Functional Independence predicts patients with stroke more likely to be discharged to the community after inpatient rehabilitation

ABSTRACT

Background

Functional domain predictors of discharge destination following inpatient rehabilitation for stroke have not been thoroughly identified.

Objectives

1) Determine the relationships between intrinsic variables (demographic; comorbidities; functional independence at admission to and at discharge from an inpatient rehabilitation facility (IRF)) and discharge to home. 2) Determine cut scores for Functional Independence Measure® (FIM) subscales and domains that predict discharge to the community.

Methods

This study was a secondary analysis of a large, multi-IRF dataset from the Uniform Data System for Medical Rehabilitation. Participants were adults with stroke who were discharged from an IRF in 2019 (n = 92,153).

Results

Correlations with discharge to the community were strongest for discharge FIM scores (r = 0.330 to 0.580), followed by admission FIM scores (r = 0.245 to 0.411), which were stronger than the demographic and comorbidity variables (r = 0.005 to 0.110). Logistic regression analysis indicated 5 of 6 FIM domains (Social Cognition, Self-care, Sphincter, Transfer, and Locomotion) scored at admission and at discharge were predictive of discharge home. Receiver operating characteristic curve analyses determined the best cut point for each domain. For each FIM measure, the area under the curve was greater when the measure was obtained at discharge than it was at admission.

Conclusions

Clinicians may consider the cut points presented for each domain at admission and at discharge when setting goals or making recommendations for patients with stroke who aspire to a discharge from an IRF to a community setting.

KEYWORDS:

• Stroke
• rehabilitation
• functional independence
• prediction
• patient discharge
• community

Disclosure statement

No potential conflict of interest was reported by the author(s).

Data Availability

The data set associated with this study was obtained from the Uniform Data System for Medical Rehabilitation®, a division of UB Foundation Activities, Inc.

Suppliers

^a Uniform Data System for Medical Rehabilitation, a division of UB Foundation Activities, Inc.

^b IBM SPSS Statistics (v. 26)

^c MedCalc (v. 19.7.2)

^d Microsoft Excel (v. 2016)

Additional information

Funding

This work was supported by the Elon University Faculty Research & Development Fund.