Health system performance will not simply improve over time. In fact, challenges are likely to escalate due to growth in the aging population, increasing public demands, and predicted resource constraints. Historically, quality has largely been addressed through professional registration and licensure, accreditation of hospitals, and less formal professional peer review processes. While these traditional methods of quality management are important and need to continue, new approaches to improving performance will depend on better diagnosing the problems and implementing more effective solutions.

A key lesson from other industries is that most performance problems can be traced to flawed systems, inadequate training, and perverse incentives that hinder people from performing optimally. The Institute of Medicine has called for a redesign of local health delivery systems accompanied by new policies to promote high-quality health care (IOM 2001b, 2004). Given the complex nature of the health care system, systematic improvement will require an integrated strategy that sets national priorities and targets for improvement, provides infrastructure support through technical assistance and information technology, and monitors performance and rewards its improvement (Leatherman and Sutherland 2003).

Putting knowledge into practice
Closing the gap between what is scientifically known and what is actually done by health care providers is not simply a matter of exhorting everyone to work better and harder. The sheer volume of the medical literature makes it humanly impossible for physicians to master all of the emerging new knowledge. Systematic reviews of the literature and the development of evidence-based practice guidelines are important starting points for knowledge transfer, but disseminating guidelines alone is rarely enough to change practice (Grimshaw et al. 2001).

Educational strategies alone may sometimes be sufficient to change straightforward practices such as selection of an appropriate antibiotic (Steinman et al. 2006). To make more complex changes, health care professionals must be convinced that a new approach is worth adopting and given practical ways of implementing it along with feedback and support to reinforce change (Torrey et al. 2001). Effective methods and systems for improvement need to be routinely identified and disseminated in "user-friendly" formats such as learning aids and collaborations, skills-based training, peer mentoring, and educational outreach (Kilo 1999; Resar et al. 2005; Thomson O'Brien et al. 2001; 2000a; 2000b; 2000c).

Practice change can be aided by tools such as:

• checklists and standardized order sets to ensure that critical steps in a process are followed and that critical information is communicated (Lappe et al. 2004; Mehta et al. 2002; Pronovost et al. 2003);
• disease registries to track patients in need of follow-up care (Griffin and Kinmonth 2000; Stroebel et al. 2002); and
• computerized prompts that remind physicians to deliver recommended services during patient visits (Balas et al. 2000).

Making these approaches commonplace will not be easy. A recent survey of physician practices provides a sobering assessment that "physicians have not yet fully embraced [quality improvement] principles and methods" (Audet et al. 2005). An international review found that the U.S. lags behind other advanced nations by more than a dozen years in adopting health information technologies, such as electronic health records, that can help improve the quality, safety, and efficiency of health care (Anderson et al. 2006; Chaudhry et al. 2006).

Making performance information available
One reason for the uneven quality of health care is the lack of systematic reliable reporting that objectively describes the nature of the problem and monitors progress. That gap is being addressed, in part, by the National Healthcare Quality Report and the National Healthcare Disparities Report, mandated by Congress and published annually by the Agency for Health Care Research and Quality (AHRQ). The Institute of Medicine recommends that Congress create a National Quality Coordination Board to guide and manage the selection and use of a consistent set of quality measures nationally and regionally (IOM 2006).

Performance measurement and feedback also need to take place at the front lines, where the patient and health care provider interact, so that clinicians, hospitals, and health systems can understand where local improvement is needed (Jamtvedt et al. 2003). Organizations that collaborate to share performance data for benchmarking purposes can identify and spread processes of care that are associated with better outcomes (Horbar et al. 2001; O'Connor et al. 1996).

The patient is increasingly regarded as an important audience for performance data. For example, a public-private collaboration known as the Hospital Quality Alliance is reporting on the quality of care at more than 4,000 participating U.S. hospitals. The Medicare Modernization Act of 2003 established financial incentives for hospitals to report these data to the federal government. Similarly detailed comparisons of quality are now available for other types of health care organizations, including nursing homes, home care agencies, and health plans.

Some experience suggests that well-designed public reports on provider performance can motivate health care organizations to improve processes and outcomes of care, especially when baseline quality is poor (Casalino et al. 2003; Hibbard et al. 2005; Marshall et al. 2000). Following release of the New York State Cardiac Surgery Reports, for example, heart bypass surgery mortality declined significantly more than the national average and reached the lowest level of any state in the nation (Peterson et al. 1998). Methodological and presentation issues must be carefully considered to ensure that public reports are fair, accurate, and usable (Hibbard et al. 2002).

Improving performance through coverage
The title of a recent report from the Institute of Medicine says it succinctly: "Coverage Matters" (IOM 2001a). Those who lack health insurance coverage are less likely to get needed health care and are at risk for poor health outcomes (IOM 2002)—at an economic cost to the nation of $65 billion to $130 billion in lost life-years and productivity (IOM 2003a). Experience suggests that when uninsured individuals gain coverage, the quality of their health care improves. For example:

• Formerly uninsured near-elderly individuals who gain Medicare coverage upon reaching age 65 increase their use of preventive care closer to the level of those who were continuously insured before Medicare enrollment (McWilliams et al. 2003).
• Uninsured children who enrolled in a State Children's Health Insurance Program had fewer unmet needs and improved quality of care one year later (Szilagyi et al. 2006; 2004). Moreover, preexisting racial and ethnic disparities in access to care were eliminated (Shone et al. 2005).

Almost 47 million Americans, or 16 percent of the population, did not have any type of health insurance coverage during 2005. The rate was 11 percent among non-Hispanic whites, 18 percent among Asians, 20 percent among blacks, and 33 percent among Hispanics (U.S. Census Bureau 2006). Researchers judged that one-quarter of the uninsured were eligible but not enrolled in existing public programs, but only 20 percent could afford to purchase coverage themselves—meaning more than one-half of the uninsured need financial assistance to purchase coverage (Dubay et al. 2006). Recent coverage expansion plans in Massachusetts and Vermont offer hope that other states might follow suit.

It is not yet known what effect the recent trend to enroll individuals in high-deductible health plans and health savings accounts might have on the quality of health care. One early study found that consumers were less satisfied with such plans and were more likely to delay or skip care compared with those with more comprehensive coverage (Fronstin and Collins 2005). Health savings accounts could provide incentives for consumers to take greater responsibility for their health. On the other hand, these plans might lead to poorer outcomes, especially among those with chronic illnesses and low incomes (Davis 2004).

Rewarding performance improvement
Attempts to exploit consumer market forces have had only a marginal effect on quality to date and the need to implement more effective incentives—both financial and nonfinancial—is increasingly recognized. For example, one study found that physician organizations subject to external incentives for quality were more likely to adopt disease management programs associated with improved quality (Casalino et al. 2003). The most potent incentive was public recognition.

Paying for performance is a concept of increasing interest in both publicly financed and private health systems (Rosenthal et al. 2004). Financial incentives must be designed carefully to reinforce positive performance while avoiding unintended consequences. For example, early experience suggests that incentives should reward improvement as well as attainment of high performance (Rosenthal et al. 2005). The Agency for Healthcare Research and Quality has developed an evidence-based guide to help purchasers implement effective pay-for-performance programs (AHRQ 2006).

Other types of incentives include paying hospitals to participate in improvement collaborations and report on their results, and offering health care providers a discount on malpractice liability insurance for engaging in activities that enhance patient safety (McCarthy 2005; McCarthy and Staton 2005).

Designing high performance into the system
In many cases, changing the way that health care is organized and delivered is the most—or even the only—effective means of improving practice. For instance, a growing body of experiential learning suggests that health care organizations can best ensure patient safety through a "safety culture" that encourages vigilance in reporting and analyzing errors so that system failures can be diagnosed and corrected. They take account of the propensity for human error by promoting skills and behaviors, such as effective teamwork and communication, that minimize the opportunity for harmful mistakes (McCarthy and Blumenthal 2006).

The health care system also must be designed to eliminate existing racial and ethnic disparities in care. The Institute of Medicine called for increasing the number of minority health care providers, the use of cross-cultural training, and the availability of interpretation services, among other recommendations (IOM 2003). For example, cultural competency policies and training have been associated with better patient adherence to treatment (Lieu et al. 2004).

Long-term restructuring programs are needed to build understanding about the factors that enable certain organizations to consistently achieve high performance and to apply and adapt this knowledge to redesign local delivery systems based on local needs. At the tactical level, practitioners need usable guidance on where to focus their improvement efforts to achieve the greatest impact on patients' health, and on what kinds of cross-cutting organizational and system changes will support improvement across a range of common health conditions.

Yet experience shows that even when cost-effective improvement interventions have been demonstrated, they often do not diffuse into routine practice and policy (Seow et al. 2006). Encouraging more implementation-oriented research may help to overcome this gap (Rubenstein and Pugh 2006). Ultimately, successful transformation will require leadership, collaboration, and system-wide culture change supported by market incentives, professional education, and policy initiatives. The Commonwealth Fund recently established a Commission on a High Performance Health System with a mission to promote this kind of systemic change (Davis 2005).

Conclusion
Health system performance is a challenging issue, defined differently by various stakeholders and impacted on many levels from the complexity of large organizations to the dynamics of interpersonal relationships. Even with the multitude of perspectives, there is little argument that Americans are increasingly concerned about the effectiveness, responsiveness, and accessibility of health care services. Performance information is a prerequisite to identify the priority areas, create a common understanding of the gaps, and point us to reasonable strategies for improving the American health care system.