Association of physical and sexual assault with mortality in two British birth cohorts

Abstract

The association of assault in adulthood with all-cause mortality, and the relevance of intermediate psychological distress, alcohol use and cigarette smoking, is poorly understood. We used data from British birth cohorts (the 1958 National Child Development Study referred to as the 1958 birth cohort and the 1970 British Birth Cohort Study) to investigate association between assault and mortality, employing a formal approach for the identification of psychological distress, alcohol use and cigarette smoking as mediators.Associations (HRs), between assault and mortality were estimated with Cox regressions, adjusting for potential confounders. Mediation via intermediate psychological distress, alcohol use and cigarette smoking was explored using the gformula approach. The birth cohorts were analysed separately, and together estimating interaction between exposure and cohort year.Results were based on 353 deaths in 19 725 individuals. Based on multiply imputed data, the fully adjusted estimate for assault on mortality was 1.72 (95% CI 1.22 to 2.42) in the combined cohorts, 1.53 (95% CI 0.97 to 2.40) in the 1958 birth cohort and 2.05 (95% CI 1.20 to 1.50) in the 1970 birth cohort. The fully adjusted estimate for the association of sexual assault with mortality was 3.17 (95% CI 1.17 to 8.60) in the combined cohorts, 1.36 (95% CI 0.19 to 9.81) in the 1958 birth cohort and 6.02 (95% CI 1.84 to 19.69) in the 1970 birth cohort. The fully adjusted mortality HR for one additional assault was 1.46 (95% CI 1.23 to 1.73) in the combined cohorts, 1.34 (95% CI 0.99 to 1.82) in the 1958 birth cohort and 1.53 (95% CI 1.25 to 1.87) in the 1970 birth cohort. Greater need for medical treatment for assault was associated with a fully adjusted mortality HR of 1.56 (95% CI 1.19 to 2.05) in the combined cohorts, 1.43 (95% CI 1.00 to 2.05) in the 1958 birth cohort and 1.79 (95% CI 1.18 to 2.74) in the 1970 birth cohort.There was statistical evidence on combining the two birth cohorts, and on analysing the 1970 birth cohort, that assault in adulthood is associated with mortality. Understanding mechanisms underlying this relationship could benefit violence reduction strategies for public health.