Evidence-Based Performance Measurement

Motivation for Physician Performance Measurement

Professional, political, and societal interest in measuring quality in health care continues to be a dominant theme in medicine; it affects clinical care, physician education, research, and health policy. The demand for quality initiatives and measurement of outcomes of health care delivery has been motivated by multiple events and conditions. Most prominent among these is the sustained rapid increase in health care costs, at five times the average rate of inflation, and the observation from health services research that, despite spending more per capita than any other nation in the world on health care, the United States lags in many areas of public health and wellness.

Despite persistence of economic and market forces that incentivize volume and intensity of service, research shows that higher volume and intensity of health care services do not lead to better aggregated quality of life or public health. Overall, only about half of all Americans receive the recommended health interventions identified by consensus standards of care. Even more striking than the low overall proportion of those who receive recommended care is the wide variation that exists across health conditions, races, genders, and socioeconomic divisions. Dissatisfaction with the health care system is higher in the United States than in parallel western nations. Likewise, the percentage of U.S. citizens who did not get health care because of cost constraints is higher than in many other western nations. A huge gap exists between the consensus recommended appropriate care and the care that is actually delivered for easily identifiable and definable conditions. Large geographic variations in care, unexplainable by patient demographics and characteristics, are easily observable over a broad range of conditions. These geographic variations are far more significant than even the health care disparities seen as a result of ethnic or health literacy differences in the population. Unacceptably high rates of mortality and morbidity related to medical error have been the subject of many reports from both federal agencies and independent health services researchers. Finally, there seems to be no correlation between the per capita cost of health care and the quality of health care delivered on a range of observations and bases. The combination of all these factors has led to the current need for physician performance measures.

What Is Quality and Who Defines It?

Quality improvement and physician performance have taken center stage, yet there is no way to consistently define quality and its measurement. Each stakeholder, including the patient and the physician, has a reasonable perspective for viewing quality differently. A patient might define quality care as the relief of symptoms, perception of cure, or an improvement in lifestyle. However, the physician might define it as the achievement of a particular desired or expected medical or surgical outcome. An employer may see quality care as a return on investment for premiums paid, reduced liability for injury, and a workforce that is healthy, productive, and present in the workplace. A health plan purchaser may look at global health outcomes and the need to spread vast resources over large populations with competing needs. Therefore defining what constitutes quality and, hence, deciding exactly what to measure to determine whether quality is being delivered, continues to be debated.

In oversimplified terms, most measures of clinical quality or performance today fall into the following categories:

• Outcomes measure

• Process measures

• Capacity and structure measures

• Administrative measures

• Cost and efficiency measures

• Patient experience or satisfaction measures

The measurement of each of these segments has value but also pitfalls. Purchasers of health care have access to voluminous claims and economic data, making administrative, cost-effectiveness, and capacity measures attractive. Although many physicians opposed to administrative or efficiency measures clamor for outcomes measures as the only valid assessment of physician performance, in truth physicians rarely have complete control over all the factors that determine medical outcomes. As a result, many issues arise when it comes to measuring individual physician performance within a system of care or when the physician is operating within a team environment. Additionally, for valid outcome measures, effective risk adjustment must occur to reflect differences in the case mix of the patients served; this is often neglected, which results in misleading outcomes data. Process measures are easier to define and are more attributable to the practitioner; however, focusing primarily on processes of care can be deceptive when no one takes responsibility for the final outcome. Levels of evidence for different types of interventions can vary greatly, especially when medical care for chronic conditions primarily involving medication management are being compared with acute surgical care, for which randomized, double-blind, controlled studies may not exist or even be feasible. Because many elements—such as availability of support services and tertiary care, patient compliance, comorbidities, ethnic and religious practices, and preferences—can all influence the assessment of medical outcomes, measuring performance attributable to and under the control of the physicians being measured must be a common basic theme if fairness and true patient-centered quality improvement are to be achieved.

The concept of patient-reported outcomes and the related issue of “shared accountability” for health outcomes have gained traction in recent years. Gathering data directly from patients is one method of validating physician performance without relying on self-reported physician data or expensive external chart reviews. Potentially this could include aggregated global or population data for health care systems or large group practices or data at the level of the individual physician. The setting of standards for data integrity and validity is an extremely challenging process. It introduces concern for the role of the patient and his or her accountability for personal health choices and behaviors that influence desired health outcomes. Some social experience with patient accountability has been gained through employer programs. Although employer-sponsored wellness and healthy-lifestyle incentives are not new, very little has been done to measure the global effect and report on the patient's accountability for his or her health and for the public health in general. Current health services researchers are calling for specific standards of measuring and holding patients accountable for decisions and health choices they make that influence the quality of care. Diet and exercise; risky behaviors; tobacco, alcohol, and drug use; compliance with physician-directed care; and medication adherence are only a few examples of ways in which patient behavior affects health care outcomes. It makes little sense to hold physicians accountable for patient choices they cannot control; however, through patient education, an effective doctor-patient relationship, and appropriate communications and follow-up, physicians do have some influence on patient behavior. Therefore it is difficult to draw the public policy line on accountability.

The purpose for which measures are developed has a powerful influence on measures structure and what kinds of measures are used. Among other reasons, performance is measured today for the following overlapping purposes:

• Research, development, and improvement of the effectiveness of an intervention

• Reduction of medical error

• Improved patient safety

• Certification of achievement to meet standards for maintaining board certification

• Credentialing or accreditation to document training, competence, or proficiency for privileging, payment, or inclusion in a plan, group, or tier

• Licensing and discipline to identify, limit, and punish poor performance

Both overlap and synergy are found among these categories, and distinct subcategories further separate these types of measures. The American Medical Association (AMA) House of Delegates addressed criteria and standards for acceptable elements of any pay-for-performance system. As the quality movement matures and value-based purchasing takes different forms, embracing such criteria will become increasingly important to ensure patient centeredness.

The Process of Building a Coherent System of Performance Measurement

Three basic principles must underscore the roles of physicians and their organizations in addressing performance and quality improvement. First, it is essential that practicing physicians —not just methodologists and health policy scientists—actively and formally engage in prioritizing, developing, field testing, and implementing quality initiatives and performance measures. Second, demand for quality and its definition and measurement must be aggregated. Third, physicians and their organizations must be unified in their response to this demand.