American health care is respected worldwide in terms of its training and education, technological sophistication, focus on the consumer, and relative ease of access for those with insurance coverage or ability to pay. Yet, whether through the eyes of the caregivers or the patients, recent surveys show widespread concerns regarding the performance of the American health care system. In part, these ratings reflect the very high expectations that Americans have for their health care.

Performance problems that exist cannot be cured simply by more money. In fact, the U.S. already spends 50 percent more on health care per capita than the next highest-spending countries (Norway and Switzerland). Health care expenditures consumed 15 percent of gross domestic product in the U.S. in 2003, compared with an average of about 9 percent across 30 democratic countries (OECD 2005).

Yet in a recent Commonwealth Fund study that surveyed patients in six nations, the U.S. ranked first on only one of six performance domains and ranked last on four of the domains (Davis et al. 2006). The authors concluded that the "United States often stands out for inefficient care and errors and is an outlier on access and cost barriers" (Schoen et al. 2005b). We can and should do much better in terms of the value we receive for the dollars we spend.

Improvement in five areas of performance could significantly affect the process and outcomes of health care: 1) providing appropriate and effective care based on evidence; 2) reducing unjustified variation in care; 3) eliminating disparities and inequalites in access and treatment; 4) preventing harmful errors; and 5) improving responsiveness to patients. A brief overview of each area follows.

Appropriate and effective care for acute and chronic disease
A comprehensive study conducted by the RAND Corporation in 12 U.S. communities compared patients' medical records and interview responses with 439 measures of quality for 30 different health conditions. Overall, the patients' care adhered to expert recommendations only 55 percent of the time. More than one of 10 patients received care that could do more harm than good. Almost one-half of the patients failed to receive care that would have been beneficial for their health condition (McGlynn et al. 2003).

These findings illustrate the problems of overuse and underuse. For example, physicians overuse antibiotics by continuing to prescribe them for the common cold even though the evidence is clear that antibiotics are ineffective. And in more complex conditions such as heart disease, physicians underuse medications such as beta-blockers that are known to be effective in preventing the reoccurrence of a heart attack. In numerous other clinical situations even when effective treatment is known and agreed upon, the care does not reflect good science. Moreover, even when health care providers achieve high quality of care for one clinical condition, such as heart attack, they may not perform equally well in treating another condition, such as pneumonia (Jha et al. 2005).

Unwarranted Variation in Health Care
During the last several decades, researchers have produced a large amount of evidence showing that the use of various medical services, treatments, and procedures varies substantially depending on where one lives or receives care, even for patients with similar illnesses and needs (Dartmouth Atlas Project 1999, 2006). For example, one study found that hospital use and physician visits varied up to 14-fold among chronically ill patients treated at 77 different hospitals in the last six months of life (Wennberg et al. 2004).

In the aggregate, more intensive use of services in an area (and correspondingly higher spending) does not appear to translate into higher quality of care, better patient outcomes, or greater satisfaction (Baicker and Chandra 2004; Fisher et al. 2003a, 2003b; Sirovich et al. 2006; Teno et al. 2005).

Unwarranted variation results from "care that is not consistent with a patient's preference or related to a patient's underlying illness" (Wennberg and Wennberg 2003). Experts who study variation point to three main factors that likely contribute to this quality problem (Dartmouth Atlas Project 2005a, 2005b, 2005c):

1. underuse of effective treatments when there is clear scientific evidence about what works best;
2. failure to fully educate patients on the risks, benefits, and tradeoffs of treatment alternatives, which allows decisions to be influenced by idiosyncratic local norms or assumptions (such as that more care is better care); and
3. overuse of discretionary services in areas with a relatively larger supply of hospitals and specialists, especially when there is a lack of clear evidence on what treatment approach works best.

At a policy level, unwarranted variation implies inequity and waste in the use of resources. For example, Dartmouth researchers estimate that the federal Medicare program could reduce spending on care at the end of life by up to 30 percent if all hospitals and physicians treated severely ill patients in the same manner as the most efficient providers (Dartmouth Atlas Project 2006).

Medical errors and patient safety
The Institute of Medicine estimated that medical errors lead to thousands of preventable patient deaths annually (IOM 1999). A recent poll found that nearly one-half of Americans are very concerned that a medical mistake will harm them or a family member during a hospital stay (KFF 2003). Many people report experience with potentially harmful mistakes such as being given the wrong medication or wrong medication dose (Schoen et al. 2005b).

Medical errors may be a rare event, such as amputation of the wrong limb, or the much more common but potentially lethal situation of prescribing contraindicated drugs or a drug overdose. Patients also can be harmed when health care professionals fail to observe infection control procedures and other safety precautions. Advances in medicine may lead to more opportunities for error, such as an increase in medication errors as more drugs are being used (Lesar et al. 1997).

A review of patient safety failures concluded that "preventing future failures depends on cultural as much as structural change in health care systems and organizations" (Walshe and Shortell 2004). Although medicine has a strong commitment to patient welfare, human beings will never be infallible. Current thinking emphasizes the need for a systems perspective to identify the underlying flaws in training, skills, behaviors, equipment, and procedures that give rise to harmful error. This proactive approach makes it possible to design countermeasures to prevent patient harm.

Access issues
Access problems are many and diverse. Those who are uninsured, underinsured, or who experience gaps in coverage have generally worse self-reported access to care and quality of care compared with those who are continuously and adequately insured (Hadley 2003; Schoen and DesRoches 2000; Schoen et al. 2005a). In particular, they are:

• less likely to have a regular care provider;
• more likely to delay or forgo needed medical care, preventive services, and prescription drugs;
• more likely to have poor health outcomes; and
• less likely to rate the quality of their care as good or excellent.

Even when people have insurance coverage, they might not be able to obtain care because of barriers in language, culture, transportation, or geography. For example, among those who have private insurance or Medicaid coverage, Spanish-speaking Hispanic adults have worse access to health care compared with non-Hispanic white adults, whereas access to care is similar among English-speaking Hispanic adults and non-Hispanic white adults (Fiscella et al. 2002).

Responsiveness/patient-centered care
Respect, dignity, autonomy, kindness, and convenience are all critical from a patient's perspective. Patient-centeredness has been defined as the degree to which patients and their health professionals establish a partnership "to ensure that decisions respect patients' wants, needs, and preferences and the patients have the education and support they need to make decisions and participate in their own care" (IOM 2001c).

A recent survey of patients found that many reported suboptimal communication and engagement with their physician (Schoen et al. 2005b). Likewise, a survey of physicians found that many did not have mechanisms to promote patient-centered care (Audet et al. 2006).

A great deal of activity is now focused on providing information to patients that would enable them to make more informed decisions about where to get health care and how to participate more actively in treatment choices. Research has shown that when patients are given good information that is tailored to their needs they use it and make prudent decisions (Coulter 2003).

Quality of care for vulnerable populations, such as nursing home residents, is sometimes overlooked, but the problems can be serious. One survey showed generally positive views from people that have a friend or family member in a nursing home. However, a significant minority indicated that they, or someone they know, received poor-quality care, including about one-quarter who reported incidents of abuse or being poorly treated by staff (NewsHour/KFF/HSPH 2001).

Disparities and inequalities in treatment
Gaps in performance affect the entire U.S. population, but not equally. Research reveals disparities in care by race and ethnicity, gender, and socioeconomic factors such as income and education (AHRQ 2005a; IOM 2003b). Understanding the root causes of disparities is important to identify those that are amenable to change by the health system or that require changes in wider social policies.

• Differences in insurance coverage, income, and, to a lesser degree, the availability of community safety net resources (charity care and emergency rooms) explain most of the disparities in access to care between Hispanics and whites and much of the access disparity between blacks and whites (Hargraves and Hadley 2003).
• Many studies find that racial and ethnic disparities in treatment and outcomes remain significant even after clinical and socioeconomic factors have been controlled. Disparities are best documented in cardiac care but have also been reported in cancer diagnosis and treatment, pain management, HIV treatment, diabetes care, kidney transplantation, child and maternal health care, mental health care, and nursing home care (IOM 2003b).
• Potential causes for these disparities include mistrust in the health system caused by poor treatment, miscommunication due to cultural differences between clinicians and patients, and physician biases due to stereotyping and uncertainty. Although minorities are sometimes more likely to refuse treatment, patient preferences do not fully account for disparities in treatment (IOM 2003b).

The increasing diversity of the U.S. population—with projections that minority groups will outnumber whites within 50 years, as they do already in California—makes the issue of disparities all the more urgent (Johnson 1999; Smith and Edmonston 1997). In the words of Martin Luther King, Jr., "Of all the forms of inequality, injustice in health care is the most shocking and inhumane" (King 1966).