Are different measures of self-rated health comparable? An assessment in five European countries

Self-rated health (SRH) is widely used to compare population health across countries, but comparability is often hampered by the use of different versions of this item. This study compares the WHO recommended version (ranging from 'very good' to 'very bad') with the US version (ranging from 'excellent' to 'poor') in European countries.

Data came from the Survey of Health, Ageing and Retirement in Europe (SHARE). Both the WHO and US versions of SRH were measured in representative samples of Europeans aged 50+ (n = 11,643) in five countries. Concordance between the two SRH versions and differences in their associations with demographics, chronic diseases, functioning and depression were assessed using ordered probit regression.

The US version has a more symmetric distribution and larger variance than the WHO version. Although the WHO version discriminates better at the positive end, the US version shows better discrimination at the positive end of the scale. Sixty-nine percent of respondents provided literally concordant answers, while only about one-third provided relatively concordant answers. Overall, however, less than 10% of respondents were discordant in either sense. The two versions were strongly correlated (polychoric correlation = 0.88), had similar associations with demographics and health indicators, and showed a similar pattern of international variation.

Health levels based on different measurements of SRH are not directly comparable and require rescaling of items. However, both versions represent parallel assessments of the same latent health variable. We did not find evidence that the WHO version is preferable to the US version as standard measure of SRH in European countries.