Validity and reliability assessment of mental health literacy scale adapted into Turkish in patients' relatives

AIM: Aim of this study was to evaluate validity and reliability of the Turkish version of Mental Health Literacy Scale (MHLS) in a different population. METHOD: The data was collected by interviewing 401 adult family members that companied their relatives who applied to Istanbul University Istanbul Medical Faculty as patients. The data was collected via Personal Information Survey and Turkish version of Mental Health Literacy Questionnaire (MHLQ). Initially item-total point correlation was tested, then both exploratory and confirmatory factor analyses were performed. After modifications were applied to achieve significant improvement in model fit, Cronbach alfa and Guttman Split Half coefficent were calculated. Mann- Whitney U, Kruskall- Wallis tests and Spearman correlation coefficent were used to evaluate variables that may be in relation with mental health literacy total point. Logistic regression was used to investigate mulivariate analysis of the data. RESULTS: Kaisser Mayer Olkin value was 0,864 and Barlett test was found to be significant (p<0,001). Four items ( 9,10,12,15) which correlated less than 0,2 after item-total point correlation were excluded from the scale. Exploratory factor analysis revealed that the same four items had their factor loads under 0,4. Five factored structure was found to be most suitable for the remaining 31 items with no items less than 0,4 factor load or cross-loading. Extraction value of items was over 0,3. After confirmatory factor analysis, another item which found to have loading less than 0,3 (item number 20) was also excluded. When the model was constructed after defining three covariants that provided significant improvement in fit index, fit indices of the model was found to be acceptable or appropriate. Female gender, being single, educated, perception of high income, having mental illness, receiving treatment for mental illness or having a relative with mental illness were found to be variables that are related to having higher mental health literacy. Level of education, perception of income and having a relative who has a mental illness were found be related to mental health literacy in multivariant analyses. CONCLUSION: Turkish version of Mental Health Literacy Scale is found to be valid and reliable in general population when used in its 30 questions form. We advise its usage in this form when asessing mental health literacy.