Estimating the minimal important change of WHODAS 2.0 in individuals with chronic stroke: a methodological comparison

Assessing the state of functioning and disability post-stroke is essential for managing treatment and allocating appropriate resources. To explore the minimal important change (MIC) of the World Health Disability Assessment Schedule 2.0 (WHODAS 2.0) in individuals with chronic stroke using different statistical approaches. WHODAS 2.0 was administered at baseline and after six months. MIC was estimated through distribution-based methods, anchor-based approaches, and ROC curve analysis using the Youden's index. A logistic regression model was also applied to estimate a predictive MIC (MIC_pred). Positive and negative predictive values (PPV and NPV) were estimated for three MIC thresholds. Fifty participants with chronic stroke were included (48% male; mean age approximately 60 years). Most (78%) had mild disability, and 80% showed functional improvement after six months. MIC estimates varied: 1.6 points (standard error of measurement), 5.28 points (0.33 × SD), and 8 points (0.5 × SD). ROC analysis yielded an AUC of 0.67, Youden's index of 0.35, sensitivity of 75%, specificity of 60%, and MIC ≤ 1.89 points. MIC_pred was 10.3 points (Nagelkerke R2 = 0.078). The threshold of 5.28 points showed the best PPV (90.67%) and modest NPV (26.32%). MIC values for WHODAS 2.0 in chronic stroke differ by method. The 5.28-point threshold offered the most balanced predictive value but should be interpreted cautiously. The findings support using multiple methods and emphasize the need for patient-centered assessment and methodological consistency in MIC determination.