Development of Thai Health Literacy Assessment Based on the Assessment of Ability to Use Nutrition Label

Abstract

Objective: To develop and test the screener of health literacy (HL) for Thai people using nutrition label in the measurement, called the Thai Health Literacy Assessment Instrument: Nutrition label (THLA-N) and to find cut-off for interpret the scores. Methods: The THLA-N was developed to evaluate HL by using 7 questions that measured the ability to use nutrition labels in the same approach the Newest Vital Sign, together with 3 HL screening questions. Five experts (researcher on construction of measurement and evaluation, physician, pharmacist, nurse and public health official) assessed face validity of the scale. Subsequently, the scale was tested to find potential misunderstandings by using think aloud techniques in 15 subjects with different levels of education. Later, the researchers conducted pilot-test of the scale in 30 outpatients and, subsequently, in 497 outpatients at a regional hospital. Validity was assessed by examining differences of the THLA-N between people with different levels of education and the relationship between the THLA-N and reading ability and the ability to understand health documents. Cut-off point for interpreting scores was determined using analysis of ROC (receiver operating curve) with the sum of reading ability scores and the ability to understand health documents as gold standard 1 (gold standard 1: GS1) and  ability to understand health documents as GS2. Results: Two questions in the THLA-N having less than 5 percent of the respondents with correct answers were removed from the scale. The score calculated from 8 remaining questions was called THLA-N8. The 8 item scale’s reliability was 0.70. Validity of the scale was evident by those who have higher education levels having greater THLA-N8 scores than those with lower education levels (P <0.001). The THLA-N8 showed positive and statistically significant correlation with 2 indicators of HL, i.e. understanding of health documents and reading ability (r = 0.48 and 0.21, respectively). In the ROC curve analysis, the THLA-N8 scale had area under the curve at 0.76 and 0.81 when using GS1 and GS2 as gold standards, respectively, indicating that the scale could distinguish those with sufficient and insufficient HL. The THLA-N8’s score less than or equal to 4.83 indicated inadequate HL. Sensitivity of the scale was 0.71 and 0.75 when using GS1 and GS2 as gold standards, respectively. The specificity of the scale was 0.67 and 0.73 when using GS1 and GS2 as gold standards, respectively. Conclusion: The THLA-N8 shows satisfactory reliability, validity, sensitivity and specificity.  However, there should be further studies in other groups of subjects to confirm its psychometric properties and cut-off point for interpretation of the scores.