Hospital readmission rates do not vary widely among primary care physicians

Among 4,230 primary care physicians in Texas, one had a 30-day readmission rate significantly higher than the mean of 12.9%, while none had a significantly lower rate.

Hospital readmission rates have very low variation among primary care physicians, and therefore federal policy and other programs holding them accountable for readmission rates may be ineffective, a new study indicates.

Researchers performed a retrospective cohort study of Texas Medicare claims from 2008 to 2015. Patients were included if they were discharged alive from the hospital between Jan. 1, 2008, and Nov. 30, 2015, and had had a primary care physician in the previous year who had at least 50 admissions in the study period. The study defined a primary care physician as a generalist (that is, a general practitioner, family physician, internist, or geriatrician) who saw a patient two or more times in an outpatient setting in the year before hospitalization. The study's main outcome measures were readmission within 30 days of discharge and follow-up visits with a primary care physician within seven days of discharge, with separate cohorts used to evaluate each. The study results were published May 21 by Annals of Internal Medicine.

The mean risk-standardized rate of 30-day readmissions was 12.9% between 2012 and 2015. Among the 4,230 primary care physicians included in both cohorts, one had a readmission rate significantly higher than the mean, while none had a significantly lower rate. The 10th and 90th percentiles were 12.4% and 13.4%, each of which differed from the mean by only 0.5 percentage point, and the 99th percentile was 14.0%, which was 1.1 percentage points higher than the mean. The authors calculated that more than 3,500 admissions per primary care physician per year would be required in order to detect a difference of 1.1 percentage points from the mean. For the 3,603 physicians in the seven-day follow-up cohort, the overall rate of follow-up within seven days among 413,527 patients was 20.4%.

The study included only fee-for-service Medicare patients in one state, and readmissions could not be classified as avoidable versus unavoidable, among other limitations, the authors noted. However, they concluded that in analyses controlling for patient characteristics, there appears to be “negligible variation” in primary care physicians' readmission rates.

“Our finding of minimal variation in risk for readmission among [primary care physicians] calls into question any pay-for-performance program that aims to reduce readmissions and assumes variation by [primary care physician],” the authors wrote. “Furthermore, our analysis indicates that the threshold used in the Merit-based Incentive Payment System of 200 or more readmissions per year is far too low to distinguish real-world differences among [primary care physicians] or group practices.”