Development of a health literacy prediction model: Using HLS-EU-Q16 questionnaire data from a population-based survey

BackgroundHealth literacy (HLIT) has been linked to both individual- and population-level outcomes; however, the large-scale measurement of HLIT presents practical challenges. This study aimed to develop a predictive model for HLIT using the 16-item European Health Literacy Survey Questionnaire (HLS-EU-Q16), a standardised survey instrument created by the HLS-EU consortium, to facilitate assessment and comparison of HLIT across countries. MethodsData were derived from the Korea Health Panel Survey, conducted between March and July 2021, involving a final sample selection of 9,510 participants. Variables were selected based on previously demonstrated theoretical and empirical relevance, and backward multiple linear regression was conducted to develop the prediction model on 70% of the data, which was later validated on the remaining 30%. ResultsThe final model identified age, sex, education level, and annual household income as significant predictors of HLIT scores while maintaining model parsimony. The model explained approximately one-third of the variance in HLIT within the test set with a weighted R2 of 0.34 and showed a root-mean-square error of approximately 3.73 points in predicting the 0-16 HLIT score. Pearsons correlation analysis further supported the validity of the model, showing a strong positive correlation between the observed and predicted HLIT scores (r = 0.63, p < 0.001). ConclusionThe findings of our study provide a feasible model for predicting HLIT scores from population-based data, in this case, Korea Health Panel Survey data in combination with the HLS-EU-Q16, allowing for the effective measurement of HLIT, which would otherwise be time-consuming. This model has implications for use in cross-country comparisons of HLIT and in helping to inform public health policies and planning.