Psychiatric disorders are relatively common, with 1 in 4 people worldwide displaying symptoms of a diagnosable psychiatric disorder (Mental Health Task Force, 2016). Previous research has provided adequate preliminary evidence of the association between psychiatric disorder and violence. For example, one previous study (Witt et al, 2013) revealed that violence, and homicide in particular, is approximately four times greater among males and eight times greater among females who have been diagnosed with schizophrenia or other psychoses. Importantly, however, the majority of these associations were accounted for by substance use comorbidity. In another study (Sariaslan et al, 2019) examining Swedish population registers and employing a rigorous, sibling-comparison research design revealed a link between psychiatric disorders and both perpetrating and also being the victim of violence.
Collectively, these findings provide preliminary evidence to expect a link between psychiatric disorders and violent behaviours. Despite this evidence, however, the role of psychiatric disorders in the development of violence has been continuously called into question, both in the academic literature and among policy makers. For example, summaries of the literature often cite the extremely low number of violent offenders that suffer from psychiatric disorders (see for example Varshney et al., 2016). Further, governmental agencies often cast doubt on this association, with the Department of Health and Human Services in the United States even listing this connection (albeit as a more sensationalised version) as a “Myth” on their “Mental Health Myths and Facts” page (U.S. Department of Health and Human Services, 2017), instead citing the increased rate of victimization among those with mental illness.
These inconsistencies are problematic as they may prevent the delivery of potential treatment and intervention programming aimed at improving quality of life and, at least potentially, preventing violence. In an effort to examine the potential association between psychiatric disorders and violence, Whiting et al. (2020) performed a structured, meta-analytic review of the literature examining this association. The authors clearly state the goal of their study:
In this Review, we summarise evidence examining associations between individual psychiatric diagnoses and violent outcomes, with an emphasis on high quality methods, which could lead to improvements in the assessment and management of violence and related risks.
If mental health anti-stigma messages say that people with severe mental illness are not more likely, on average, to commit violent acts, are they unintentionally stigmatising some people who need help?
Methods
The authors performed a systematic review of the literature with a specific emphasis on studies that directly examined the independent influence of specific psychiatric diagnoses on violent criminal convictions. The authors seem to have taken the Critical Appraisal Skills Programme (CASP) checklist into account in their review, as they have directly focused on high-quality, rigorous studies.
The examined psychiatric diagnoses were based on the International Classification of Diseases (ICD) and Diagnostic and Statistical Manual of Mental Disorders (DSM) criteria and included:
- Schizophrenia and related disorders
- Bipolar disorder
- Depression
- Attention deficit hyperactivity disorder (ADHD)
- Autism spectrum disorders (ASD)
- Personality disorder
- Post-traumatic stress disorder (PTSD)
- Substance misuse
- Alcohol use disorder
For the outcome, the authors relied on official, register-based records of violent criminal convictions.
One of the primary strengths of this study is that for each of the examined psychiatric diagnoses selected, the authors have deliberately selected studies that fall within three categories:
- Systematic review
- Sibling-control study (compares affected and unaffected siblings from the same household)
- Population-based epidemiological study (compares affected and unaffected siblings to controls in the general population)
Results
For each of the primary diagnoses examined, the authors reported the relative risks for the three types of studies examined (when available). The results revealed consistent associations between virtually all of the examined disorders and violence, with the only exception being autism spectrum disorders.
Relative risk of violent outcomes
Here are the results for the disorders, broken down by study type:
| Mental health condition | Systematic review | Sibling control study | Population-based epidemiological study |
| Schizophrenia and related disorders | OR = 4.0 (95% CI = 3.0 to 5.3) | OR = 7.4 (95% CI = 7.1 to 7.8) | Men: OR = 3.0 (95% CI = 2.7 to 3.4);
Women: OR = 7.5 (95% CI = 6.0 to 9.3) |
| Bipolar disorder | OR = 4.6 (95% CI = 3.9 to 5.4) | OR = 2.8 (95% CI = 2.5 to 3.1) | OR = 3.7 (95% CI = 2.2 to 6.0) |
| Depression | OR = 3.0 (95% CI = 2.8 to 3.3) | OR = 2.9 (95% CI = 2.1 to 4.0) | |
| ADHD | OR = 3.6 (95% CI = 2.3 to 5.7) | OR = 2.7 (95% CI = 2.0 to 3.8) | OR = 1.8 (95% CI = 1.1 to 2.7) |
| Personality disorder | OR = 3.0 (95% CI = 2.6 to 3.5) | OR = 2.7 (95% CI = 2.2 to 3.2) | |
| PTSD | OR = 3.2 (95% CI = 2.8 to 3.8) | OR = 2.2 (1.4 to 3.6) | |
| Substance misuse | OR = 7.4 (95% CI = 4.3 to 12.7) | OR = 16.2 (95% CI = 14.6 to 17.9) | |
| Alcohol use disorder | OR = 9.0 (95% CI = 8.2 to 9.9) |
The authors also found significant associations between some additional disorders and violence. It is important to note however, that these results should be viewed with more caution, as the methodologies that resulted in these findings are not nearly as robust as those presented for the primary disorders. With that in mind, here are the results for the other diagnostic categories:
| Diagnostic categories | Longitudinal register-based | Retrospective cohort study |
| Anxiety disorders | HR = 2.0 (1.7 to 2.3) | |
| Intellectual disability | study 1: IRR = 1.7 (1.3 to 2.3)
study 2: OR = 3.6 (2.7 to 4.8) |
|
| Gambling disorder | OR = 2.2 (1.4 to 4.5) | |
| Conduct disorder | Females: HR = 1.4 (95% CI = 1.2 to 1.5);
Males : HR = 1.2 (95% CI = 1.2 to 1.3) |
Conclusions
The authors of the study summed up their findings as:
In this Review, we have shown some individual psychiatric disorders, particularly schizophrenia spectrum, personality, and substance use disorders, are clearly associated with increased relative risks of violence. Better research designs and large-scale replications have led to more conservative estimates of this association.
A substantial body of reliable evidence suggests that violence appears to be an adverse outcome of many psychiatric disorders.
Strengths and limitations
This review is an excellent contribution to the existing literature and provides some of the strongest and most comprehensive evidence of the link between psychiatric disorders and violence to date. Some of the key strengths of the study are:
- The authors’ approach of identifying the most rigorous methodologies available and summarising the results of the examined disorders across these types of studies provides some of the most convincing evidence of a causal connection between psychiatric diagnoses and violence.
- The authors not only examine the collective influence of mental disorders and the accompanying effect sizes, but they also examine the unique influence of specific diagnoses and violent criminal convictions.
- Examining results across an assortment of rigorous methodologies not only provides a general effect size for the examined associations, but also demonstrates the robustness and replicability of the reported findings.
With that said, however, this study, just like any other, is not without limitation. Below, I have identified some of the limitations I found most concerning:
- Aside from the evidence from the systematic reviews, the majority of the cohorts, registries, and samples from which the results are reported appear to be from a limited number of countries, primarily represented by Sweden, Denmark, and The Netherlands. Again, there are exceptions noted throughout the text, but the bulk of the findings seem to stem from this limited number of countries. With that in mind, external validity may be limited, as this cross section of countries possess important demographic, cultural, social, and economic differences when compared to other countries.
- The authors focus exclusively on violent criminal convictions as their outcome. There are many benefits to this approach but relying on official records also results in important limitations. Most notably, this approach only captures violent acts that are: 1) reported to or detected by official agents of the criminal justice system; 2) result in an official arrest; 3) are prosecuted; and 4) result in a criminal conviction. On the one hand this measure is precise, but on the other hand it is overly stringent, omitting a significant number of violent offences that do not meet these criteria.
- The study does not consider comorbidity of disorders. I understand that the primary objective was to examine the independent effect of each of the included disorders, but previous studies have revealed significant comorbid effects (particularly with substance abuse) in the development of violent behaviours.
Ignoring this evidence does not help people who have severe mental health problems and also commit acts of violence.
Implications for practice
The findings of this review provide the most stringent evidence to date indicating that violence is an adverse outcome of a relatively wide range of psychiatric disorders. From a public health perspective this finding is important, as it suggests that proper treatment of psychiatric disorders could potentially have the dual benefit of improving mental health outcomes and also preventing future violence. Not only would this improve overall quality of life and provide safer communities, but it would also result in a significant reduction in expenditures, as both mental health services and crime control practices are extremely costly.
Second, and as the authors discuss, these findings have implications for risk assessment models, which are typically used to guide treatment plans and gauge violence risk in institutional settings. While an improvement in these prediction models would be useful in preventing future violence, this approach overlooks the fact that a significant number of violent offences are committed before such tools are implemented for the first time. Previous studies have revealed that a sizeable number of individuals already meet the criteria for a mental disorder during childhood and adolescence (Caspi et al., 2020). These findings indicate that routine screening procedures during these same stages of the life course (similar to physical health wellness checks) may result in pronounced preventative benefits. This practice, coupled with the improved predictive models outlined by the authors, would likely result in a significant improvement over the current system, which allocates extremely limited resources for prevention and provides an almost exclusive focus on more reactive forms of treatment.
A greater emphasis on prevention efforts could reduce the risk of violence.
Statement of interests
None.
Links
Primary paper
Whiting D, Lichtenstein P, & Fazel S (2020) Violence and mental disorders: a structured review of associations by individual diagnoses, risk factors, and risk assessment. The Lancet Psychiatry 2020.
Other references
Care and Quality Commission (2015) Right Here, Right now: People’s Experiences of Help, Care and Support During A Mental Health Crisis. London: Care Quality Commission.
Caspi A, Houts RM, Ambler A, et al (2020). Longitudinal Assessment of Mental Health Disorders and Comorbidities Across 4 Decades Among Participants in the Dunedin Birth Cohort Study. JAMA Network Open 2020 3(4) e203221-e203221 doi:10.1001/jamanetworkopen.2020.3221.
Center for Behavioral Health Statistics and Quality (2020). 2019 National Survey on Drug Use and Health: Methodological summary and definitions. Rockville, MD: Substance Abuse and Mental Health Services Administration.
Judge J (2014). Long duration of untreated psychosis is associated with a range of poor outcomes. The Mental Elf. 2014 October 22.
Mental Health Taskforce (2016) The Five Year Forward View for Mental Health.
Northfield J (2012) Mental health disorders in young people with learning disabilities still under-recognised and often untreated in clinical practice. The Mental Elf. 2012 September 26.
Sariaslan A, Arseneault L, Larsson H, Lichtenstein P, Fazel S. Risk of Subjection to Violence and Perpetration of Violence in Persons With Psychiatric Disorders in Sweden. JAMA Psychiatry. Published online January 15, 2020. doi:10.1001/jamapsychiatry.2019.4275
Steeg S (2020) In harm’s way: Psychiatric diagnosis and risks of being subjected to and perpetrating violence. The Mental Elf. 2020 February 20.
Tomlin A (2013) New review confirms the strong association between criminal history and violence risk in psychosis. The Mental Elf. 2013 February 15.
U.S. Department of Health & Human Services (2017 August 29) Mental health myths and facts. MentalHealth.gov.
Underwood R (2014) Meta-review presents the risks of all-cause and suicide mortality in mental disorders. The Mental Elf. 2014 August 12.
Varshney M, Mahapatra A, Krishnan V et al (2016). Violence and mental illness: what is the true story? J Epidemiol Community Health 2016 70(3) 223-225 .
Walton M (2013) People with disability are more likely to be victims of violence and to suffer mental illness as a result. The Mental Elf. 2013 March 12.
Witt K, van Dorn R, Fazel S (2013) Risk Factors for Violence in Psychosis: Systematic Review and Meta-Regression Analysis of 110 Studies. PLoS ONE 8(2): e55942. doi:10.1371/journal.pone.0055942
Photo credits
- Photo by Salomé Guruli on Unsplash
- Photo by Evan Dennis on Unsplash
- Photo by Nick Fewings on Unsplash
- Photo by Possessed Photography on Unsplash
- Photo by Franck – Free-Hotspot.com on Unsplash
- Photo by Kate Kalvach on Unsplash
- Photo by Andrew Gook on Unsplash
- Photo by chester wade on Unsplash
Really good summary.
Important point about the need to consider the full range of the evidence in order to prevent and treat violence risk. Prevention will reduce stigma, not ignoring the evidence.
One point about the generalisability of the review. We brought together the systematic reviews (based on general population controls) and the sibling control studies to examine the consistency, strength and quality of the evidence. The included systematic reviews draw on research from many countries (although almost entirely high income, which is a limitation) – including the US, Australia, New Zealand, and the UK. The sibling control studies, as the blog points out, are mostly based in a few countries where these investigations are possible, so mostly Nordic countries for now.
And the blog makes another good point about the outcome measures used. Many of these studies use violent conviction – which has advantages (including that it captures the highest impact on morbidity and mortality, also service use and costs, reliably captured). At the same time, many primary studies in the systematic reviews use self-report and informant-report measures. One notable thing about these is that relative risks don’t seem to change compared to when you use non-criminal outcomes. In other words, the degree of underreporting of violence using criminal outcomes is the same in the cases and controls.