From Norman Bates in Psycho (1960), to Jack Torrance in The Shining (1980), to characters in more recent productions such as Split (2016) and Joker (2019), a well-trodden motif in film and literature is that of the mentally-ill antagonist whose psychological state makes them dangerous, unpredictable and violent. This is not merely a narrative trope that sells film tickets: it has the potential to feed into public attitudes and stereotypes about mental health in ways that can be misleading and stigmatising.
Viewing mental illness as a precursor to violent behaviour neglects the fact that the individuals affected may themselves be at increased risk of being victimised. A previous blog post for The Mental Elf (Bhavsar, 2015) has summarised research showing that people with severe mental illness were more likely than people in the general population to be victims of violence (Khalifeh et al, 2015).
This is not to deny that research has found higher rates of violent behaviour among people with severe mental illness, often linked to substance use (Schwartz, 2021; Steeg, 2020). But studies that look at both sides find that being a victim of violence is also far more common (Paulino and Ogonah, 2024).
A recent systematic review by Rocha and colleagues (2025), published in the journal Trauma, Violence & Abuse, sought to provide further clarity on this matter by synthesising the current breadth of knowledge in the area. Their aim was to present a comprehensive picture of how common victimisation is among adults suffering from severe mental illness.
Methods
Rocha and colleagues searched five databases of scientific publications, in order to locate studies that reported the prevalence of victimisation in the adult population. For the purpose of the review, they defined ‘victimisation’ as any form of violence exposure, and their working definition of ‘severe mental illness’ comprised diagnoses such as major depressive disorder, bipolar disorder, schizophrenia, post-traumatic stress disorder and substance use disorders.
Starting with more than 1,500 initial search hits, the authors shortlisted thirty articles that met their inclusion criteria. In addition to reviewing and summarising the findings of the research, the authors extracted the reported statistics where possible, and used random effects meta-analysis to calculate ‘pooled’ effects based on data from across multiple studies. They also evaluated the methodological quality of the studies using a standardised appraisal tool, the Joanna Briggs Institute’s Critical Appraisal Checklist for Studies Reporting Prevalence Data (Munn et al, 2015).
Results
The thirty publications identified were highly heterogeneous in terms of the study design and sample makeup. The majority (twenty-two) were cross-sectional in nature, and in the eight longitudinal studies, the follow-up period varied in duration, from a few weeks to several years. Sample sizes were also highly variable; ranging from 150 to 4,168. Other sample characteristics were more consistent: the average age tended to fall within the 35-45 range, and the gender distribution was roughly balanced in most studies.
Using the appraisal tool developed by the Joanna Briggs Institute, the authors assessed the methodological quality and risk of bias in the included studies, based on criteria such as sampling and measurement. They found a mixed picture: most of the studies fell short on at least one criterion, and 70% were deemed to have samples that were insufficiently representative of the general population.
A headline finding from the meta-analysis was that approximately one-third of people with severe mental illness experienced violence victimisation within a 12-month time period. It should be noted that this is a ‘pooled’ result across all of the studies with usable data, and the prevalence of victimisation in individual studies varied widely (from 13% to 94%). Among the handful of studies that asked about longer periods, the pooled estimate of victimisation prevalence was 27% for studies covering 13-60 months, and 41% for studies covering more than 60 months (including lifetime).
When examining correlates of victimisation, a number of factors were identified, including: young age, female gender, homelessness, substance abuse, severity of mental health symptoms, and previous history of victimisation. While these associations do not establish causal pathways (the interrelationship between such risk factors is complex), they demonstrate a constellation of co-occurring vulnerabilities among people suffering from severe mental illness.
Conclusions
The authors concluded that people with severe mental illness are at significantly elevated risk of being victims of violence; something which they argue is relatively under-researched, particularly in low-income countries. They suggest that these findings can inform mental health practice, public policy, and efforts to increase awareness in the wider population. However, there is still room for greater methodological consistency, and research in more diverse populations, to improve the evidence base on this topic.
Strengths and limitations
Systematic reviews and meta-analyses can generate some of the strongest forms of scientific evidence. Because they synthesise a large body of work to produce a big-picture overview of the current state of the literature, they are valuable in guiding clinicians and policymakers in their decision-making. They can also highlight gaps and limitations in the existing evidence that can inform new research directions.
Rocha and colleagues followed the best-practice procedures that typify a systematic review; pre-registering their study protocol and reporting it in line with the PRISMA guidance (Page et al, 2021). The selection of studies was based on agreement between two of the authors, with the input of a third author if agreement could not be reached. The statistical methods, using random-effects models to account for heterogeneity across studies, were appropriate.
When searching the literature for a systematic review, researchers must make careful decisions about the search query they use. If too specific, important publications may be missed; if too broad, the number of results becomes unmanageable and bulked out with irrelevant papers. Here, Rocha and colleagues struck a reasonable balance in their search string. However, the victimisation-related search terms they entered were limited to “victimisation” and “victim”. They might have considered including related terms such as “abuse” and “maltreatment”, which could have picked up additional search results.
Nonetheless, the authors aimed for a good amount of breadth in their review, searching both medical and psychological databases as well as more general scientific ones. The literature reviewed was international in scope: it included thirteen studies from European countries, seven from the United States, four from Australia, three from Asia, and three from Africa. Surprisingly absent were any studies from South and Central America, even though the databases searched included SciELO (which is dedicated to hosting research from that region of the world).
An additional point is that the population under investigation (those with ‘severe mental illness’) comprised a range of different conditions, such as depression, psychosis and substance abuse. These represent different clinical sub-groups who may have quite diverse experiences of victimisation. Integrative reviews such as this are valuable for providing the ‘bigger picture’, but it is also important for studies to look at differences across diagnoses (and across various types of victimisation).
Implications for practice
Risk assessments of individuals with severe mental illness carefully consider the risk of harm posed by the affected individual; either to themselves or to others. This is an essential part of clinical due diligence. Rocha and colleagues note that clinicians generally assess the risk patients pose to others but rarely ask about their experiences of being victims of violence, and they emphasise the importance of preventative measures to ensure that violence victimisation does not exacerbate negative outcomes in people with severe mental illness.
The review highlights that discussions around safety in mental health practice, and the policy decisions that are made on the basis of these discussions, should give due consideration to violence that people with severe mental illness are exposed to. This is an important part of the bigger picture, with regard to ensuring individuals have access to the appropriate services and safety nets.
For researchers, a number of evidence gaps are revealed. Certain geographical regions, in particular Latin America, appear to be under-represented in this research space. In addition, the authors recommend further research into factors that can increase or mitigate the risk of victimisation among people with severe mental illness. As less than one-third of the studies reviewed were longitudinal in design, this is a clear area in which the literature could be strengthened.
Notably, the pooled prevalence estimates did not show a consistent pattern by the length of time participants were asked about; studies covering the past 13-60 months showed a lower pooled prevalence than studies covering shorter or longer periods. As each estimate comes from a different, small set of studies, it is difficult to infer whether there is an increase or decrease in victimisation rates over time, and future longitudinal research should seek to explore this.
Lastly, the paper serves as a reminder that social attitudes towards mental health can be one-sided; over-emphasising the idea of people with mental illness as a ‘danger’. By conducting a thorough deep-dive of a body of literature spanning multiple disciplines and multiple countries, Rocha and colleagues present the other side of the story: that individuals with severe mental illness are significantly more likely than average to be victimised themselves. Hence, already at-risk people can find their health and recovery further impacted by the harmful experiences they encounter.
The Joe Wright film The Soloist (2009) tells the true story of Nathaniel Ayers, a talented cellist who was forced to abandon his studies at a prestigious music school after his mental health deteriorated. Following unsuccessful treatment in a psychiatric hospital, Ayers faced homelessness, social isolation, and exploitation. The Soloist did not receive quite the same accolades as the films mentioned above, but it is a work that invites us to reflect on the sorts of vulnerabilities highlighted by Rocha et al’s review.
Statement of interests
Timothy Matthews has no conflicts of interest to declare.
Editor
Edited by Laura Hemming.
Links
Primary paper
Isabella Carvalho Oliveira Rocha, Alexandre Martins Valença, Evandro Silva Freire Coutinho, Pedro Barcelos Banhara, William Berger (2025) Prevalence of victimisation among adults with severe mental illness worldwide: A systematic review and meta-analysis. Trauma, Violence, & Abuse, 12:15248380251386799.
Other references
Bhavsar V (2015) People with severe mental illness are more likely to be victims of violent and non-violent crime. The Mental Elf.
Khalifeh H, Johnson SC, Howard LM et al (2015) Violent and non-violent crime against adults with severe mental illness. British Journal of Psychiatry, 206(4), 275-282.
Munn Z, Moola S, Lisy K et al (2015). Methodological guidance for systematic reviews of observational epidemiological studies reporting prevalence and incidence data. International Journal of Evidence-Based Healthcare, 13(3), 147–153.
Page MJ, McKenzie JE, Bossuyt PM et al (2021). The PRISMA 2020 statement: An updated guideline for reporting systematic reviews. International Journal of Surgery, 88, 105906.
Paulino A, Ogonah M (2024) Mental illness is linked to being a victim and/or perpetrator of violence: time to face up to some inconvenient truths? The Mental Elf.
Steeg S. (2020) In harm’s way: psychiatric diagnosis and risks of being subjected to and perpetrating violence The Mental Elf.
Schwartz J. (2021) Violence and mental illness: does ignoring this blog stigmatise some people who need our help? The Mental Elf.