Volume 52 | Number 3 | June 2017

Abstract List

James E. Graham Ph.D., D.C., Janet Prvu Bettger Sc.D., Steve R. Fisher Ph.D., P.T., Amol M. Karmarkar Ph.D., M.P.H., Amit Kumar P.T., Ph.D., M.P.H., Kenneth J. Ottenbacher Ph.D., O.T.R.


To examine changes in facility‐level risk‐standardized rehospitalization rankings for postacute inpatient rehabilitation facilities after modifying two model parameters.

Data Sources

We used national Medicare enrollment, claims, and assessment data to study 522,260 patients discharged from inpatient rehabilitation in fiscal years 2010–2011.

Study Design

We calculated risk‐standardized 30‐day unplanned rehospitalization rates for 1,135 inpatient rehabilitation facilities using four approaches. The first model replicated the current postacute risk‐standardization methodology and included patients discharged from acute hospitals up to 30 days prior to postacute admission and excluded patients transferred directly back to acute hospitals following rehabilitation. Our alternative models excluded patients with delayed admissions (>1 day between acute discharge and postacute admission) and counted direct transfers back to acute as rehospitalizations.

Principal Findings

Excluding patients with delayed admissions and counting direct transfers back to acute care as rehospitalizations substantially impacted rankings of more than half the postacute providers: 29 percent had better and 27 percent had worse quintile rankings.


Changing the timeframes for duration to admission and rehospitalization will have profound effects on postacute provider quality performance ratings. Reporting rehospitalization rates is an important issue with the explicit goal of improving the quality of postacute care. Research is needed to understand and minimize potential unintended consequences of this quality metric.