Reliability，Validity and Agreement of the Short Form Modified Extended Barthel Index for Assessing Activities of Daily Living of Patients with Stroke

ObjectiveTo construct a short form Modified Extended Barthel Index (MEBI), and to verify its reliability and validity in assessing activities of daily living (ADL) of patients with stroke and its agreement with MEBI.MethodsA total of 267 stroke patients from the Kunshan Rehabilitation Hospital were included. Within 3 days of admission, MEBI, Barthel Index (BI), Modified Barthel Index (MBI), Extended Barthel Index (EBI), and Functional Independence Measure (FIM) assessments were used to evaluate ADL of 175 patients as the analysis set for short-form development, while the remaining 92 patients completed the MEBI alone as the validation set for agreement analysis. Multiple linear regression models (stepwise method) were used to extract core MEBI items to construct short-form MEBI (model 1-16). The optimal short form MEBI was determined through internal reliability analysis and validity evaluation (content validity, criterion validity, and construct validity). In the validation set, paired mean differences and intraclass correlation coefficients (ICC) were calculated to assess agreement between short form MEBI and full MEBI scores.ResultsShort-form models (model 5-10) were constructed. In the analysis set, the internal reliability Cronbach's α and Spearman-Brown coefficients of the short-form MEBI (models 6-10) were greater than 0.8, and the content validity ρ correlation coefficients between the scores of each assessment item and the total score of the model were greater than 0.4 (<italic>P</italic>&lt;0.001); criterion validity ρ correlation coefficients between short-form MEBI (models 6-10) and MEBI, EBI, FIM, MBI, BI were greater than 0.8 (<italic>P</italic>&lt;0.001). The short-form MEBI (Model 7) demonstrated optimal construct validity (KMO=0.837; Bartlett's test <italic>P</italic>&lt;0.001), with two factors cumulatively accounting for 79.3% of the total variance: factor 1 (bed-wheelchair transfer, limb-trunk hygiene, facial hygiene, ambulation), factor 2 (problem-solving, comprehension, bladder control); the seven items of the two common factors constituted the evaluation items of the short form MEBI. In the agreement validation set, the mean paired difference between the short form MEBI and the full MEBI was 4.13 (95% <italic>CI</italic>: 2.70, 5.56), and the mean ICC was 0.969 (95% <italic>CI</italic>: 0.919, 0.984).ConclusionThe short form MEBI exhibits excellent reliability, validity, and agreement with the full MEBI, which can provide a user-friendly tool for ADL assessment in stroke patients.