Abstract Background Understanding and utilizing health information has become increasingly demanding whereby people's health literacy becomes important. European Health Literacy Survey Questionnaire (HLS-EU-Q16) measures health literacy (HL). We aimed to study the psychometric properties of the Danish version in the general population and a population with coronary artery disease. Methods A national, cross-sectional survey of individuals from the general Danish population and patients with Myocardial Infarction (MI) was conducted. The general population sample consisted of 7,653 individuals. The MI sample of 4,440 individuals had previously been hospitalized with MI. To assess the influence of large sample sizes, we examined five randomly drawn subsamples of 500 from each population. The Rasch analysis was conducted using RUMM-2030 consisting of inspection of unidimensionality, local dependency, Differential item Functioning (DIF), and item fit. Results The results are preliminary and analyses still ongoing. The HLS-EU-Q16 did not fit the Rasch model, neither for the total sample, subsamples, the general population sample, nor for the MI sample. For the MI sample, there were signs of local dependence and item 6 had the largest item misfit. DIF was observed for item 6 and 7 (age and gender). The general population sample also showed signs of local dependence, and item 1 indicated the largest item misfit. DIF was observed for item 1 (age) and item 1, 7, 8, 13 and 16 (gender). Conclusions The Danish version of HLS-EU-Q16 did not fit the Rasch model neither in the general population sample nor for the MI sample. Our large population-based study indicated scalability problems of the HLS-EU-Q16. The validation procedure is a step on the path to ensure the use of valid measures of HL in both populations and patient groups to support the raising awareness of HL as a necessary element to ensure equity in our health system and thereby initiating an urgent public health activity. Key messages We studied the construct validity of the Danish version of the European Health Literacy Survey Questionnaire HLS-EU-Q16 with a Rasch model. Our large population-based study indicate problems with the psychometric properties of the current Danish version of the HLS-EU-Q16.