Risk of common psychiatric disorders, suicidal behaviours, and premature mortality following violent victimisation: A matched cohort and sibling-comparison study of 127,628 people who experienced violence in Finland and Sweden.

Abstract

Background: Associations between violent victimisation and psychiatric disorders are hypothesised to be bidirectional, but the role of violent victimisation in the aetiologies of psychiatric disorders and other adverse outcomes remains unclear. We aimed to estimate associations between violent victimisation and subsequent common psychiatric disorders, suicidal behaviours, and premature mortality while accounting for unmeasured familial confounders.

Methods and findings: Using nationwide registers, we identified a total of 127,628 individuals born in Finland (1987 to 2004) and Sweden (1973 to 2004) who had experienced violent victimisation, defined as either hospital admissions or secondary care outpatient visits for assault-related injuries. These were age- and sex-matched with up to 10 individuals in the general population (n = 1,276,215). Additionally, we matched those who had experienced violent victimisation with their unaffected siblings (n = 132,408). Outcomes included depression, anxiety, personality disorders, alcohol use disorders, drug use disorders, suicidal behaviours, and premature mortality. Participants were followed from the victimisation date until the date of the outcome, emigration, death, or December 31, 2020, whichever occurred first. Country-specific associations were estimated using stratified Cox regression models, which also accounted for unmeasured familial confounders via sibling comparisons. The country-specific associations were then pooled using meta-analytic models. Among 127,628 patients (69.0% male) who had experienced violent victimisation, the median age at first violent victimisation was 21 (interquartile range: 18 to 26) years. Incidence of all outcomes was larger in those who were exposed to violent victimisation compared to population controls, ranging from 2.3 (95% confidence interval (CI) [2.2; 2.4]) per 1,000 person-years for premature mortality (compared with 0.6, 95% CI [0.6; 0.6], in controls) to 22.5 (95% CI [22.3; 22.8]) per 1,000 person-years for anxiety (compared with 7.3, 95% CI [7.3; 7.4], in controls). In adjusted models, people who had experienced violent victimisation were between 2 to 3 times as likely as their siblings to develop any of the outcomes, ranging from adjusted hazard ratio [aHR] 1.7 (95% CI [1.7; 1.8]) for depression to 3.0 (95% CI [2.9; 3.1]) for drug use disorders. Risks remained elevated 2 years post-victimisation, ranging from aHR 1.4 (95% CI [1.3; 1.5]) for depression to 2.3 (95% CI [2.2; 2.4]) for drug use disorders. Our reliance on secondary care data likely excluded individuals with milder assault-related injuries and less severe psychiatric symptoms, thus suggesting that our estimates may be conservative. Another limitation is the possibility of residual genetic confounding, as full siblings share on average about half of their co-segregating genes. However, the associations remained robust even after adjusting for both measured and unmeasured familial confounders.

Conclusions: In this longitudinal cross-national cohort study, we observed that those who had experienced violent victimisation were at least twice as likely as their unaffected siblings to develop common psychiatric disorders (i.e., depression, anxiety, personality disorder, and alcohol and drug use disorders), engage in suicidal behaviours, and to die prematurely. Importantly, these risk elevations remained 2 years after the first victimisation event. Improving clinical assessment, management, and aftercare psychosocial support could therefore potentially reduce rates of common psychiatric disorders, suicidality, and premature mortality in individuals experiencing violent victimisation.