Population norms for the EQ-5D-5L for Hungary: comparison of online surveys and computer assisted personal interviews

Background and objectives The aims of this study were to provide population norms for EQ-5D-5L in Hungary and investigate the differences in EQ-5D-5L normative data by survey mode, i.e. online surveys and computer assisted personal interviews (CAPI). Methods A pooled database was built comprising six online (N = 7,034) and two CAPI (N = 3,020) population-based studies with the EQ-5D-5L. Descriptive statistics were performed. Multinominal logistic and linear regression analyses were applied to compare the online and CAPI samples. Traditional and machine learning regression tools were used to investigate the determinants of EQ-5D-5L index values. Results ‘No problems’ in any of the five EQ-5D-5L domains were reported by 33.9% (online) and 58.9% (CAPI) of the participants. Most problems were reported on the pain/discomfort domain in both study types (51.9% and 33.6%, respectively). Men and more educated respondents had significantly higher average EQ-5D-5L index values. EQ-5D-5L index values and EQ VAS scores were significantly higher in the CAPI sample, except in age groups 65–74 (no difference) and 75+ (online scores were significantly higher). Only 7–10% of variance in the EQ-5D-5L index values was explained by the variables survey mode, education, sex and age, with age having the largest and sex the smallest effect. Conclusions EQ-5D-5L population norms derived from online and CAPI studies may differ significantly from each other. It is recommended to consider the survey mode, sampling and sociodemographic characteristics of the participants when choosing population norms as reference set. Further comparative studies investigating EQ-5D-5L population norms by different study designs and administration modes are encouraged.