Educational inequalities in Global Activity Limitation Indicator disability in 28 European Countries: Does the choice of survey matter?

OBJECTIVES: To assess the sensitivity of prevalence and inequality estimates of Global Activity Limitation Indicator (GALI) to the choice of survey in European countries. METHODS: We use logistic regression to estimate adjusted risk ratios, quantifying differences in prevalence and educational inequalities, the impact of survey characteristics and Kendall’s tau to assess similarity in country rankings between surveys. We include the European Health Interview Survey (EHIS), European Social Survey (ESS) and European Union Statistics on Income and Living Conditions (EU-SILC). RESULTS: EHIS estimates higher prevalence than EU-SILC 17% (men) and 23% (women), and ESS 24% (men) and 29% (women). Prevalence does not differ significantly between EU-SILC and ESS. EU-SILC estimates 52.5% (men) and 28.1% (women) higher inequalities than EHIS and 63.2% (men) and 32.7% (women) higher inequalities than ESS. Survey characteristics do not account for differences in prevalence or inequalities. Country rankings do not agree for prevalence or inequalities. CONCLUSIONS: Survey choice strongly impacts estimates of GALI prevalence and educational inequalities. Further study is necessary to understand these discrepancies. Caution is required when using these surveys for cross-country comparisons of (educational inequalities in) GALI disability. ELECTRONIC SUPPLEMENTARY MATERIAL: The online version of this article (10.1007/s00038-018-1174-7) contains supplementary material, which is available to authorized users.