Development of a Short Version of the Thai Health Literacy Assessment: Application of the Item Response Theory (IRT)

Abstract

Objective: To develop and assess the properties of a short version of the Thai Health Literacy Assessment Using Word List with Extended Questions to Test Comprehension (STHLA-W+) by using the Item Response Theory (IRT). Method: The study analyzed data from a previous study which used the THLA-W+ to survey health literacy (HL) of 1,310 subjects in 7 southern provinces in order to select items by considering their IRT based parameters. Subsequently the selected items were tested for validity, reliability, and cut-off score of STHLA-W+ for interpreting HL level in a new group of subjects consisting of 200 residents of Yala. Results: All 48 questions passed the tests for basic assumptions of the IRT and were consistent with the two-parameter logistic model. Item selection based on discrimination (a), difficulty (b) and item elimination based on differential item functioning (DIF) from the influence of education and age resulted in a 19-items short version of the measure. Evaluation of the STHLA-W+ in a new group of subjects revealed its good reliability with good Cronbach’s alpha 0.808. Its validity was supported by positive correlation coefficients between the STHLA-W+ score and other HL indicators ranging from 0.20 to 0.37. Additionally, the subjects with higher education levels scored higher on the STHLA-W+. The measure showed a moderate ability to distinguish individuals with adequate and insufficient HL (AUC = 0.67 to 0.73). The cut-off point was 16, i.e., those with the score less than or equal to 16 points out of 19 were considered to have insufficient HL. Conclusion: The application of the IRT in the development of a short version of the measure resulted in a scale with item length at 40% of the total number of items in the original version. The scale took an average of no more than 4 minutes to complete test and can be used for a preliminary HL assessment in the public.