Citation

Wallinius M, Nilsson T. Lancet Public Health 2023; 8(6): e398-e399.

Copyright

(Copyright © 2023, Elsevier Publishing)

DOI

10.1016/S2468-2667(23)00099-3

PMID

37244669

Abstract

Violence is a global public health priority affecting the lives of people worldwide, either as victims, perpetrators, or members of society. It can lead to traumatic experiences, and studies of military personnel have linked post-traumatic stress disorder (PTSD) to subsequent violent behaviours, indicating a potential bidirectional association between trauma exposure and violence. These effects are especially pertinent in relation to war. Furthermore, traumatic experiences during childhood have been linked to subsequent violence and substantial health challenges for individuals. In familial contexts, these subsequent effects have been referred to as a cycle of violence. However, there are few studies investigating mental disorders related to the substantial and long-term effects of trauma, such as PTSD, in relation to violence, and findings from the existing studies have methodological limitations. In this issue of The Lancet Public Health, Anabelle Paulino and colleagues use a Swedish general population sample with matched controls and sibling comparisons from linked registry data to show that, regardless of previous psychiatric comorbidity and familial factors, PTSD is associated with an increased risk of violent crime conviction (hazard ratio [HR] 6·4, 95% CI 5·7-7·2). However, they also show that substance use disorder (SUD) and a history of violent crime, both prominent risk factors for violent offending, confound this association, as was shown when analyses were stratified by history of violent crime (2·2, 1·7-3·0) and SUD (1·9, 1·5-2·6). Although these results imply a risk increase at the population level for violent crime in people diagnosed with PTSD who do not have previous SUD or a history of violent crime, possible moderation or mediation effects remain unclear.


Observational studies of registry data can be informative and include broader perspectives on public health challenges, especially when conducted in settings with high coverage and validity of national registers. Such studies are statistically well powered, have general representativeness for the investigated setting, and use objective measures (eg, official records) instead of subjective measures (eg, self-report). However, the inherent limitations of this design include requirements for a disorder or behaviour to be registered in official records. The main outcome in this study, violent crime conviction, has been described as largely under-reported when used as an estimation of violent crime.7
Reliance on official records might also have affected the prevalence of PTSD diagnoses. Qualitative research in the UK found a reluctance to seek health care for mental distress in male offenders.8
Thus, the magnitude of an association between PTSD and violent crime might not be adequately captured in studies using registry data only. These issues are acknowledged by the authors but should be considered when translating their results to health-care interventions.

Language: en