Abstract

The recent paper by Dunn et al showed that the positive relationship between US state-level income inequality and mortality was small in the 1950s, rose to a large value around 1990 but had largely disappeared by 2019. We consider these findings in the context of the mechanisms that have been advanced for reasons why a positive relationship might be expected, and in relation to studies using alternative methods included in systematic reviews that fail to confirm an independent inequality/mortality relationship. Ecological studies, such as by Dunn et al, using subnational data have advantages compared with similar studies using cross-national data, but controls are typically confined to those available from sources such as decennial census, so scope for incorporating lagged effects and life course factors is limited. However, they are often the only studies with the statistical power to identify subnational differentials and time trends so they are complementary to rarely available sources such as high-quality long-term individual-level microdata data required for causal analyses. Income equality can arise not only due to citizens’ positive preferences but also to external choices such as economic decline and globalisation, so examining the wider context is important when explaining excess levels of ‘deaths of despair’ in low-inequality US states. The apparent increasingly strong association between income levels and low mortality with a weakening inequality/mortality relationship has implications for policy recommendations.