Cross-Cultural Adaptation and Psychometric Evaluation of the Persian Version of the ADL Taxonomy Questionnaire in...

Cultural adaptation of assessment tools is essential to ensure their accuracy and relevance across different populations.

The study focused on cultural adaptation and psychometric evaluation of the Persian ADL Taxonomy for assessing daily activities in persons with stroke.

Data were collected in a clinical setting and a university hospital in Tehran.

The ADL Taxonomy was translated into Persian in accordance with standard guidelines. Face validity was tested with 20 stroke patients and convergent validity with 104. Validity and test-retest reliability were analyzed using SPSS, Spearman correlation with MBI scores, and Pearson correlation for reliability(p < 0.05). Rasch analysis was performed in R (Version 4.4.1) using the eRm and TAM packages with the 1PL model for dichotomous data, analyzing each of the 46 actions separately. Item difficulty, standard errors, model fit (Infit MnSq ≤ 1.3; t = -2 to +2), and discrimination indices (≥ 0.30) were calculated. Person ability estimates were obtained using the EAP method, with PSI (≥ 2.0) and PR (≥ 0.80) calculated and raw scores examined to assess the scale's reliability, discriminative ability, and potential ceiling or floor effects.

Face validity confirmed item clarity, with minor edits. Impact scores ranged from 2.70 to 5.00. Content validity indices were high (CVR = 0.54-1.0; $\begin{array}{c}\text{I}-\text{CVI}=\mathrm{0.841.0}-\end{array}$ ; $\begin{array}{c}\text{S}-\text{CVI}/\text{UA}=91.3\%\end{array}$ ; $\begin{array}{c}\text{S}-\text{CVI}/\text{Ave}=0.99\end{array}$ ; modified kappa = 0.84-1.0). Convergent validity showed a strong correlation with the MBI (r = 0.89, p < 0.001), and test-retest reliability was excellent (ICC = 0.98). Rasch analysis of items showed a wide range of difficulty (-4.77 to 6.61 logits) with acceptable measurement precision, generally good model fit, and adequate discrimination for most items, although a few (e.g., communication and transportation and traveling) displayed misfit and low discriminatory power. Person ability estimates ranged widely (2-46, median = 34, mean = 32.05, Q1 = 25.75, Q3 = 42), with a person separation index of 3.95 and person reliability of 0.94.

The Persian version of the ADL Taxonomy demonstrated excellent validity and reliability, with most items showing good fit and discrimination. The scale reliably distinguished between five levels of ability, and no ceiling or floor effects were observed in the Rasch analysis, supporting its use for assessing daily activities in Persian-speaking stroke patients.