The intensity of the patient safety movement calls attention to the potential downside of focusing too much on a single disease or health theme: Concentrating on 1 niche of medicine can ultimately compromise the overall health of the population if it comes at the expense of other problems that are more threatening to health. A balanced approach to improving health would transcend theme areas, such as patient safety, and examine the big picture. If failure to provide preventive care claims 100 times as many lives as do lapses in safety, then a system that is preoccupied with safety and does little about preventive care will see more of its patients die. Too much of a good thing in quality improvement paradoxically threatens health by giving inadequate attention to more important needs.

For example, the U.S. Agency for Healthcare Research and Quality, the principal funding source for research on the uninsured, racial disparities, and other aspects of quality, allocates 60% of its grants to patient safety. Funding for all other categories of health services research, now described as "non–patient safety" grants, was cut by 15% to finance this investment (2). As of April 2003, Congress was considering 20 bills on patient safety (3), but only minimal progress was expected on improving care for disadvantaged individuals (4, 5).

The Notion of Proportionality

The issue is not whether patient safety should receive attention, as it surely must, but whether that attention is proportionate. Unless attention and resources are allocated to safety and other quality improvement areas in proportion to their relative effect on public health, an excess of the population may die or sustain morbidity. For every patient harmed by lapses in patient safety, more will experience or die of deficient health care and flawed delivery systems, which are problems that a perfect safety record will not take away. People are less likely to die of an overdose of warfarin (a lapse in safety) than of not receiving warfarin at all (6). The attention that policymakers give to safety should be coupled with a proportionately larger effort to deal with defects in health care that affect more lives.

Before this discussion proceeds further, the central premise for this concern—that there is only so much attention to go around—must be defended. Some might argue that it is unnecessary for patient safety to hinder other forms of quality improvement and that one has little to do with the other. But, in fact, resources for quality improvement are often fixed. Providers, practices, and health systems can give only so much time, human energy, and money to making health care better, and they cannot fix all problems at once.

This means that choices must be made, with resources applied to one quality arena not going to another. Systems preoccupied with safety may find little energy remaining to recognize, let alone rectify, deeper design flaws that compromise caring. As task forces convene to work on patient safety, are other aspects of quality left unattended? At the national level, is the focus on patient safety coming at the expense of efforts to help the uninsured (7) or reduce disparities in the care of minorities (8)?

Patient Safety in the "Big Picture"

Imagine 4 concentric circles representing all deficiencies in the quality of health care (Figure). At the center is the problem of patient safety, the most extreme violation of quality, in which the attempt to heal has the perverse effect of causing harm. The center blurs into the larger circle of medical errors, which connotes a larger category of mistakes that endanger patients in domains other than "safety," such as mistakes that create deficiencies in managing chronic disease (9). That circle of errors, in turn, blurs into the larger circle of gaps in quality, encompassing not only individual errors but also the broader organizational causes of inadequate care, including restricted access to care (7), racial and ethnic disparities (8), systemic quality defects introduced by insurance and management policies, faulty information systems, and flawed system designs. This circle encompasses underuse, overuse, and misuse of care (10). Some envision a fourth circle: the extent to which the technical elements of care seem good on the basis of performance indicators but ultimately fail to be caring because of deficiencies not captured by these measures (11, 12). The rudeness or insensitivity that patients encounter or the frustrations they experience in obtaining information and control over treatment decisions illustrate gaps in quality, of deep concern to the public (13), that often are not measured under normative standards.

View larger version (43K): [in this window] [in a new window]   Figure. Organizational framework for deficiencies in health care service. Center. Lapses in patient safety—mistakes in the provision of health care that expose patients to "additive" risk—inducing risks for complications or overt injuries that did not exist before the clinical encounter (for example, amputating the wrong limb or prescribing a toxic drug dose). Second circle. Medical errors—mistakes that encompass not only lapses in safety (center) but that also include inattention to extant risks that patients bring to the encounter (for example, not offering pneumococcal vaccination or colorectal cancer screening to eligible patients or not achieving optimal blood pressure control). Third circle. Lapses in quality—care that does not reach desired standards not only because of mistakes made by individuals (first and second circles) but also because of flaws in the design and operating procedures of systems and organizations (for example, failure to provide access to care, insurance coverage, timely reminders for overdue services, or acceptable waiting times). Fourth circle. Lapses in caring—unsatisfactory care resulting not only from failure to meet normative benchmarks for quality (center, second, and third circles) but also from experiences that leave patients feeling uncared for, affecting them in domains that are less easily measured (for example, feeling unheard, rushed, inconvenienced, or humiliated; or being unable to access desired information, instruction, or reassurance).

Implicit in this model is the proposition that patient safety is a subset of everything else. Quite the opposite has been suggested of late, with "patient safety" being used as shorthand for medical errors (or worse, all deficiencies in care) (14-17). But safety, defined as "freedom from injury" (1), connotes a certain kind of injury brought on by clinical intervention, such as a drug overdose or a surgical mishap, in which patients find themselves with a new problem that did not exist before the clinical encounter. This additive risk differs from errors of inadequate treatment for problems brought to the clinical encounter, in which injury occurs by default. If a patient arrives with a glycohemoglobin level of 10% and the provider does not act, the provider does not give the patient a new problem (making the visit "unsafe") but causes harm by not reducing an existing risk.

Putting Health Priorities in Epidemiologic Perspective

These errors in addressing extant risks arguably are more threatening to health than lapses in safety. Although "To Err Is Human" (1) suggested that 44 000 to 98 000 Americans die each year because of medical errors, more careful analyses suggest that only a fraction of these deaths, perhaps fewer than 5%, are causally linked to errors (18-20). Only a subset of adverse drug events, the first concern of the patient safety movement, causes serious harm (21). A cohort study of Medicare beneficiaries noted 5 life-threatening or fatal preventable adverse drug events for every 1000 person-years of observation (22).

By comparison, modeling studies estimate that hundreds of thousands of Americans die because of inadequate treatment of cardiovascular disease, cancer, and other conditions (23, 24). Failure to address extant risks in just 1 niche of medicine, the use of ß-blockers after acute myocardial infarction, claims 4300 to 17 000 lives each year (23, 25). The numbers grow larger when deaths due to omissions in other cardiac drug regimens and in managing other diseases (for example, stroke, cancer, and diabetes) are added (26). The National Committee for Quality Assurance recently estimated that 57 000 deaths occur annually because of shortcomings in delivering recommended care (27). Many more deaths—potentially more than 700 000 per year (23)—result from gaps in screening, immunizations, and reducing risk factors (for example, obesity and unhealthy diet) in the population (28-30). Tobacco use alone accounts for more than 400 000 deaths per year (31).

This matters because getting the balance wrong can cost lives. If poor control of blood pressure or serum lipid levels accounts for more deaths than do illegible drug prescriptions (21, 23, 32), a quality improvement program that is preoccupied with computerized prescription entry but ignores the large proportion of patients with uncontrolled hypertension or hyperlipidemia costs more lives than a program with reverse priorities.

This contention relies to some extent on a false dichotomy, because common etiologies often cause lapses in both safety and quality and similar solutions apply. The same reminder system that corrects drug errors can also recognize overdue mammography (33). Work conditions that cause confusion induce mistakes, whether it is reaching for the wrong vial (threatening safety) or overlooking a needed treatment (not reducing existing risks). Understanding what causes individuals to commit errors can therefore bear fruit in both domains. Although safety initiatives can thereby identify system solutions that kill 2 birds with 1 stone, they may overlook system defects unrelated to safety that are more threatening to health.

Designing Systems for Much More than Safety

Ominous flaws in the structure of the health care system severely compromise the health of Americans (34). The system delivers only 55% of recommended care (35). Patients cannot access providers, services are not covered, and features of the delivery system impede efforts to obtain what patients need precisely when they need it. System solutions to improve access, reduce inequities, and provide the infrastructure for health promotion and chronic disease management (36, 37) may do little for safety but are urgent public health priorities that could lower morbidity and mortality on a scale unattainable by safety interventions (10).

The urgency of these redesigns is apparent to those who take a global perspective and measure progress in terms of overall population health. The urgency may be less palpable to those preoccupied with health niches—such as patient safety, asthma, or bioterrorism—whose narrower perspective may obscure larger priorities. Clinicians or researchers battling a single disease may not consider whether expending the same effort on more threatening conditions or solving deeper, systemic root causes may be more beneficial (38). They are satisfied with doing a good thing without it being the best thing, but patients pay the price for the difference.

Patient safety experts who redesign systems to reduce drug errors, catheter sepsis, or anesthesia mishaps do good work. However, by not addressing larger deficiencies in quality, they may fix problems in the branches and twigs while preserving proximal disease in the trunks (39, 40). They may reduce accidents but leave deeper defects in place, making health care safer (less likely to add new risks) but still ineffective in caring for extant risks.

The greatest good for the health of the population comes from a global perspective that views the system as a whole, judges its performance by its effect on population health rather than on parochial domains, and prioritizes interventions in a rational scheme to optimize outcomes. Such an approach, albeit rational, faces potent challenges. Individual diseases, not global health outcomes, are what motivate policy and medical leaders. The dramatic changes that a global approach demands would be resisted by power centers that face financial, political, and administrative consequences (41, 42).

Methodologic tools for such an approach are also limited. A global approach for prioritizing interventions requires a common metric for contrasting the relative effect of interventions on health. Quality-adjusted or disability-adjusted life-years can serve this purpose, and policymakers have used these tools to prioritize the relative importance of preventive and other services (43). However, the methods and data sources for applying this metric across diverse health care interventions are not straightforward. Nonetheless, the path toward overcoming these challenges begins with the proper vision.

Safety, the center of the 4 circles, is an essential but uninspired goal for health care. Patients deserve far more than not to be harmed by their physicians. Improving safety is vital, but it should not distract from the larger mission of helping people maintain their health and cope with illness, of which safety is only 1 component. The medical profession should focus on this larger purpose and not on fragments. Precise thinking aimed at optimizing the health of the population should guide the prioritization of care improvements and system redesigns. Anything less tolerates a disproportion in which the public does better with portions of their care but ultimately suffers greater illness and death.