Understanding the relationship between income and health: How much can be gleaned from cross-sectional data?

The aim of this paper is to develop a better understanding of the relationship between income and health using a cross-sectional survey of the general British population, It is divided into two parts. First, it examines a number of methodological inconsistencies in the existing literature and assesses their consequences for the inferences that can be drawn about the income health association. These issues include: the measurement of income and its functional form; health selection; and the role of confounders. Second, it explores the relative strength of the income-health association in contrast to that of other socioeconomic measures. The relationship between income and health is complex. However, having taken into account a range of methodological problems, income is still significantly associated with health. The association appears to be non-linear and is attenuated but not removed by controlling for health selection effects. The inclusion of a wide range of confounders into models of income and health reduces the association between them, but does not make it insignificant. In comparison to other socioeconomic measures, income appears to be a better discriminator of health status than education or occupation. However, tenure and car ownership seem to be at Last as good if not better than current income.