Research Activities, February 2011: Primary Care Research: Primary care physicians' performance ratings depend on the makeup of their patient population

Primary Care Research
Primary care physicians' performance ratings depend on the makeup of their patient population


Primary care physicians (PCPs) with a large proportion of underinsured, minority, and non-English-speaking patients have lower care quality rankings than their colleagues with more advantaged patients, concludes a new study. This could lead to inaccurate physician rankings with implications for physician compensation and allocation of resources, note the Harvard and Massachusetts General Hospital researchers who conducted the study. To evaluate PCP clinical performance, they used a composite quality measure based on 9 Healthcare Effectiveness Data and Information Set measures (for example, regular mammography, Pap screening, colonoscopy, and cholesterol testing). Physicians were ranked by percentile (1-100) and then grouped into thirds.

Patient panels of physicians who were ranked in the top third were older, had more illnesses, and made more frequent primary care visits (more than three per year). These patients were also less likely to be minority, non-English-speaking, have Medicaid coverage, or lack insurance. After accounting for practice site and visit frequency differences, adjusting for patient panel factors resulted in a relative mean change in physician rankings of 7.6 percent per physician, with 36 percent of primary care physicians reclassified into different thirds.

The researchers conclude that since patient panel characteristics affect the relative measured quality of physicians, efforts to improve quality of care must address fairness of the assessment of physician clinical performance. It must also address the design of incentive schemes to provide equitable distribution of resources and reduce disparities in care to vulnerable patients. The study took place in the Massachusetts General practice-based research network that includes 181 primary care physicians working in 9 hospitals and 4 community health centers. These physicians saw over 125,000 patients during a 3-year period.

The ongoing and future phase of the study is supported by the Agency for Healthcare Research and Quality (HS18161). For more information on this project and other AHRQ health information technology projects, go to http://healthit.ahrq.gov.

See "Relationship between patient panel characteristics and primary care physician clinical performance rankings," by Clemens S. Hong, M.D., M.P.H., Steven J. Atlas, M.D., M.P.H., Yuchiao Chang, Ph.D., and others in the September 8, 2010, Journal of the American Medical Association 304(10), pp. 1107-1113.
Research Activities, February 2011: Primary Care Research: Primary care physicians' performance ratings depend on the makeup of their patient population