Validation of a modified MOCA translation for use among Arabic-speaking immigrants in the U.S.

Abstract

BackgroundThe only validated, widely-used dementia screen that has Arabic language norms/cutoffs is the Montreal Cognitive Assessment (MoCA). Yet, Arabic translations of the MoCA vary across countries. This study considers various Arabic translations of the MoCA, and presents a modified translation for use among Arabic-speaking immigrants in the U.S.MethodThe modified translated version of the MoCA was administered to 32 Arabic-speaking adults age 65+ living in metro-Detroit. Eight (25%) had an ADRD diagnosis. To assess the reliability of the MoCA, each item was standardized and Cronbach’s alpha was calculated. To assess the similarity of ADRD and non-ADRD respondents with regards to each MoCA item as well as with regards to demographics we used Fisher’s exact test for binary variables and t-test for continuous variables. Ordinary least squares models were used to examine how an ADRD diagnosis predicts the MoCA score, adjusting for demographics.ResultThe mean age of the sample is 73 years old. Sixty-two percent (62%) are female and 28% have a high school education or more. The alpha was acceptably high at .87. The MoCA item-level scores for respondents with and without ADRD diagnoses showed that all respondents correctly identified the picture of a camel. There are also five items for which none of the ADRD respondents gave correct responses: trail-making, verbal fluency F, both abstraction questions, and the cube-copy test. Bivariate analyses further indicate that ADRD respondents are older than non-ADRD respondents (p<.001). There is no significant difference in gender or education level. Those with ADRD diagnosis scored lower overall on the MoCA ( = -.35; se = .37), with 58% of the variation explained by the diagnosis and demographics. We interpret the result as a small effect size as indicated by the Cohen’s heuristic for the difference between means (0.2-0.49).ConclusionThe Arabic language MoCA is able to distinguish respondents with an ADRD diagnosis. Decision points about translations should consider national contexts to maximize equivalencies across samples.