Schizophrenia and Violence: What Research Really Shows
Author: Ukrainian Psychological Hub · Published: October 9, 2026 · Editorial Policy
Schizophrenia and violence are often discussed together in headlines, television dramas, and everyday assumptions. Population research gives a much more specific answer. Some well-designed studies find that people diagnosed with schizophrenia spectrum disorders have a higher relative rate of certain violent outcomes than matched community comparison groups. Yet the large majority of people with schizophrenia do not perpetrate violence, and many face substantial risks of being threatened, assaulted, exploited, or otherwise victimized themselves. The distinction between a statistical association and an individual's behavior is fundamental.
For readers concerned about someone they know, the practical question is usually more urgent than the headline: What actually raises risk, which signs matter, what does treatment change, and how can everyone remain safe without reducing a person to a diagnosis? The available evidence points toward a combination of prior behavior, substance use, current clinical state, environmental exposure, and opportunities for timely care. Diagnosis is one piece of clinical context; it is not a prediction about a particular person's actions. The National Institute of Mental Health explicitly states that most people with schizophrenia are not violent and are more likely than people without the illness to be harmed by others.
I. The short answer: what research establishes
The strongest overview comes from a 2022 systematic review and meta-analysis in JAMA Psychiatry, which combined 24 studies from 15 countries and included 51,309 people with schizophrenia spectrum disorders. Compared with community controls, the pooled odds of a measured violent outcome were elevated in male and female groups, although the estimated size of the association varied greatly across studies. The authors also reported long-term absolute outcomes that were much less frequent than dramatic relative-risk ratios might suggest. These observations belong together: elevated group-level odds can coexist with most people in the diagnosed group never experiencing the outcome.
The newer 2025 longitudinal systematic review by Lagerberg and colleagues examined risk factors within people with psychotic disorders. Its evidence base included 47 longitudinal studies representing 41 independent samples and 203,297 people. A history of violence and co-occurring substance misuse showed consistent associations with later violent outcomes. Certain treatment-related factors were associated with lower risk. Other proposed explanations, including simply counting psychotic symptoms, were less reliable in pooled analyses. This is why a serious discussion of prevention examines changeable circumstances, specific behaviors, and effective care rather than equating psychosis with dangerousness.
The World Health Organization's September 2026 schizophrenia fact sheet describes substantial stigma, discrimination, and human-rights violations experienced by people with schizophrenia. Public descriptions that treat a whole population as threatening can worsen those problems and discourage treatment engagement. Accurate discussion must protect two interests at once: prevention of actual harm and the dignity, safety, and rights of people living with the diagnosis.
II. What counts as violence in schizophrenia research?
Research studies measure different outcomes under the same general word. An official record of a violent conviction is not equivalent to a clinician's note about verbal hostility, a report of pushing someone during an argument, a threat, a physical assault requiring medical care, sexual violence, or homicide. Some investigations ask participants or relatives about behavior; others use police, hospital, or court registers. Observation windows may last months, years, or decades. These choices have a direct effect on the rate reported.
The 2022 JAMA Psychiatry review included various forms of interpersonal violence and violent offending. That breadth makes it useful for detecting associations but means its combined odds ratios are not the probability of any one specific act, such as assault or homicide. The rates of rare severe outcomes must never be presented as though they describe everyday behavior. A risk statistic tied to a conviction also reflects reporting, policing, access to support, and legal-system practices; it does not capture every occurrence of aggression or every act of victimization.
1. Aggression, hostility, and criminal violence are different outcomes
Aggression can refer to hostile words, threatening gestures, or physical acts. Hostility is an emotional or behavioral tendency, and agitation can be a state of intense distress or motor restlessness. Violence usually denotes physical force or a serious threat of harm, while violent crime additionally requires a legal classification and detection. A person may become frightened, shout, or withdraw during an episode of psychosis without intending or committing physical harm. Careful language matters because confusing these outcomes inflates perceived risk and blurs clinical decision-making.
2. Relative risk is not the same as absolute risk
If an outcome is uncommon in a comparison group, an odds ratio of four or even ten can still correspond to a minority of people in the exposed group experiencing that outcome. The 2022 review reported pooled odds ratios of 4.5 among men and 10.2 among women compared with population controls; heterogeneity was substantial in both analyses. These figures refer to comparisons of groups and cannot be interpreted as the percentage of men or women who will become violent. The higher relative odds estimated for women do not mean that women with schizophrenia are more likely than men with schizophrenia to perpetrate violence. Baseline rates differ between the comparison groups.
In register-based studies reviewed by Whiting and colleagues, recorded violent outcomes affected fewer than one in twenty women and fewer than one in four men over observation periods extending up to 35 years. Those broad long-term ceilings are not one-year probabilities and are not individual forecasts. Actual risk varies with follow-up length, study inclusion criteria, setting, type of event, and local social conditions. A person looking for a single universal percentage will find that no honest single number answers every meaning of 'violence.'
3. Why studies disagree
Studies have different participants, definitions, and settings. A sample selected from an acute psychiatric hospital or a forensic service is already unlike the population of all people living with schizophrenia in the community. A retrospective question about any aggression since illness onset is different from a five-year register of convicted offenses. Some studies account for substance use and socioeconomic circumstances; others cannot. The 2009 systematic review by Fazel and colleagues demonstrated substantial heterogeneity, and the later 2022 synthesis still found substantial variability. Study context is part of the result.
III. Does schizophrenia itself cause violence?
Schizophrenia is a clinical disorder involving disturbances of perception, thought, motivation, and functioning. Psychosis is a clinical syndrome that can arise in schizophrenia and also in mood disorders, substance-related conditions, delirium, and other medical circumstances. A hallucination is a perceptual experience without a corresponding external stimulus; a delusion is a fixed false belief evaluated in clinical context. Neither symptom is an act of violence. A confirmed diagnosis requires a professional longitudinal assessment; a reported symptom, internet quiz, or behavior that seems unfamiliar does not establish schizophrenia.
Observational evidence supports an association between schizophrenia spectrum diagnoses and certain violent outcomes. It cannot support the blanket causal claim that schizophrenia makes people violent. Risk is distributed unevenly across individuals and time, and some of the strongest correlates are shared with violence in the broader population: previous harmful behavior, substance misuse, exposure to violence, and destabilizing circumstances. The 2025 review also found that general positive-symptom severity was not consistently associated with violence across the available longitudinal studies.
1. Prior violence and current behavior matter
A documented history of violence was associated with subsequent violence in the 2025 synthesis, with a pooled odds ratio of 2.91, although study results varied. Prior behavior is relevant to assessing immediate safety because it is a concrete history rather than an assumption about a diagnostic label. Even here, history changes a probability and never makes the future inevitable. Clinicians must consider recency, seriousness, circumstances, protective factors, and the individual's present behavior. Historical categories should not become a reason for indefinite exclusion or coercion.
2. Substance use is one of the clearest modifiable correlates
Alcohol misuse, drug misuse, and co-occurring substance use disorders repeatedly appear in studies of psychosis and violence. The 2009 review found much higher violence associations among people with psychosis and substance-use comorbidity than among those without that comorbidity. The 2025 longitudinal review reported an overall pooled odds ratio of 2.36 for the substance-misuse domain; its secondary analysis also identified an association with a history of cannabis use, with limited replication for that specific estimate. These figures should not be added or multiplied together, because they arise from different samples and models.
Substance use may increase intoxication, impair judgment, worsen sleep, interfere with treatment, or interact with existing symptoms and social adversity. These are plausible pathways, not universal mechanisms proven by a single statistical association. The corresponding clinical response is to assess and treat both psychosis and substance-related problems in an integrated, nonpunitive way. Alcohol or drug use alone cannot determine whether a person is dangerous; a diagnosis of substance use disorder also does not mean violence will occur.
3. Symptoms, perceived threat, and acute distress
Persecutory beliefs, intense fear, severe disorganization, agitation, and the content of some hallucinations can be clinically relevant in specific cases, especially when a person perceives an immediate threat or feels compelled to act. The crucial issue is the person's interpretation, distress, intent, capacity to control actions, and behavior in the moment. A voice-hearing experience does not automatically lead to obedience, and a delusion does not automatically produce aggression. Assessing individual meaning and surrounding circumstances is more precise than assuming that a particular symptom predicts violence on its own.
Psychotic symptoms can also lead to withdrawal, frightened avoidance, help-seeking, or increased vulnerability to exploitation. Severe changes in mental state warrant compassionate assessment because they may signal need for treatment or medical evaluation. They do not warrant labeling an entire diagnostic group as threatening. In an acutely confused person, clinicians must also consider medical problems, intoxication, withdrawal, delirium, neurological conditions, and medication effects rather than assuming that schizophrenia explains every change.
4. Social adversity and interrupted care
Housing instability, repeated victimization, isolation, disrupted access to services, and poverty may interact with distress and affect opportunities for safety and recovery. These conditions can also change whether incidents reach police or healthcare records. Some are risk markers rather than proven direct causes. The practical implication is to improve stable housing, trusted relationships, continuity of care, and access to physical and mental healthcare. The WHO emphasizes the extensive social exclusion and rights violations that people with schizophrenia experience, making respectful community support a central concern.
IV. People with schizophrenia also experience violence as victims
Any discussion that counts only violence committed by people diagnosed with schizophrenia gives an incomplete picture. The NIMH states that people with schizophrenia are more likely to be harmed by others than people without the condition. A systematic review and meta-analysis in Schizophrenia Bulletin identified 27 studies of adults with psychotic disorders and reported a median violent-victimization prevalence of 20% across study samples, with estimates roughly four to six times higher than in the general community. The review also found associations with homelessness, substance problems, and several clinical and social circumstances.
An additional 2025 systematic review and meta-analysis combined 30 studies of adults with severe mental illness, a broader group than schizophrenia alone, and reported a pooled victimization prevalence of 33.3% over periods of up to 12 months. The authors emphasized substantial heterogeneity and inconsistent methods for measuring victimization. That number should not be misrepresented as a schizophrenia-specific rate, as the reviewed diagnostic populations were broader. Taken together, the evidence supports asking about abuse, threats, exploitation, and trauma during routine care.
1. Why victimization deserves direct attention
People experiencing psychosis may encounter unsafe housing, discrimination, manipulation, interpersonal abuse, sexual exploitation, or barriers to having their reports believed. Women and men can face different forms and contexts of victimization, and service planning must respond to an individual's circumstances rather than stereotypes. The risks can worsen when someone is homeless, isolated, dependent on an unsafe caregiver, or reluctant to contact authorities because of prior stigma. Victimization can in turn affect psychological distress, treatment continuity, and trust.
Clinicians should ask sensitively and privately about current safety, past experiences of harm, coercion, threats, and unmet practical needs. Responses should offer actual routes to protection and support; they should avoid assuming that any report is a delusion merely because the person has schizophrenia. The appropriate evaluation considers corroborating information when available, respects confidentiality, and takes a disclosure seriously while examining all relevant facts.
2. Harm within families can be discussed without blaming families or patients
Family caregivers can face fear, exhaustion, conflict, or sometimes actual aggression, and relatives may also experience threats or abuse from other people in the home. A 2023 systematic review of violence toward relatives in severe mental illness documents why family safety deserves attention, while also showing the limitations of pooling heterogeneous studies. A person who asks for safety planning should be listened to. At the same time, family membership, caregiving burden, or an individual's illness alone cannot establish which person is at risk or responsible for harm.
V. How clinicians assess violence risk in schizophrenia
Clinical assessment seeks to understand the current situation, prevent specific foreseeable harm, and identify services that may help. It normally covers recent threats or acts, access to potential weapons, substance intoxication or withdrawal, severe agitation, relevant symptom content, treatment interruption, sleep loss, previous serious incidents, and available supports. Professionals also assess whether a person is at risk of suicide, self-neglect, abuse, or urgent medical illness. These are distinct outcomes and need distinct responses. Diagnosis is documented, but assessment centers on the individual and the time-specific circumstances.
Some information may come from the person's own account, trusted caregivers when consent and law permit, records, and examination of mental and physical state. Risk formulation should remain dynamic: changes in sleep, acute distress, treatment access, substance use, and situational stress can alter what support is appropriate. Assessment is not simply an inventory of demographic traits or a sum of symptoms. Considerations of dignity and privacy should accompany the clinical need to protect others.
1. Why a screening score cannot predict a person's actions
The Oxford Mental Illness and Violence tool (OxMIV) derivation and validation study30109-8) demonstrated the value and limits of structured prediction. In its validation setting, a threshold designed to identify higher-risk people had a positive predictive value of 11%. That means most people flagged by the threshold did not have a recorded violent offense during the specified follow-up. The tool was designed as an adjunct to professional decision-making, especially for identifying low risk, not as a stand-alone verdict about an individual.
A 2023 external validation of OxMIV in first-episode psychosis services examined 1,145 people in two UK services. At a 10% risk cutoff, positive predictive value was 22% and negative predictive value 95% after model recalibration. These figures describe one clinical population, outcome definition, and follow-up period. They do not justify applying a risk score to strangers, excluding people from opportunities, or making coercive decisions without a complete individualized assessment. Apparent prediction accuracy depends on underlying event rates, and models may require validation across populations.
2. Immediate safety decisions require current facts
When someone has made a credible specific threat, is actively assaulting another person, has access to a weapon in an escalating situation, or is severely confused and unable to remain safe, immediate behavioral and environmental facts take priority. Move to a safe location, avoid confrontation, and contact local emergency services if danger is imminent. Communicate observable events rather than a diagnostic assumption. Clinical services can then assess psychiatric, substance-related, neurological, and medical contributors while addressing safety, legal requirements, and the least restrictive effective intervention.
VI. What treatment and support can change
The clearest clinical priority is effective, person-centered treatment for schizophrenia and co-occurring conditions, combined with practical support and continuity. The NICE guideline on psychosis and schizophrenia in adults recommends coordinated pharmacological, psychological, social, educational, and occupational interventions. Treatment is chosen according to the person's goals, diagnosis, prior response, benefits, side effects, and preferences. Risk reduction can be a relevant goal in specific cases, but recovery also encompasses distress, relationships, housing, employment, physical health, autonomy, and quality of life.
1. Antipsychotic medication
Antipsychotic medicines can reduce psychotic symptoms and are an established component of treatment for many people with schizophrenia. In a 2014 Swedish within-individual observational study60379-2), recorded violent-crime rates were lower during periods when individuals were prescribed antipsychotic medication than during periods when the same individuals were not; the reported hazard ratio was 0.55 in the study population. The design reduced confounding from characteristics that did not change within each person, yet changes over time and treatment selection could still influence results. The finding supports an association, rather than a guarantee that medication prevents a particular act.
Choice and monitoring matter. Weight and metabolic effects, movement symptoms, sedation, hormonal effects, cardiovascular considerations, and the person's subjective experience influence clinical decisions. The NICE guidance recommends shared selection of antipsychotic treatment and baseline and follow-up monitoring. People should discuss side effects, missed medication, desired changes, or treatment refusal with their prescribing clinician rather than making abrupt changes because of a generalized statement about violence risk.
2. Clozapine and persistent aggression
A 2024 systematic review by Faden and Citrome describes evidence supporting clozapine for persistent aggression and hostility among selected patients with schizophrenia or schizoaffective disorder. Clozapine has a particular role in treatment-resistant schizophrenia, and clinicians may consider its anti-aggressive effects when clinically relevant. Its potential benefits require balancing serious safety issues and mandatory monitoring, including blood counts and other adverse effects. It is a specialist-prescribed medication, not a general intervention for anyone described as irritable or intimidating.
3. Psychological, family, and community care
Cognitive behavioral therapy for psychosis, family intervention, psychoeducation, supported employment or education, and coordinated community care can help people manage distress, build coping strategies, strengthen communication, and stay connected with treatment. Family education can also support shared understanding of warning signs and collaborative problem-solving. Evidence for improvements in symptoms or functioning should not automatically be converted into a quantified claim about preventing violence; direct evidence on violent outcomes is more limited and varies by intervention and sample.
A person who has experienced violence, whether as a victim or perpetrator, may need trauma-informed care, a safe environment, and clear boundaries. For someone living with repeated crises, a coordinated plan can cover relapse warning signs, whom to call, medication consultation, access to a safe place, and the person's stated preferences. The broader evidence and options for schizophrenia treatment are covered in Schizophrenia Treatment: Medication, Therapy, and Coordinated Care.
4. Early intervention and treatment continuity
First-episode psychosis is an opportunity for prompt diagnostic assessment and coordinated support. The NICE guideline recommends assessing people presenting to early intervention services without delay and providing access to a broad range of treatments. These services can address co-occurring substance problems, social needs, and clinical deterioration before crises escalate. Their purpose includes stabilization and recovery; research into specific violence-prevention effects should be evaluated separately from evidence that coordinated specialty care improves other outcomes.
Sustained engagement may reduce periods in which a person has severe untreated symptoms or loses access to stable services. Relapse prevention should remain collaborative because trust, side effects, practical barriers, and preferences influence whether treatment is workable. See Schizophrenia Relapse: Warning Signs, Prevention, and Care for the condition-specific pathway.
VII. Practical guidance for relatives, friends, and caregivers
Relatives often need advice that respects the diagnosed person's autonomy while protecting everyone in the household. Begin with observable changes and immediate needs. A family can notice loss of sleep, increasing fear, intense agitation, intoxication, or a specific threat without claiming to know what diagnosis or outcome will follow. A calm conversation may identify a request for help, a frightening misunderstanding, an unmet medical need, or an actual safety concern. Ongoing family conflict, manipulation, and violence should be addressed based on behavior regardless of whether anyone has a psychiatric diagnosis.
1. When the situation is tense but not immediately dangerous
Use clear, respectful language; allow physical space; lower unnecessary stimulation; listen to distress without arguing aggressively about the content of a belief. If the person agrees, contact the treating team, a primary care clinician, or an early psychosis service to review worsening symptoms and practical support. Discuss a written crisis plan at a stable time. It can name the person's preferred contacts, previous helpful responses, and steps to protect privacy and safety. These actions are supportive strategies, not validated predictors or substitutes for professional assessment.
2. When a person is frightened by unusual experiences
Acknowledge that an experience is frightening without confirming an unverified claim. For example, a relative can respond to the person's fear and ask what would help them feel safe while encouraging assessment. The response may need to change if there is intoxication, escalating behavior, acute confusion, or a new medical concern. A conversation should not turn into a struggle to prove who is right at the moment of greatest distress.
3. When threats or physical violence have occurred
Take credible threats and assaults seriously and act on the actual behavior. Prioritize distance and safety, seek urgent help when necessary, and document factual incidents for appropriate services. Do not pressure a victim to remain in an unsafe home because the other person has an illness. A person diagnosed with schizophrenia also has a right to protection from violence, exploitation, or abusive caregiving. A competent assessment can address both treatment needs and the rights of everyone affected.
4. When unusual behavior suggests a medical emergency
Sudden disorientation, fluctuating attention, altered consciousness, fever, a seizure, a possible overdose, severe withdrawal, head injury, or rapidly changing behavior warrants prompt medical evaluation. These presentations can involve delirium or other acute medical problems and should not automatically be attributed to schizophrenia. Catatonia, severe inability to eat or drink, and suicidal intent also warrant urgent professional assessment. Emergency response depends on the person's symptoms and actual danger; the general emergency-care pathway sits beyond the narrower research question addressed here.
VIII. Media coverage, stigma, and the public understanding of risk
Dramatic crimes are memorable and comparatively uncommon events are easily overgeneralized. When a headline foregrounds a diagnosis even though other factors are more relevant, readers may incorrectly infer that the diagnosis predicts the act. Media reporting often omits denominators, comparison groups, observation periods, and the frequency of victimization among people living with mental illness. A good report explains what happened without treating every person with schizophrenia as a potential suspect. It also distinguishes a confirmed clinical diagnosis from a journalist's speculation about behavior.
Statements such as 'schizophrenia causes violence' or 'people with schizophrenia never pose risk' both lose information. Research supports a narrower and more useful account: the population-level association is measurable, absolute risks are context-dependent, most individuals do not perpetrate violence, victimization is substantial, and specific modifiable factors create opportunities for prevention. The WHO documents the consequences of stigmatization and discrimination for social participation and access to care. Precision supports public safety because it directs attention to actual needs instead of myths.
1. Questions to ask when reading an alarming statistic
Identify the outcome and time period. Was the study about reported aggression, self-report, arrest, conviction, or homicide? Did researchers compare with the general population, siblings, people with other mental disorders, or a clinical subgroup? Was the estimate an odds ratio, an absolute proportion, or a hazard ratio? Were substance misuse, previous violence, and socioeconomic factors measured? A quote without these details may be technically true yet profoundly misleading. Never turn an aggregate research result into a verdict about a known individual.
IX. Frequently asked questions
1. Are people with schizophrenia more violent than other people?
Some large observational studies show higher relative rates of defined violent outcomes in groups with schizophrenia spectrum disorders than in community comparison groups. That is a population-level association, accompanied by substantial study-to-study variation. Most people with schizophrenia do not perpetrate violence. Individual risk must be assessed from history, current behavior, substance use, clinical condition, and circumstances rather than diagnosis alone. Whiting et al., 2022.
2. What percentage of people with schizophrenia are violent?
There is no single reliable percentage without specifying what counts as violence, how the sample was selected, and over what period it was measured. In part of the 2022 meta-analysis, register-based violent outcomes occurred in fewer than 5% of women and fewer than 25% of men with schizophrenia spectrum disorders over observation extending up to 35 years. These are neither annual probabilities nor estimates for every setting. Other studies measuring broader aggression can yield different rates.
3. Does hearing voices make a person dangerous?
No. Hearing voices is a possible hallucination, not evidence of violent intent or action. Clinicians explore what the voices mean to the person, associated distress, their response, whether they perceive threats, and the broader clinical situation. Voice-hearing also occurs in clinical contexts other than schizophrenia, and a hallucination alone does not establish a diagnosis. Concerns become actionable when actual threats, loss of control, or imminent harm are present.
4. Can persecutory delusions increase risk?
In some circumstances, intensely feared or threatening interpretations may form part of an escalating crisis, especially alongside agitation, intoxication, or previous harmful behavior. A persecutory belief by itself is not proof that someone will act violently. Longitudinal research finds that simple measures of total positive-symptom severity are weak and inconsistent violence predictors compared with factors such as prior violence and substance misuse. Lagerberg et al., 2025.
5. Is schizophrenia associated with homicide?
Some studies observe an elevated relative association with homicide, an extremely rare outcome. Because baseline rates are very low, a large relative ratio does not describe how most people with schizophrenia behave. Homicide data are additionally influenced by ascertainment, jurisdiction, and co-occurring factors. Presenting a striking odds ratio without absolute frequency or context distorts the findings. Whiting et al., 2022.
6. Are people with schizophrenia more likely to be victims of violence?
Yes, research consistently identifies elevated victimization in people with psychotic disorders and severe mental illness. A schizophrenia-specific comparison is not interchangeable with studies of broader diagnostic groups, but both NIMH guidance and systematic-review evidence support treating victim safety as a priority. Clinicians should ask about abuse and threats rather than focusing only on the possibility that someone might harm others. NIMH; de Vries et al..
7. Does medication eliminate the risk of violence?
No medication can guarantee that a person will never behave violently. Antipsychotics are used to treat schizophrenia, and some observational studies associate their use with lower recorded violent-crime rates. The choice to use or change medication is a clinical decision based on symptoms, preferences, prior response, and safety monitoring. A systematic review of clozapine supports a potential role in selected cases with persistent aggression, with substantial monitoring requirements.
8. What should I do if someone with psychosis threatens me?
Respond to the seriousness of the threat and the immediate circumstances. If harm is imminent, get to safety and contact local emergency services. Avoid confrontation and do not attempt to manage an armed or escalating situation alone. For concerning changes without immediate danger, seek prompt evaluation through the person's treating clinician or local mental health services. Describe actual threats, behavior, intoxication, or medical symptoms; a suspected diagnosis is not an emergency assessment.
9. Can someone with schizophrenia have safe relationships, work, and a fulfilling life?
Yes. Outcomes vary, and people with schizophrenia can build meaningful relationships and participate in education, employment, and community life. Treatment and social supports can improve functioning and recovery. The WHO reports that at least one in three people with schizophrenia can achieve full recovery, while others experience varying needs over time. A diagnosis cannot determine a person's character, moral worth, capacity for connection, or future.
X. What the evidence means for clinical practice and everyday life
The most effective interpretation of the evidence is concrete: address threats and harmful behavior whenever they occur; identify previous violence, substance problems, acute deterioration, and real-world stressors; protect people from victimization; improve access to sustained, respectful care; and distinguish clinical risk assessment from stigma. For clinicians, that means a formulation centered on behavior and modifiable factors. For families, it means boundaries, support, and a safety plan where warranted. For public discussion, it means naming the actual measured outcome and denominator before repeating a dramatic statistic.
Schizophrenia can involve profound distress and functional disruption, yet individual trajectories differ. Research on violence is most useful when it informs prevention and protection while preserving the rights of people affected by psychosis. The broader diagnostic foundation is described in Schizophrenia: Symptoms, Causes, Diagnosis, Treatment, and Recovery; Schizophrenia Spectrum Disorders: Types, Differences, and Clinical Diagnosis explains why diagnostic categories and symptom presentations must remain distinct.
Related Articles
References
de Vries, B., van Busschbach, J. T., van der Stouwe, E. C. D., et al. (2019). Prevalence rate and risk factors of victimization in adult patients with a psychotic disorder: A systematic review and meta-analysis. Schizophrenia Bulletin, 45(1), 114–126. https://doi.org/10.1093/schbul/sby020
Faden, J., & Citrome, L. (2024). A systematic review of clozapine for aggression and violence in patients with schizophrenia or schizoaffective disorder. Schizophrenia Research, 268, 265–281. https://doi.org/10.1016/j.schres.2023.11.008
Fazel, S., Gulati, G., Linsell, L., Geddes, J. R., & Grann, M. (2009). Schizophrenia and violence: Systematic review and meta-analysis. PLOS Medicine, 6(8), e1000120. https://doi.org/10.1371/journal.pmed.1000120
Fazel, S., Wolf, A., Larsson, H., Lichtenstein, P., Mallett, S., & Fanshawe, T. R. (2017). Identification of low risk of violent crime in severe mental illness with a clinical prediction tool (OxMIV): A derivation and validation study. The Lancet Psychiatry, 4(6), 461–468. https://doi.org/10.1016/S2215-0366(17)30109-830109-8)
Fazel, S., Zetterqvist, J., Larsson, H., Långström, N., & Lichtenstein, P. (2014). Antipsychotics, mood stabilisers, and risk of violent crime. The Lancet, 384(9949), 1206–1214. https://doi.org/10.1016/S0140-6736(14)60379-260379-2)
Lagerberg, T., Lambe, S., Paulino, A., Yu, R., & Fazel, S. (2025). Systematic review of risk factors for violence in psychosis: 10-year update. The British Journal of Psychiatry, 226(2), 100–107. https://doi.org/10.1192/bjp.2024.120
National Institute for Health and Care Excellence. (2014, updated). Psychosis and schizophrenia in adults: Prevention and management (CG178). NICE guideline
National Institute of Mental Health. (n.d.). Schizophrenia. U.S. National Institutes of Health. NIMH patient information
Rocha, I. C. O., Valença, A. M., Coutinho, E. S. F., Banhara, P. B., & Berger, W. (2025). Prevalence of victimization among adults with severe mental illness worldwide: A systematic review and meta-analysis. Trauma, Violence, & Abuse. https://doi.org/10.1177/15248380251386799
Whiting, D., Gulati, G., Geddes, J. R., & Fazel, S. (2022). Association of schizophrenia spectrum disorders and violence perpetration in adults and adolescents from 15 countries: A systematic review and meta-analysis. JAMA Psychiatry, 79(2), 120–132. https://doi.org/10.1001/jamapsychiatry.2021.3721
Whiting, D., Mallett, S., Lennox, B., & Fazel, S. (2023). Assessing violence risk in first-episode psychosis: External validation, updating and net benefit of a prediction tool (OxMIV). BMJ Mental Health, 26(1), e300634. https://doi.org/10.1136/bmjment-2022-300634
Wildman, E. K., et al. (2023). Prevalence of violence by people living with severe mental illness against their relatives and its associated impacts: A systematic review. Acta Psychiatrica Scandinavica, 147(2), 155–174. https://doi.org/10.1111/acps.13516
World Health Organization. (2026, September 11). Schizophrenia. WHO fact sheet