A systematic review of studies reporting multivariable models to predict functional outcomes after post-stroke inpatient rehabilitation

Abstract

Purpose: This systematic review summarizes the utility of variables available at acute discharge after stroke for predicting functional independence at discharge from inpatient rehabilitation. Methods: A systematic review of four electronic databases (Medline, EMBASE, PsycINFO and CINAHL) was conducted to identify studies reporting multivariable models predicting post-rehabilitation Barthel Index (BI) or Functional Independence Measure (FIM®) scores. In studies meeting inclusion criteria, the frequency with which candidate predictors were found statistically significant was calculated and summarized. Results: A total of 3260 articles were screened, of which 27 were included and 63 multivariable models of discharge BI or FIM® were reported. In all, 126 candidate predictors of BI or FIM® were explored. Variables found to be significant most frequently included admission functional level (BI or FIM®), National Institute of Health Stroke Scale (NIHSS), dysphasia, impulsivity, neglect, previous stroke, and age. Conclusions: Only a selected group of variables have repeatedly proven to be significant predictors of functional ability after post-stroke inpatient rehabilitation.

Implications for Rehabilitation

• This review identifies, and summarizes, studies that have used a multivariable model to predict Barthel Index (BI) or Functional Independence Measure (FIM®) after post-stroke inpatient rehabilitation.

• Clinicians making decisions about admission to inpatient rehabilitation should consider age and a measure of stroke severity, functional status and progress to date.

• Variables that have been demonstrated to be useful most commonly included age, admission functional level (BI or FIM), National Institute of Health Stroke Scale, dysphasia, impulsivity, neglect, and previous stroke.

• Barthel Index
• cerebrovascular accident
• functional independence measure
• stroke,

Acknowledgements

The authors would like to thank John Costella from the Allyn and Betty Taylor Library at Western University, London, Ontario, for his guidance in designing the search strategy used in this study.

Declaration of interest

The authors report no declaration of interest.