Robert W. Yeh, MD, MSc, from the Smith Center for Outcomes Research in Cardiology at Beth Israel Deaconess Medical Center in Boston, Massachusetts, and colleagues, conducted a cross-sectional population-based study of hospitals participating in the all-payer Nationwide Readmission Database in 2013 or 2014, in order to examine whether hospital level readmission measures over 30 days for Medicare participants with publicly-reported conditions are similar to those for participants hospitalized with unreported conditions.
The study included 2101 hospitals and 953,086 Medicare patient hospital admissions; reported conditions included heart failure, acute myocardial infarction, and pneumonia. There were more than 7 million (n=7,121,223) Medicare patient hospital admissions for unreported conditions and 446,250 admissions for non-Medicare patients (risk-standardized unplanned readmission rate [RSRR], 19.8%, 17.4%, and 14.6%, respectively.)
Excess readmission ratios (ERRs) were also compared for Medicare hospitalizations for reported and non-reported conditions. Within-hospital differences between these groups ranged from -0.59 to 0.32 overall; the ranges were largest in metropolitan non-teaching hospitals, private hospitals, and medium-sized hospitals (-0.59 to 0.30; -0.59 to 0.26, and -0.59 to 0.30, respectively).
Researchers found that among hospitals with high readmission ratios, ERRs for the Medicare-reported group “tended to overestimate ERRs for the non-Medicare group but underestimate those for the Medicare unreported group.”
The study ha several limitations. The data set did not allow for the use of a risk-adjustment strategy similar to the one used by the Center for Medicare and Medicaid services, and hospitals could not be linked over the course of multiple years due to changes in hospital identification numbers, leading to double inclusion of some hospitals.
Dr Yeh and colleagues noted that there were “sizeable differences in hospital performance when we compared 30-day readmissions after hospitaliz[ed]ation for Medicare reported conditions with those for unreported conditions and non-Medicare reported conditions.” They continued, “These discrepancies translated into substantial disparities in the numbers of hospitals penalized or not penalized when different groups were used.”
“Current public measures of hospital performance for 30-day readmission rates may not reflect hospital quality of care for unreported populations and conditions,” they concluded.
Butala NM, Kramer DB, Shen C, et al. Applicability of publicly reported hospital readmission measures to unreported conditions and other patient populations [published online March 26, 2018]. Ann Intern Med. doi:10.7326/M17-1492