An Algorithm, Integrating a Short Form of the Functional Autonomy Measurement System, to Predict Discharge Destination After Acute Care Post-Stroke

Abstract

Purpose: This study develops a short form of the Functional Autonomy Measurement System (SMAF), the SF-SMAF, for measuring functional capacity in patients undergoing acute care post-stroke, identifies predictors of the discharge destination chosen by the care team, and derives an algorithm that integrates the SF-SMAF and other predictors to guide discharge planning. Method: This multisite prospective cohort study involved 200 patients assessed with the SMAF within 8 days post-stroke. Sociodemographic and clinical data were extracted from patients’ medical records. We performed linear regressions to identify subsets of SMAF items that closely approximate the SMAF total score and asked a panel of experts to make the final selection. We used logistic regression to develop an algorithm that predicts discharge destinations using the SF-SMAF and other predictors. Results: The SF-SMAF includes four items: “washing”, “walking inside”, “judgment”, and “budgeting”. It is highly correlated with the SMAF ( R2 = 0.94) and, alone, predicts 71% of discharge destinations. Adding obstacles to returning home, support required from caregivers, and the ability to communicate, raises the prediction of the proposed algorithm to 82%. Conclusions: The SF-SMAF results closely approximate those of the SMAF in the first week post-stroke. Following further validation, the proposed algorithm could guide clinicians in using the SF-SMAF for discharge planning.