Telephonic Barthel Index (BI) assessment is less time-consuming and more feasible than a face-to-face interview. The aim of this study was to test the validity as well as reliability of the BI administered by telephone in comparison with face-to-face assessment in a multi-centric study. The study was conducted during the course of a randomized controlled trial in which 120 patients with subacute strokes from five teaching hospitals from different parts of India were recruited. Central telephonic follow-up and face-to-face assessment of BI and modified Rankin Scale (mRS) at 3 and 6 months were done by trained and certified blinded researchers. Kappa or weighted kappa (wK) was estimated. Sensitivity and specificity at various cutoff levels of telephonic BI were calculated. Concurrent validity of the telephonic BI was assessed by correlating it with the mRS and National Institutes of Health Stroke Scales (NIHSS) at 3 and 6 months. We observed high sensitivity and specificity at various cutoff levels of BI. Moderate to substantial agreement was observed between the two methods at 6 months wK 0.72 (95% CI 0.70-0.77). Item-wise and center-wise kappa also reflected substantial agreement. The study shows that telephonic assessment of activities of daily living with the BI in moderate to severely disabled stroke patients is valid and reliable compared to face-to-face assessment. Our study shows that telephonic assessment requires smaller sample size compared to face-to-face assessment of BI.