Suicide rates and voting choice in the UK's 2016 national Brexit referendum on European Union membership: cross-sectional ecological investigation across England's local authority populations

Population health, not individual health, drives support for populist parties

Recent electoral shifts toward populist parties may have been partly driven by deteriorating health, although empirical evidence on this link is primarily confined to ecological designs. We performed both ecological- and individual-level analyses to investigate whether changes in health are associated with changes in the support for populist parties. Data were used on the strategic Dutch case, the only liberal democracy featuring leftist and rightist populist politicians in parliament for over a decade. We used: (a) fixed effects models to examine whether changes in the standardized mortality ratios and self-assessed health (SAH) in municipalities were associated with changes in the populist vote share in four parliamentary elections (2006/2010/2012/2017); and (b) 10 waves of panel data collected in 2008 to 2018 to investigate if changes in individual-level SAH were linked to movement in the sympathy, intention to vote, and actual voting for populist parties. The ecological analyses showed that: changes in municipality mortality ratios were positively linked to changes in the vote share of right-wing populist parties, while changes in the prevalence of less-than-good SAH were negatively associated with changes in the vote share for left-wing populist parties. The individual-level analyses identified no such associations. Our findings imply that support for populist parties may be driven by health concerns at the ecological, but not the individual, level. This suggests that sociotropic (e.g. perceiving population health issues as a social problem), but not egotropic (e.g. relating to personal health issues like experienced stigma), concerns may underlie rising support for populist parties.

Brexit and the geography of depression: A reply to Liew et al. (2020)

Liew et al. (2020) recently published a paper in this journal that analyzed antidepressant prescription trends in the context of the 2016 Brexit referendum and the sociopolitical discord that followed. They present a novel finding that Leave-majority constituencies in England seemed more adversely affected by that discord than Remain-majority constituencies. I offer criticism of their findings and methodology. Using the complete set of available NHS prescription data shows that the trend the authors detect dates from at least mid-2010 and is not associated with the referendum. In terms of methodology, I critique the potential ecological fallacy and issues of false equivalence in their study design. The former stems from the inability to adequately control for demographic heterogeneity within constituencies, and the latter stems from the fact that the populations from which they draw their data are not equivalent in potentially important ways. Finally, I conclude that the key trend the authors detect seems to merely be a geographic artifact. The set of Remain-majority constituencies unintentionally oversamples the areas of England with the lowest rates of antidepressant prevalence, Greater London and the Southeast. Remain-majority constituencies outside of those two regions have roughly the same antidepressant prescription levels as Leave-majority constituencies in all of England. In itself, that is a troubling fact of social epidemiology, but Brexit is associated with it neither spatially nor temporally.