually every other country. Instead, care should be organized into Integrated Practice Units (IPUs).6 A patient may be cared for by more than one IPU at the same time, which greatly simplifies the coordination challenges compared to today’s structure in which each specialist or intervention for each medical condition must coordinate with all the others involved in the patient’s care. Health care today is organized by facility (e.g., hospital, clinic), by specialty, and/or by discrete intervention (e.g., imaging, surgery, office visits). This means that both outcomes and processes tend to be mismeasured. Also, faulty organization of care creates many hurdles to actually achieving excellent outcomes.

Measurement today usually focuses on the individual providers or specialists, despite the fact that the intervention of one provider is not the sole or even the primary determinant of the overall outcomes. Measurement focuses on the discrete intervention, despite the fact that the intervention is one of many that determine outcomes. Measurement covers short episodes, which tells an incomplete story in understanding the overall outcome. For example, mortality measures often track mortality after surgery or for the duration of the patient’s stay, but not after a short post-discharge period. Outcomes from a few discrete interventions, or in a few medical conditions, tend to be used as proxies for the overall outcomes of the provider.

Current organizational structure in healthcare delivery makes it difficult to measure value correctly. Indeed, this is one of the most important reasons why it is poorly measured, or not measured at all. Providers, particularly, have a tendency to measure only what is under their direct control in a particular intervention, even if this is not what actually determines value. What is measured is what is easy to measure, rather than what matters for outcomes. What is measured is also what is billed, even though the unit of reimbursement is misaligned with overall value.7 Or, measurement is centered for the hospital or facility as a whole, even though this unit of analysis has little relevance for value delivered.

Gathering long-term, longitudinal data on outcomes is surely challenging, but the cost of doing so is unnecessarily high because of the current organizational structures and practice patterns. If practice structures were realigned to cover the care cycle, the cost of long-term outcome measurement would fall dramatically. Moreover, the assumption of joint responsibility for outcomes would be natural.

All these observations also apply to measuring costs. To understand the true costs of heath care delivery, one must measure the costs of all the interventions and services involved in determining the outcome. Today each unit or department is seen as a separate revenue or cost center; no one measures the cost of the entire care cycle. Entities such as rehabilitation centers and counseling units are all but ignored in cost analysis. Many costs, such as those borne by the patient or within primary care practices, are not counted in measuring procedure-centric care. Treating drugs as a separate cost, for example, only obscures the overall value of care. All costs must be included to measure the total cost of delivering outcomes, and overall value.

While the unit of value creation is the medical condition over the cycle of care, a given patient may have multiple medical conditions. This often occurs, for example, in older patients who might have congestive heart failure and breast cancer and osteoarthritis of the hip. Such patients are best cared for by integrated practices for each condition that coordinate with each other. Value is best measured for each medical condition, with the presence of other medical conditions a risk factor in each one. In this way, it is possible to compare one patient’s results with others, and measure how well a provider deals with complicated patients. The alternative, defining a different measure of value for each patient, defeats the whole purpose of measurement.

Measuring Health Outcomes

Outcomes are the core of value in healthcare delivery. There is growing attention to measuring outcomes, which is a most welcome development. However, the practice of outcome measurement suffers from a number of problems. One of these is a tendency to look for a single ideal outcome measure for a given medical condition. Often, the measure chosen is one which is easy to agree upon and/or easy to measure. However, there is never one outcome measure in any field or endeavor, and health care is no exception.

For every medical condition, there are multiple outcomes that collectively define patient value. One commonly measured outcome is survival or death. This is just one outcome, albeit an important one. Outcomes related to safety, such as the incidence of medical errors along with their consequences, are an additional type of outcome measure. To think holistically about outcome measurement for a medical condition, outcomes can be can be conceptualized in a hierarchy, with the most fundamental outcomes, survival and patient health, achieved at the top, and other outcomes arrayed in a natural progression, such as those related to the nature and speed of the recovery process and those related to the sustainability of the results. Although there is not time to explore the details in this discussion, it should be possible to characterize the set of outcomes for a medical condition in a fashion that lends to objective and quantitative outcome measurement.8

For many patients, trade-offs may exist among individual outcomes. For example, a more complete recovery could require treatment with a greater risk of care-induced illness. Or, more complete recovery could require treatment that is more discomforting. Where there are trade-offs, individual patients may differ in the weight they place on different levels of the hierarchy, and on specific measures. The discomfort of treatment willingly endured will be affected, for example, by the degree and sustainability of health achieved. For example, cosmetic considerations may weigh heavily against risk of recurrence, such as in the choice of the amount of the breast resected for breast cancer patients, or long-term sustainability of recovery may matter less to older patients. A complete understanding of all aspects of such an outcome hierarchy matters more, not less, when different groups of patients value individual outcomes differently.

Thus, the first step to a systematic approach to value improvement is a disciplined approach to defining and describing the total set of outcomes. In most fields, including medicine, progress in improving value is iterative and evolving. Excellent performance on one quality attribute leads to attention on improving others (although improvements may also occur simultaneously). Over time, innovations seek to relax trade-offs among quality dimensions.

In healthcare delivery, the concept of an outcome measures hierarchy emphasizes that progress can be made at different rates at different levels. As survival rates improve, for example, more attention can be focused on the speed and discomfort of treatment. Once effectiveness in recovery reaches an acceptable level, attention can shift to relaxing trade-offs between effectiveness and risk of complications, as in cancer therapy. By measuring the entire outcome hierarchy, such improvement is not only encouraged but made more transparent and systematic. And viewing outcomes in a hierarchy reveals opportunities for dramatic value improvements in existing therapies as well as in the development of more cost-effective therapies that address disease earlier in the causal chain. This is a potential source of great optimism for the future in terms of cost containment.

Some Analytic Challenges

If we posit a hierarchy of outcome measures for a medical condition, this raises the question of how the importance of each one should be determined. Also, should the set of outcome measures be aggregated to determine an “overall” outcome? These are important questions, which can easily derail outcome measurement. They have led to the effort to monetize outcomes by, for example, calculating the value of human life or measuring the monetary benefits of improved productivity. If outcomes can be monetized, they can be aggregated and directly compared to costs to determine benefit-cost of value.

Seeking to monetize individual outcomes is tempting, but unnecessary and even misleading and distracting in value measurement. Monetizing even tangible outcomes such as improved survival is fraught with complexity, and often arbitrary. Monetizing more subjective or intangible outcomes is problematic. How should less arduous or less discomforting treatment be monetized? How should cosmetic or appearance improvements be valued? With multiple outcomes, as we have noted, the value and weights will also vary by patient.

Attempting to calculate a single aggregate outcome measure for all patients, or for each patient, is not the right approach to outcome measurement, at least given the current state of practice. Instead, the focus should be on improving the set of outcomes (and value) in the sense that some outcomes improve without sacrificing others. For this purpose, outcomes need not be monetized, and individual outcome measures need not be aggregated.

Similarly, factoring initial health state into outcome measurement is an important issue. Because a patient’s initial conditions affect the outcomes that can be achieved, measuring initial conditions is needed and outcomes must be stratified based on the important initial conditions or risk-adjusted. Several efforts to gather and report outcomes have failed because inadequate risk adjustment led to resistance and rejection by the medical community (Porter and Teisberg, 2006). Even in its current imperfect state, however, getting on with understanding the relevant initial conditions and adjusting for them is essential to improving value itself. For example, the lack of case adjustment methods is a root cause of the underpayment of providers for more complex cases, both in the United States and elsewhere. Flawed reimbursement for complex versus simpler cases has many adverse consequences for value, ranging from inadequate care to excessive fragmentation of “profitable” services.

Finally, the task of appropriately measuring costs requires close attention. Cost measurement needs to follow some essential principles, including: measuring the full costs of care, not the portion of costs borne by any one actor or the portion of costs taking any one form (e.g., drugs, tests, office visits); endogenizing costs borne by employers or families; aggregating costs for medical conditions over the cycle of care, by aggregating the costs of all the interventions involved; measuring the actual costs, not charges; aligning the unit of reimbursement with the unit of value—which will require a shift to bundled reimbursement models for medical conditions, in which all physician fees, services, facilities, and drugs over the care cycle are included in a single price; and varying prices according to patient initial conditions.


Value must be the fundamental goal of any healthcare system. Measuring value, and improving it, must become the driving force for every participant in the system. Today, in the U.S. healthcare system and in systems around the world, value is measured incompletely, if at all, which is the single biggest weakness standing in the way of healthcare improvement. The fact that healthcare delivery is not organized around value impedes excellent care and drives up its cost. The fact that reimbursement is not aligned with value cripples the process of value improvement, and renders the profit motive a destructive force rather than a value driver.

Proper measurement of value is the single most powerful lever for improving healthcare delivery. While current organizational structures and practice standards surely create obstacles to value measurement, there are promising efforts to overcome these obstacles. While current measurement efforts are highly imperfect, at least the process of measurement has begun. Health plans, providers, employers, and government policy can all contribute to making measurement of value in health care a reality. If all actors in health care were to embrace value as the central goal, and work together to achieve good measurement, the resulting improvements in healthcare delivery would break the current logjam that threatens the nation’s human and economic health.


  1. AHRQ (Agency for Healthcare Research and Quality). The summary of outcomes research at AHCPR: Final report. 1999. [accessed June 26, 2007]. http://www​ .
  2. Birkmeyer JD, Kerr EA, Dimick JB. Improving the quality of quality measurement, Appendix F. Performance Measurement. 2006:177–203.
  3. Brook RH, McGlynn EA, Shekelle PG. Defining and measuring quality of care: A perspective from us researchers. International Journal for Quality in Health Care. 2000;12(4):281–295. [PubMed: 10985266]
  4. de Lissovoy G, Ganoczy DA, Ray NF. Relationship of hemoglobin A1c, age of diabetes diagnosis, and ethnicity to clinical outcomes and medical costs in a computer-simulated cohort of persons with type 2 diabetes. American Journal of Managed Care. 2000;6(5):573–584. [PubMed: 10977465]
  5. Donabedian A. Evaluating the quality of medical care. The Milbank Memorial Fund Quarterly. 1966;44(3 Pt 2):166–203. [PubMed: 5338568]
  6. Fonarow GC, Abraham WT, Albert NM, Stough WG, Gheorghiade M, Greenberg BH, O’Connor CM, Pieper K, Sun JL, Yancy C, Young JB. Association between performance measures and clinical outcomes for patients hospitalized with heart failure. JAMA. 2007;297(1):61–70. [PubMed: 17200476]
  7. IOM (Institute of Medicine). Crossing the quality chasm: A new health system for the 21st century. Washington, DC: National Academy Press; 2001.
  8. Krumholz HM, Normand SL, Spertus JA, Shahian DM, Bradley EH. Measuring performance for treating heart attacks and heart failure: The case for outcomes measurement. Health Affairs. 2007;26(1):75–85. [PubMed: 17211016]
  9. Porter ME, Teisberg EO. Redefining health care. Boston, MA: Harvard Business Press; 2006.



This article draws heavily on joint research with Elizabeth Teisberg (Porter and Teisberg, 2006). I am grateful to her for her fundamental contribution to the line of thinking discussed here. This paper also draws on subsequent research by Benjamin Tsai, Saquib Rahim, and Jennifer Baron at the Institute for Strategy and Competitiveness at Harvard Business School. Jennifer Baron, Sachin Jain, Joan Magretta, Michael McGinnis, and Elizabeth Teisberg provided helpful comments.


The same is true for health plans, who can profit through shifting costs and restricting revenues to providers, patients, or the government without actually offering patient outcomes or true efficiency.


We treat non-economic costs of care, such as the patient’s discomfort, anxiety, and time, as part of patient outcomes. See below.


This does not imply, for example, that all efforts at prevention will lower costs or even improve value. However, focusing on outcomes as a way to reduce long-term costs will spur innovations in better and more cost-effective forms of prevention as well as other value improvements.


Note that the set of specialties required to care for a medical condition may differ across patient populations.


The IPU structure dramatically reduces the complexity of coordination and facilitates the true medical integration of care.


For these reasons, the use of claims data in outcome measurement can be misleading unless it aggregates claims at the medical condition level.


See Michael E. Porter, “What Is Value in Health Care?” Working paper, Harvard Business School, Boston, MA, 2008.

Copyright © 2008, National Academy of Sciences.
Bookshelf ID: NBK52830


  • PubReader
  • Print View
  • Cite this Page
  • PDF version of this title (1.9M)

Related information

Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...