Media release

From: BMJ Group

Voting behaviour in elections strongly linked to future risk of death

Likely stronger determinant of health than education, suggest researchers

Voting behaviour in elections is strongly linked to the future risk of death, and is likely a stronger determinant of health than education—considered a key influence on health—suggests research published online in the Journal of Epidemiology & Community Health.

Over the past several years, voting in national and local elections has increasingly been seen as a potential social determinant of health—the non-medical factors that influence health and wellbeing—explain the researchers.

And previously published research suggests that voters generally have better health than non-voters, but it’s not clear if electoral participation might also be a predictor of future death risk.

To explore this further, the researchers drew on electoral participation information in the 1999 parliamentary elections for the entire electorate of Finnish citizens living in mainland Finland and aged at least 30.

Turnout among this age group was 71.5% for men and 72.5% for women. In all, 3,185, 572 people (1,508,824 men; 1,676,748 women) were included in the study and their survival was tracked from 21 March 1999 (election day) to death or the end of 2020, whichever came first.

Between 1999 and 2020, 1,053,483 people died: 95,350 from external causes (accidents, violent and alcohol-attributable causes); 955,723 from other underlying causes; 2410 people whose cause of death wasn’t known were excluded from the final analysis.

Not voting was consistently associated with a 73% heightened risk of death from any cause among men and a 63% heightened risk of death among women.

After adjusting for education level (basic or unknown; secondary; or tertiary) this reduced to a 64% heightened risk among men and a 59% heightened risk among women.

The difference in the risk of death between voters and non-voters was larger than it was between those with basic and tertiary level education.

The association between voting participation and death was stronger for external causes of death rather than for other causes. And when adjusted for age, this risk was twice as high among both men and women who didn’t vote than it was among those who did.

The difference in the relative risk of death between voters and non-voters was strongest among men under 50. Among 75 to 94-year olds, women who didn’t vote had a higher risk of death than men who voted, whereas the reverse was true for younger age groups.

And among men in the lowest 25% of household income, the risk of death associated with not voting was 9–12% higher than it was for other income groups.

This is an observational study, and as such, can’t establish cause and effect. And the researchers acknowledge various limitations to their findings, including that some people might have wanted to vote, but were unable to do so, or simply chose not to vote.

Nevertheless, the fact that voting was more strongly associated with the risk of death than educational level “suggests a strong association between voting and mortality, given that a substantial education gradient in mortality is among the most robustly established sociodemographic mortality patterns,” they suggest.

“Voting, as a form of participation, is a type of social capital, which is linked to health benefits. In addition, voting may increase other forms of civic participation,” they explain.

“Health problems and related difficulties in functioning may also negatively affect many important prerequisites of participation, including enhancing resources, motivation to vote, and political mobilisation,” they add.

They conclude: “Information on voting may be useful in clinical settings – for example, abandonment of voting habits may be an early signal of significant health decline – and in monitoring population wellbeing, health, and health disparities. Additionally, a strong association between voting and mortality raises concerns about equal political representation.”