Background: A number of previous studies have compared individuals' valuations of their own health with societal valuations of the same states. Most studies have not examined these differences at the individual level (i.e. between a single respondent's valuation of both self and a hypothetical state with matching description), and there has been little empirical work on valuations in developing countries. We examined differences between self- and hypothetical valuations in a multi-country sample survey study.

   Methods: Population-representative surveys were conducted during 2000-2001 in fourteen countries: China, Colombia, Egypt, Georgia, India, Iran, Lebanon, Indonesia, Mexico, Nigeria, Singapore, Slovakia, Syria, and Turkey, as part of a World Health Organization study on health systems. In a health-state valuation module, respondents considered 10 hypothetical conditions with brief descriptions (e.g. “total blindness, acquired as an adult”), in addition to their own current health. For both “self” and hypothetical conditions, respondents provided “profiles” comprising categorical ratings along six dimensions (affect, cognition, mobility, pain, self-care and usual activities), and assigned values to these states using a visual analog scale (VAS). For individuals assigning a self-matching profile to at least one hypothetical condition, we compared VAS values between the matching states and examined determinants of the differences using multivariate regression.  

   Results: Out of 52,050 respondents 7,498 assigned at least one hypothetical state the same 6-dimensional profile as their own health. The distribution of differences in VAS scores (self minus hypothetical) had a mean of 21, with a strong positive skew (only 4% of differences were ≤0) and considerable variation across countries (highest mean = 35 in Colombia; lowest = 16 in China). The hypothetical conditions most often assigned self-matching profiles were “mild vision problems” (44%) and “mild hearing problems” (10%).  In multivariate analyses, most indicator variables for countries were significant predictors of self-hypothetical differences; sex, age and years of education were not significant, and most variables relating to levels on the six health dimensions were not significant.

   Conclusions: In this study, individuals across a diverse range of countries consistently rated their own health states higher than hypothetical states with the same profiles. These differences may have strong cultural determinants, suggested by significant cross-country variation; however, much of the variation remains unexplained even after controlling for country, socio-demographic characteristics, and health-state characteristics.