Schizophrenia and Crime: What Research Shows

By Fahad Hizam, fraud and risk consultant in Mexico and LATAM. Fahad works on fraud, due diligence, corporate security, and asset-tracing matters for legal, business, and institutional clients.

Schizophrenia can be associated with increased risk of violent offending at the population level, especially when substance misuse, prior violence, treatment disruption, or acute symptoms are present. People with schizophrenia also face elevated exposure to harm, neglect, homelessness, exploitation, victimization, and criminal justice contact. For investigators, security teams, and risk practitioners, the correct standard is evidence-based risk assessment grounded in documented behavior, treatment records, and current case facts.

Schizophrenia is a serious mental health condition involving impaired perception of reality and changes in thought, feeling, and behavior. It may include hallucinations, delusions, disorganized thinking, reduced motivation, social withdrawal, and problems with daily functioning (National Institute of Mental Health, n.d.; World Health Organization, 2025).

Population studies show elevated relative risk of violence among people with schizophrenia spectrum disorders. In the largest recent review, fewer than 1 in 20 women and fewer than 1 in 4 men in the included studies had violent outcomes (Whiting et al., 2022). Elevated relative risk and low absolute risk can exist in the same data.

Population-level evidence

Early research on schizophrenia and violence often used clinic samples, small groups, or case reports. Fazel, Gulati, Linsell, Geddes, and Grann (2009) changed the discussion by pooling data in a systematic review and meta-analysis. Their work found an association between schizophrenia and violent behavior, with risk strongly affected by substance misuse. It remains useful as historical background because it pulled scattered findings into one evidence base.

Whiting, Lichtenstein, and Fazel (2022) later reviewed schizophrenia spectrum disorders and violence across 24 studies from 15 countries. That paper is the main population-level source for schizophrenia violent crime risk because it covers different settings, study designs, and legal systems. The study found increased relative risk of violence perpetration among people with schizophrenia spectrum disorders. It also reported absolute-risk figures that keep the finding in scale.

A large Israeli population study adds national-level detail. Fleischman, Werbeloff, Yoffe, Davidson, and Weiser (2014) studied 3,187 patients with schizophrenia. They reported that 15.1% of patients had a violent crime record, compared with 4.0% of controls. The odds ratio was 4.3. Men and women both showed increased risk, although the absolute rate was higher among men.

The timing finding is useful for case work. Among patients with schizophrenia who had a violent crime record, 65% had their first crime before first psychiatric admission (Fleischman et al., 2014). This does not show that the diagnosis caused the later offense. It shows that criminal justice contact, symptoms, family stress, and delayed care can appear before a case has a clear medical record.

Munkner, Haastrup, Joergensen, and Kramp (2003) found a similar timing pattern in Denmark. In plain language, about 37% of male offenders with schizophrenia committed crimes before first psychiatric contact. The precise figure from their table was 36.5%. The shared point across the Danish and Israeli studies is stronger than either single number: a first registered offense can predate first hospital admission or diagnosis.

Biology research adds possible mechanisms, with limits. Yılmaz, Öner, Taşcı, and Kaya (2023) studied 130 men: 50 schizophrenia patients involved in crime, 40 schizophrenia patients not involved in crime, and 40 healthy controls. They reported lower serum oxytocin levels in the crime-involved schizophrenia group. Among patients with schizophrenia, oxytocin levels were negatively correlated with PANSS total score (r = −0.339, p = 0.016) and Buss-Perry Aggression Questionnaire score (r = −0.551, p < .001), and positively correlated with empathy (r = 0.699, p < .001) and forgiveness scores (r = 0.682, p < .001).

That finding should stay in its place. The study had a small sample of 130 men, one clinical setting, an observational design, and exclusions for substance abuse and personality disorder. It does not create a test for dangerousness. No single biomarker replaces behavioral evidence or clinical evaluation.

Substance misuse and risk

Substance misuse is one of the most important modifiable risk factors in schizophrenia and violence research. Fazel, Långström, Hjern, Grann, and Lichtenstein (2009) used Swedish national registers from 1973 to 2006. They compared 8,003 patients with schizophrenia with 80,025 general population controls and also examined 8,123 unaffected siblings.

The Swedish study found that 1,054 patients with schizophrenia, or 13.2%, had at least 1 violent offense after diagnosis. The rate among general population controls was 5.3%. The adjusted odds ratio was 2.0 when patients were compared with general population controls, and 1.6 when compared with unaffected siblings (Fazel, Långström, et al., 2009).

The substance split changed the practical meaning. Among patients with schizophrenia and substance abuse comorbidity, 27.6% had a violent offense, with an adjusted odds ratio of 4.4 compared with general population controls. Among patients without substance abuse comorbidity, 8.5% had a violent offense, with an adjusted odds ratio of 1.2. The increase was much smaller in the group without recorded substance abuse.

The sibling comparison narrowed the finding further. When patients with substance abuse were compared with their unaffected siblings (28.3% vs. 17.9%), the odds ratio dropped to 1.8. Fazel, Långström, et al. (2009) interpreted this as evidence that family background, genetics, or early environment may affect the association.

Patients whose substance abuse was recorded after their schizophrenia diagnosis had a violent crime rate more than double that of patients whose substance abuse was recorded earlier: 34.5% vs. 15.6%. The adjusted odds ratio was 6.4 in the later-substance-abuse group and 1.9 in the earlier group, compared with general population controls (Fazel, Långström, et al., 2009).

Fleischman et al. (2014) found a similar substance difference in Israel. Violent crime was recorded for 32.9% of patients with substance misuse and 8.7% of patients without it. Any serious discussion of schizophrenia crime research needs to separate diagnosis from substance use, prior violence, and current treatment status.

What case-level factors were associated with violence in a forensic sample? Lin et al. (2024) studied 308 violent offenders with schizophrenia and 139 nonviolent individuals with the same diagnosis in Hunan Province, China. Prior violence (OR = 2.88) and persecutory delusions (OR = 2.57) were associated with violence. Regular treatment in the previous 4 weeks was associated with lower risk (OR = 0.29). The study was retrospective and limited to one province, so it should support case-level thinking rather than broad population claims.

Absolute vs. relative risk

Relative risk can sound large. Absolute risk gives scale.

Whiting et al. (2022) found increased relative risk of violence perpetration among people with schizophrenia spectrum disorders. They also reported that absolute risks were “less than 1 in 20 women and less than 1 in 4 men.” Those two facts should sit next to each other without collapsing one into the other.

For a case file, this distinction is basic. A population-level association can justify careful assessment. It does not justify suspicion based only on a diagnosis.

A risk section should list dates, incident descriptions, witness names, substance-use records, treatment timelines, and collateral sources. Write the facts: recent threats, assault history, weapon access, stalking behavior, domestic conflict. Add substance use, treatment status, acute psychosis, victim reports, police contacts, and collateral statements.

The safest wording is direct: people with schizophrenia spectrum disorders show elevated relative risk of violence in population studies. Absolute-risk figures show that most individuals in these studies did not commit violent offenses.

This distinction matters in security, insurance, journalism, family disputes, and legal work. A diagnosis may be one background fact. The working question is behavior: what happened, when it happened, who observed it, what changed, and what current risk factors exist.

Public perception and legal judgment

Queen and Goncy (2026) tested how a schizophrenia diagnosis affected public perceptions of offenders in violent and nonviolent crime conditions. Participants rated offenders with a schizophrenia diagnosis as less negative, less deserving of incarceration, and less severely criminal than offenders without the diagnosis. This appeared in both violent and nonviolent crime conditions.

For case evaluation, the practical point is narrow. A diagnosis can affect legal perception. The file should rely on documented behavior, capacity evidence, treatment records, witness statements, and risk factors rather than diagnostic labels.

State what is known. Separate diagnosis from conduct. Identify who observed the behavior. Record dates, locations, threats, weapons, injuries, treatment status, substance use, and prior incidents. Avoid psychiatric conclusions unless they come from qualified clinical records or expert opinion.

Victimization

The victim side of this topic is easy to miss. Khalifeh, Oram, Osborn, Howard, and Johnson (2016) reviewed recent domestic and sexual violence against adults with severe mental illness. Across included studies, recent domestic violence ranged from 15% to 22% among women and from 4% to 10% among men or mixed samples. Median prevalence of recent sexual violence was 9.9% in women and 3.1% in men, with about 6-fold higher odds of sexual victimization compared with the general population.

Latalova, Kamaradova, and Prasko (2014) found wide variation across studies because samples, settings, and definitions differed. The broad finding still holds: adults with severe mental illness are often exposed to violent victimization, especially when homelessness, substance use, active symptoms, and prior criminal justice contact are present.

Official-record data add another layer. Short, Thomas, Luebbers, Mullen, and Ogloff (2013) compared 4,168 people with schizophrenia-spectrum disorders with 4,641 community controls in Victoria, Australia. Patients were more likely to have a record of violent victimization (10.1% vs. 6.6%, OR = 1.4) and sexual victimization (1.7% vs. 0.3%, OR = 2.77), while they were less likely to have an official victimization record overall (28.7% vs. 39.1%, OR = 0.5).

Official records can undercount harm. The undercount may be worse when a person fears police, lacks stable housing, has prior justice contact, or has trouble describing events clearly. In a real case file, victimization data belong next to perpetration data because the same person may appear as a suspect, complainant, witness, missing person, or victim.

Mexico and LATAM

For Mexico and LATAM, the same documentation standard applies: describe the case conditions before drawing any risk conclusion. The international studies above can guide risk questions, but they should not be treated as country-specific crime estimates for Mexico or Latin America.

He, Gu, Wang, Li, Li, and Hu (2022) give a limited comparison point from China. In a forensic psychiatric sample, migration status, employment, and access to care shaped crime patterns among people with schizophrenia. That finding should be used only as a comparison. It does not prove the same pattern in Mexico or LATAM.

Treatment access is a practical issue in Mexico. Cabello-Rangel, Díaz-Castro, and Pineda-Antúnez (2020) modeled treatment options for schizophrenia under universal public financing. In a hypothetical population of 1,000,000 people, they estimated 2,200 schizophrenia cases and assumed 80% coverage. Their table reported 347 DALYs avoided for typical or atypical antipsychotic treatment plus basic psychosocial treatment, and 417 DALYs avoided for olanzapine plus basic psychosocial treatment. The same paper estimated avoided out-of-pocket expenses from US $101,221.72 to US $787,498.71, depending on the intervention.

If treatment access is uneven, the file should say how the person received care, whether medication was available, who paid for it, and whether treatment stopped because of money, distance, family conflict, or system access. Those facts may matter more than the diagnosis label.

Yarris and Ponting (2019) add Mexican clinical context. Their clinical ethnography at a public psychiatric hospital in central Mexico showed how schizophrenia affected men through work, family roles, masculinity, and dependence on relatives. Their cases involved unemployment, rural or urban residence, family caregiving, and the cost of maintaining care. This source is not crime data. It is useful because it shows why Mexican case files should record family structure, work history, caregiving, shame, treatment access, and social pressure.

Documentation standards for investigators

In my work reviewing fraud and risk cases in Mexico, I have seen files where a schizophrenia diagnosis was the entire risk section. No dates, no behavior description, no treatment timeline, no substance-use check. I would not rely on that file.

Start with verified behavior. What was said or done? Who saw it? Was there a threat, assault, weapon, stalking pattern, property damage, coercion, exploitation, self-neglect, or missing-person concern? Record the source for each claim.

Document treatment with dates. Include the last known appointment, medication name when known, reported adherence, hospital discharge date, recent dose change, and any gap in access. Avoid broad phrases like “off meds” unless the source is clear.

Substance use needs its own section. Record alcohol, cannabis, stimulants, sedatives, opioids, inhalants, and other drugs when relevant. Note frequency, recent change, intoxication during the incident, withdrawal, and source of information.

Prior violence should be separated by type and proof level: alleged incidents, arrests, convictions, restraining orders, workplace reports, family reports, and medical records. Each item needs date, source, and status. Current context also matters: housing, work, sleep, treatment contact, recent discharge, debt, eviction, breakup, grief, family conflict, and weapon access.

Victimization history belongs in the review. Assault, exploitation, threats, financial abuse, neglect, coercion, and housing loss can affect risk, vulnerability, and witness reliability.

The final report should separate facts, statements, records, and opinions. Use labels such as “reported by mother,” “shown in medical record,” “observed by investigator,” “police record,” “unverified claim,” and “clinical opinion.” This protects the subject, the client, the victim, and the investigation.

For investigators, the file should contain documented behavior, treatment records, substance-use history, collateral sources, and victimization data. Diagnosis alone is not a threat assessment.

References

Cabello-Rangel, H., Díaz-Castro, L., & Pineda-Antúnez, C. (2020). Cost-effectiveness analysis of interventions to achieve universal health coverage for schizophrenia in Mexico. Salud Mental, 43(2), 65–71. https://doi.org/10.17711/SM.0185-3325.2020.010

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., Långström, N., Hjern, A., Grann, M., & Lichtenstein, P. (2009). Schizophrenia, substance abuse, and violent crime. JAMA, 301(19), 2016–2023. https://doi.org/10.1001/jama.2009.675

Fleischman, A., Werbeloff, N., Yoffe, R., Davidson, M., & Weiser, M. (2014). Schizophrenia and violent crime: A population-based study. Psychological Medicine, 44(14), 3051–3057. https://doi.org/10.1017/S0033291714000695

He, Y., Gu, Y., Wang, S., Li, Y., Li, G., & Hu, Z. (2022). Migration, schizophrenia, and crime: A study from a forensic psychiatric sample. Frontiers in Psychiatry, 13, Article 869978. https://doi.org/10.3389/fpsyt.2022.869978

Khalifeh, H., Oram, S., Osborn, D., Howard, L. M., & Johnson, S. (2016). Recent physical and sexual violence against adults with severe mental illness: A systematic review and meta-analysis. International Review of Psychiatry, 28(5), 433–451. https://doi.org/10.1080/09540261.2016.1223608

Latalova, K., Kamaradova, D., & Prasko, J. (2014). Violent victimization of adult patients with severe mental illness: A systematic review. Neuropsychiatric Disease and Treatment, 10, 1925–1939. https://doi.org/10.2147/NDT.S68321

Lin, R., Li, Q., Liu, Z., Zhong, S., Huang, Y., Cao, H., Zhang, X., Zhou, J., & Wang, X. (2024). Risk factors for violent crime in patients with schizophrenia: A retrospective study. PeerJ, 12, e18014. https://doi.org/10.7717/peerj.18014

Munkner, R., Haastrup, S., Joergensen, T., & Kramp, P. (2003). The temporal relationship between schizophrenia and crime. Social Psychiatry and Psychiatric Epidemiology, 38(7), 347–353. https://doi.org/10.1007/s00127-003-0650-3

National Institute of Mental Health. (n.d.). Schizophrenia. U.S. Department of Health and Human Services. https://www.nimh.nih.gov/health/topics/schizophrenia

Queen, M. M., & Goncy, E. (2026). Impact of offender schizophrenia diagnosis on public perceptions of crime and punishment. Behavioral Sciences & the Law. https://doi.org/10.1002/bsl.70040

Short, T. B. R., Thomas, S., Luebbers, S., Mullen, P., & Ogloff, J. R. P. (2013). A case-linkage study of crime victimisation in schizophrenia-spectrum disorders over a period of deinstitutionalisation. BMC Psychiatry, 13, Article 66. https://doi.org/10.1186/1471-244X-13-66

Whiting, D., Lichtenstein, P., & 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

World Health Organization. (2025). Schizophrenia. https://www.who.int/news-room/fact-sheets/detail/schizophrenia

Yarris, K. E., & Ponting, C. (2019). Moral matters: Schizophrenia and masculinity in Mexico. Ethos, 47(1), 35–53. https://doi.org/10.1111/etho.12226

Yılmaz, S., Öner, P., Taşcı, G., & Kaya, Ş. (2023). Low oxytocin levels in schizophrenia patients involved in crime and the relationship of these levels to aggression, empathy and forgiveness. The Journal of Forensic Psychiatry & Psychology, 34(1), 1–19. https://doi.org/10.1080/14789949.2022.2156378