Health care reform is back on the nation's political agenda. For those of us who have been working on health policy for some time, it feels like déjà vu all over again. Many aspects of reform—like the problems of how to cover uninsured Americans at sustainable costs and how to distribute these costs—never seem to change. The debates around these issues that now engage Washington often echo those heard during the Clinton effort, 15 years ago; the Nixon effort, 35 years ago; and even the Truman effort, over 50 years ago. But this year's debate does have one important difference. One of the fundamental premises underlying prior discussions of health reform—the belief that, at least for insured Americans, the quality of health care in the United States surpassed that available anywhere else in the world—has been called into question.
The anemic recent performance of the U.S. health care system is depressingly evident in our mortality and morbidity statistics, in comparisons of the quality of care for specific conditions, and even in surveys of the service quality of health care delivery (Schoen et al. 2006). We rank 19th of 19 countries assessed by the OECD in mortality from conditions amenable to health care. Our healthy life expectancy at age 60 is the lowest among 23 countries measured (Commonwealth Fund 2008). In 1976, a 40-year-old American man could expect to live nearly 33 years, about 1/2 a year less than the OECD high-income country average and more than 1/2 a year longer than a German (the OECD high income laggard). By 2006, a 40-year-old American man could expect to live nearly 38 years—an impressive 4.8 year gain over the 30-year period. But over that same period, the OECD high-income country average increased by over 6 years, and by 2006 the 40-year-old American man's life expectancy was more than 1/2 a year less than a German's (still the OECD laggard) (author's tabulations of SourceOECD, 2008). While we spend much more on health care per capita than any other country in the world, we have shockingly little to show for our expenditure. A recent economic assessment of the U.S. health care system concluded that its performance is “uniquely inefficient” (Garber and Skinner 2008).
This disheartening evidence on the outcomes of the U.S. health care system has led to an appropriate focus on quality improvement as a central feature of reform efforts. In the 2008 Presidential campaign, both Senators McCain and Obama stressed elements of their health plans aimed at improving quality: disease management initiatives, improvements of electronic record systems, and payment strategies geared toward the effective deployment of costly resources.
Health services researchers have a long record of developing and testing quality improvement ideas. Several papers in this issue provide such insights. Prior research has shown that providing health care practitioners with information on the behavior of peers can influence their behavior. Such comparisons are most effective if professionals believe that the comparator set is appropriate. Byrne et al., in this issue, provide a novel strategy for generating peer groups that are likely to be viewed as acceptable.
Health care delivery is multifaceted and not all elements of quality are likely to be measurable. If practitioners are rewarded based on their performance on measurable quality elements, they may skimp on the less-observable components of quality. In this issue, Werner, Kobetzka, and Kruse report on how performance on elements of quality of care that were not reported publicly fared when some measures of nursing home quality were reported publicly. Their results provide a mixed picture of the impact of quality measurement. For high-performing nursing homes, quality improved on both reported and unreported measures. For low-performing nursing homes, quality on unreported measures did not improve and sometimes diminished.
Health services researchers have also looked at the steps organizations take to generate quality improvement. Nembhard reports on the features of cross-institutional collaboratives that teams find helpful. She finds, for example, that teams find the high-cost, high-intensity, face-to-face “Learning Sessions” element of the collaborative particularly helpful. She also finds that those teams most effective in implementing quality improvement were most likely to find interorganizational learning helpful. She hypothesizes that teams that face fewer internal or external constraints may be best able to improve quality and to benefit from interorganizational learning. As this issue of HSR shows, there is a great deal to be learned and implemented based on the health services research literature on quality improvement.
Seen from a different perspective, however, the health services research literature on quality improvement also points to the significant limitations of quality improvement as a stand-alone strategy for improving health outcomes. As the papers in this issue, and many others, suggest, quality improvement depends on measurement of population level outcomes, professional commitment, and institutional investments. These inputs cannot be effectively assembled and deployed at a national level in the absence of universal health insurance coverage.
Universal coverage and health outcomes are intertwined at many levels. The most obvious is that uninsured people do not get health care that is timely and appropriate. They seek care less frequently and do not use needed care consistently. Their health outcomes are, consequently, worse. And since the uninsured population now comprises over 15% of all Americans, their inclusion in the statistics brings down average outcomes. But the pernicious effects of lack of coverage on health outcomes go well beyond this straightforward connection.
The uninsured do not constitute a distinct population outside the mainstream of health care delivery. Rather, nearly one third of the entire population experiences a spell of uninsurance over a 4-year period (Bhandari and Mills 2003). From a quality improvement perspective, it is both the overall number of uninsured people and this high level of churning that play havoc with efforts to systematically measure, reward, and improve quality. The magnitude of the uninsurance problem makes it very challenging to measure the quality of treatment over the course of time, which is particularly important for those with chronic illness. Churning is an obstacle to risk adjustment, because information about care during and before uninsured spells is likely to be missing. It makes comparisons across institutions and professionals difficult because some patients may fail to adhere to recommended regimens because their coverage ends, a problem for which medical professionals cannot be reasonably held accountable. Thus, the possibility that patients will become uninsured affects the care they receive even when they hold coverage.
The existence of a substantial uninsurance problem affects the ability and desirability of rewarding high-quality care. The level and nature of care provided to uninsured people depend on professional and institutional commitment. Advocates are rightly concerned that efforts to improve quality for insured populations, in response to high-powered payment incentives, will lead to reductions in the level and quality of care provided to the uninsured. The results in Werner, Kobetzka, and Kruse suggest that this is possible. Systems-level outcome improvement cannot occur if improvements in the quality of care provided to some diminish the quality and level of care available to others.
The need to serve a disconnected, unpredictable, nonpaying population also saps institutional commitments to quality improvement. Staff and leadership are unlikely to make investments in patient reminder systems, follow-up protocols, and electronic record systems if they must simultaneously support emergency rooms crowded with patients who should be receiving care in outpatient settings. They will not turn their undivided attention to quality improvement if they are busy trying to cover the costs of uncompensated care. Even an angelic health care institution cannot do two things very well at once. It should not be surprising that most of the institutions most notable for a systems-based approach to high-quality care are located in communities with relatively low rates of uninsurance.
Covering the uninsured will not solve the quality problems of our health care system, but it will certainly help. Consider the evidence from the other high-income OECD countries, where outcomes have improved much faster than in the United States. While many European countries have made recent targeted investments in quality improvement, through pay-for-performancelike initiatives (notably in the United Kingdom), the implementation of electronic medical record systems (notably in Denmark), and disease management initiatives (notably in Germany), these efforts are of recent vintage and sporadic implementation. What distinguishes the OECD countries is not their efforts at quality improvement, which vary widely across these countries; rather, it is their continuing commitment to universal access. Indeed, OECD country efforts at quality improvement have been linked with efforts to make care more accessible, through new mandates for coverage (the Netherlands, Germany) and reductions in copayments for low-income populations (France). Without such a commitment in the United States, quality improvement initiatives are likely to remain seductive models that attract policy maker attention without contributing meaningfully to population health outcomes.