The Value of Measuring Health Care Quality
The Value of Measuring Health Care Quality
Consider the following scenario:
You are contacted by the major health plan with which you contract and are told that your average length of inpatient stay is longer than their standard. You believe this is because your patients are more severely ill than average. How do you respond?
It's a fact: Physicians' practices are, and will continue to be, evaluated. It is therefore important to: 1) understand what is meant by quality, why and how it is measured and the problems with measuring it; 2) review current health care quality measurement activities; and 3) appreciate why physicians should become involved with quality endeavors.
Quality measurement is a reality. Why should health care providers become involved? Because quality data can be useful for physicians. Quality data may permit health care providers to understand how errors creep into clinical practice, and allow detailed review of all steps involved in providing care. Physicians are in an excellent position to make a case for improving quality.
If physicians cannot understand, much less lead, the debate about quality in health care, the public may lose confidence in physicians' abilities to serve and protect their patients in the face of health care system changes. By showing leadership, physicians can improve outcomes for patients and retain a measure of autonomy over the practice of medicine.
Until recently, physicians could be confident that they alone had the social mandate to judge and manage quality of care. Changes in the medical marketplace have probably had the most significant influence on the increasing prominence of health care quality measurement. As cost containment pressures have increased, quality of care issues have emerged. The field of outcomes research has grown in response to concerns about cost containment and health care quality, and in response to the belief that variation in medical practice may be due to uncertainty about appropriate types and levels of care (Palmer, 1996; Epstein, 1995).
In addition, practice patterns and the quality of medical care vary more than had been previously realized. Clinicians' interest in obtaining objective information about their practices has increased, and patients and purchasers want to know more about the quality of care available (Wennberg and Gittelsohn, 1982; Brook et al., 1996; Corrigan, 1995).
There are differing views on what constitutes quality. In 1980, Donabedian defined quality care as that which is expected to maximize an inclusive measure of patient welfare, after taking into account the balance of expected gains and losses that attend the process of care. According to the American Medical Association (1984), quality care is that which consistently contributes to the improvement or maintenance of quality and/or duration of life. The Institute of Medicine defines quality as the degree to which health services increase the likelihood of desired health outcomes and are consistent with current professional knowledge (Blumenthal, 1996).
Health professionals tend to define quality in terms of the technical quality of care and the quality of the interaction between the physician and the patient. Patients tend to take into account both traditional clinical end points and specific problems unique to a given disease or its treatment, including perceptions of physical, mental and social functioning, disability, and overall quality of life (Blumenthal, 1996; Ware, 1995).
Patients are more likely to define quality in terms of provider-patient interactions rather than clinical processes and outcomes of care (Kane et al., 1997). Health plans and organized health care purchasers emphasize the health of entire populations and those attributes of care that reflect the performance of organizational systems (Blumenthal, 1996).
There are many potential problems with quality measurement. First and foremost, the relationship between health care outcomes and the quality of care is not always clear. Many important outcomes occur long after care is provided, making it difficult to associate specific interventions or activities with those outcomes. It is difficult to relate receipt of specific health care services to a patient's quality of life or well-being.
Differences in severity of illness and patient preferences limit one's ability to make comparisons of clinical performance without some sort of adjustment method (Palmer, 1996; Epstein, 1995). In addition, distinguishing differences in outcomes requires large samples, which may not be available for a single provider (Ware, 1995; Palmer, 1996).
There are three generally accepted criteria for measurement of care quality: structure, process and outcome (Table). In the practice of medicine, the division between structure, process and outcome is not as distinct as might be theorized. In fact, process may not be an important predictor of outcomes, and differences in outcomes may be the result of factors not under the control of health care providers. If care quality criteria based on structural or process data are to be credible, one must demonstrate that variations in the attribute measured actually lead to differences in outcome.
Likewise, if outcome criteria are to be credible, one must demonstrate that differences in outcome will result if processes of care are altered. It is difficult to relate changes in patient health to specific medical care interventions. Many factors influence outcomes that are not directly attributable to the care process itself, and it is difficult to measure the individual contributions of each provider or organization to the overall care process (Brook et al., 1996; Hammermeister et al., 1995).
The challenge in designing quality measures is to decide what it is we want to measure and how we are going to measure it. First and foremost, one must have good measures: the integrity of everything that follows is driven by the measurement system. Generally, it is much better to design measures around specific questions, rather than to collect all data possible and hope to solve the problem in the process.
For data to be useful, they must have fundamental value in monitoring and improving care (O'Leary, 1995; Ware, 1995). One must understand the reliability, validity, strengths and weaknesses of different measures in order to avoid misinterpretation of the data.