Abstract
BACKGROUND: Dual-task assessments, including Walking While Talking (WWT) and Stops Walking While Talking (SWWT) tests, predict fall risk in stroke survivors. However, their effectiveness relative to established predictors, such as the Berg Balance Scale (BBS) and Falls Efficacy Scale (FES), remains unclear. This study evaluated the comparative predictive value of WWT and SWWT tests alongside BBS and FES among stroke survivors. METHODS: This cross-sectional study included 68 stroke survivors who completed WWT-Simple (WWT-S), WWT-Complex (WWT-C), and SWWT, as well as the BBS and FES. Spearman correlations assessed relationships between balance, fear of falling, and dual-task performance. Logistic regression identified fall risk predictors, and Receiver Operating Characteristic (ROC) analysis evaluated predictive accuracy. The study adhered to STROBE guidelines. RESULTS: BBS scores were strongly negatively correlated with WWT-S (r = -0.734, p < 0.0001) and WWT-C (r = -0.737, p < 0.0001), indicating poorer balance with slower dual-task completion. Positive correlations were found between WWT-S and FES (r = 0.668, p < 0.0001) and WWT-C and FES (r = 0.610, p < 0.0001), linking slower completion times with higher fear of falling. SWWT was significantly negatively correlated with BBS (r = -0.625, p < 0.0001). WWT tests had higher sensitivity (97.8%) and specificity (99%) than SWWT (sensitivity = 68.9%; specificity = 91.3%). Logistic regression identified SWWT (Positive) as a significant predictor of fall risk (p = 0.009), and ROC analysis showed an AUC of 0.911, indicating excellent predictive power. CONCLUSIONS: Findings highlight the superior predictive value of WWT tests over SWWT in assessing fall risk among stroke survivors. Incorporating dual-task measures into clinical practice may enhance fall risk evaluation, supporting targeted stroke rehabilitation.