Benchmarking length of stay for inpatient stroke rehabilitation without adversely affecting functional outcomes

Abstract

Objective: To evaluate the effects of introducing the practice of targeting a discharge date for patients admitted to an inpatient stroke rehabilitation unit on process and patient outcomes. Design: Comparison of retrospective (control group n = 69) and prospective (experimental group n = 60) patients. Methods: Rehabilitation professionals assessed both groups at admission and discharge using a standard-ized assessment toolkit. Benchmarks for length of rehabilitation stay (LoRS) were introduced based on median severity-specific LoRSs in the control group. The multidisciplinary team documented facilitators and obstacles affecting the prediction of patient benchmark attainment. Categorical variables were compared using a χ2 test with exact probabilities. Ordinal and continuous variables were analysed using rank-based non-parametric analysis of variance. Effect sizes were estimated using a relative treatment effect statistic. Results: The mean combined length of stay in acute care and rehabilitation beds for the experimental group (82 days) was shorter (p = 0.0084) than that of the control group (103 days). This 21-day reduction in combined length of stay included a 10-day reduction in the mean time between stroke onset and admission to the stroke rehabilitation unit (p = 0.000014). Improvements in 6 func-tional and sensorimotor outcomes with rehabilitation were of similar magnitude in both groups, while Functional Independence Measure (FIMTM) efficiency improved (p = 0.022). The team was 87% successful in predicting which patients were discharged on the LoRS benchmark. Conclusion: Benchmarking the length of stay in rehabilitation resulted in reduced bed occupation and system costs without adversely affecting functional and sensorimotor patient outcomes.

Keywords

• stroke
• inpatient rehabilitation
• quality improvement
• length of stay
• benchmarking