​Functional status may be a better predictor of 30-day acute care readmission than traditionally investigated variables such as demographics and comorbidities, according to an October 1, 2016, study in the Journal of Post-Acute and Long-Term Care Medicine. The authors conducted a retrospective database analysis on 4,199,002 readmissions at 1,158 U.S. inpatient rehabilitation facilities (IRFs) between 2002 and 2011. They developed three models to determine which factors best indicate readmission risk. The first model measured functional status only, the second model used only demographic and comorbidity data, and the third model ("function plus") added function data to comorbidity and demographic data. The function-only model was significantly more likely to predict readmission than the demographic-comorbidity model, while the "function plus" model demonstrated little improvement over the function-only model, the authors said. Functional status may be a better indicator than comorbidity, the authors said, because it is an early indicator of burden of care and disease severity, whereas comorbidity diagnoses do not account for either of these. One significant limitation of the study is possible selection bias, the authors said, because the subjects were inpatient rehabilitation patients, a population that is already identified as having functional impairments. Other limitations included an inability to account for repeat readmissions and an inability to account for whether the readmission was planned or unplanned.

HRC Recommends: Risk managers and quality improvement professionals responsible for acute or postacute care facilities may wish to review the study and consider whether it is feasible for their organizations to use functional measures to assess readmission risk and implement functional interventions to reduce readmissions. As the study authors point out, robust patient functional status data are not routinely collected in U.S. acute care hospital administrative data sets. (For the study, data were obtained from the Uniform Data System for Medical Rehabilitation, a data repository of IRF patients in the United States. The data set includes IRF-Patient Assessment Instrument data, which are composed of demographic, functional, medical, and facility data for approximately 70% of U.S. IRFs.)