Functional Performance Tests to Predict Limited Community Ambulation in Stroke Survivors with Supervised Indoor Walking: Discriminative and Predictive Validity

Abstract

Background

Achieving community ambulation is a critical milestone in post-stroke rehabilitation, particularly for patients who can ambulate indoors with supervision (Functional Ambulation Category [FAC] 3). However, objective criteria to predict the transition to limited community ambulation (FAC 4) during the subacute phase remain inadequately defined.

Objective

To examine the discriminative and predictive validity of selected functional performance tests for identifying the potential to achieve limited community ambulation in stroke patients with supervised indoor walking ability.

Methods

This retrospective study included 52 subacute stroke patients with FAC 3-level ambulation. Participants completed a battery of functional assessments categorized into gait speed, endurance, balance, and activities of daily living (ADL)— including the 10-Meter Walk Test (10mWT), 6-Minute Walk Test (6MWT), Berg Balance Scale (BBS), Four Square Step Test (FSST), Activities-specific Balance Confidence Scale (ABC), Modified Barthel Index (MBI), among others. Receiver operating characteristic (ROC) curve analysis and logistic regression were employed to determine optimal cutoff values and significant predictors for achieving FAC 4.

Results

The 6MWT (cutoff: 99.35 m; AUC = 0.980) and 10mWT (cutoff: 0.315 m/s; AUC = 0.953) demonstrated the highest predictive accuracy. Logistic regression identified 6MWT (OR = 1.156, p = .029) and FSST (OR = 0.838, p = .040) as significant predictors. BBS, ABC, and MBI showed moderate discriminative power.

Conclusion

The 6MWT and 10mWT are valid and clinically useful tools for predicting limited community ambulation in subacute stroke patients. Early incorporation of endurance and dynamic balance measures may support individualized rehabilitation planning.