Physician Quality Reporting and Patient Outcomes in Medicare
Principal Investigator: Bryan Dowd, Ph.D.
Division of Health Policy and Management, School of Public Health, University of Minnesota
This study examined the impact of the Physician Quality Reporting System (PQRS) on the quality and cost of care for Medicare patients. PQRS enables individual physicians and other eligible professionals to voluntarily select the measures to report data on the quality and outcomes of care provided to Medicare beneficiaries. This information goes beyond the standard diagnosis and procedure codes available from administrative claims data. It is increasingly of interest to the Medicare program, since the Affordable Care Act requires Medicare to incorporate measures of "value" into its payment systems for providers, including physicians and hospitals caring for beneficiaries in the traditional fee-for-service program. The term "value" implies consideration of both cost and quality, a major change in the philosophy underlying Medicare payments.
The study examined the effect of the onset of PQRS reporting on various measures of utilization and cost, including newly developed measures of inappropriate emergency department (ED) utilization and overuse of services. It also examined the relationship of PQRS reporting to physician specialty, a topic of current interest to policymakers.
The analysis uses a difference-in-differences model to compare changes in outcomes over time (from before implementation of PQRS in 2005 and 2006 to after implementation in 2008 and 2009) for practices that submitted PQRS reports to those that did not. Since PQRS reporting is voluntary, the study also examined whether physicians reported only "good" values on the outcome variables of interest, or whether they also reported "fair" and "poor" values.
Participation in PQRS was associated with a significant, desirable (negative) effect for two of three measures of avoidable utilization across the entire sample of beneficiaries. The magnitude and statistical significance of the desirable effects increased in subgroups of providers and beneficiaries more prone to overutilization, and among beneficiaries with heart problems, diabetes, and eye problems. Within subpopulations of interest, the onset of PQRS was associated with improvement in avoidable utilization, but not risk-adjusted cost.
The study examined selected health outcome measures for which physicians used PQRS to report good, fair, and poor values on the variables of interest. Using a composite measure of hemoglobin A1c, low-density lipoprotein, and blood pressure for a sample of diabetic beneficiaries, the study found that the reporting of "good" rather than "poor" PQRS outcomes was significantly associated with fewer ambulatory care-sensitive admissions, potentially preventable readmissions (PPRs), and avoidable ED visits, and with lower inpatient and higher outpatient expenditures. As expected, a comparison of "fair" versus "poor" outcomes produced similar but weaker results, as did a comparison of "good" versus "fair" outcomes (except that in this latter analysis, the difference in PPRs was not statistically significant).