Development of the Thai Health Literacy Assessment Based on the Testing of Essential Skills

Abstract

Objective: To develop the Thai Health Literacy Assessment based on Skills: THLA-S), and to determine the cut-off value to interpret the level of health literacy (HL). Methods: There were 5 steps in development of this measure:1)analysis of skills assessed by the HLSI (Health Literacy Skills Instrument), the measurement model of THLA-S, 2) preparation of the first draft of the THLA-S for measuring the skills equivalent to those assessed by the HLSI, 3) review of the measure by 6 experts and revision of the measure accordingly, 4) validity and reliability testing in large samples in a regional hospital with 522 outpatients and their relatives chosen by accidental sampling. The subjects completed the THLA-S and questionnaire on various indicators of HL (self-perception on skills for activities necessary for HL, reading ability from cloze test, HL screening test, and ability to use nutritional label), and THLA-N (Thai Health Literacy Assessment Using Nutrition Label), 5) determination of cut-off of the THLA-S using THLA-N as gold standard. Results: The THLA-S showed a Cronbach’s alpha of 0.81. The measure was valid because 1) In assessing self-perception on skills for activities necessary for HL, those reporting 6 activities (from all 8 activities) as easy had a higher level of the THLA-S than those reporting the tasks difficult 2) those with a higher education level showed a higher level of the THLA-S 3) relationships between the THLA-S scores with the following HL indicators were positive and statistically significant (P<0.05)--self-perception on skills for activities necessary for HL, reading ability as measured by the cloze test, scores from screening tests for inadequate HL, ability to use nutrition labels and the THLA-N with correlation coefficients of 0.34, 0.40, 0.16, 0.63 and 0.61, respectively. Analysis of the ROC curve using the THLA-N as a gold standard showed an AUC of 0.84. The THLA-S scores less than or equal to 13 indicated inadequate HL. At this cut-off, the scale had a sensitivity of 75.0 percent, specificity of 78.4 percent, PPV of 79.12 percent and NPV of 74.19 percent, with LR+ and LR- of 3.47 and 0.32 respectively. Conclusion: The THLA-S is valid, reliable, sensitive, and specific with satisfying discriminating power of HL. However, the THLA-S still needs further improvement because of its large number of items.