Mental Illness and Crime: What the Evidence Actually Shows
Author: Ukrainian Psychological Hub · Published: October 11, 2026 · Editorial Policy
Mental illness and crime have a complex relationship. Psychiatric diagnoses can be associated with higher rates of particular recorded offenses in some populations, but diagnosis alone is a poor basis for predicting whether a person will commit a crime. Most people who live with mental illness are not violent offenders. The best research examines the type of offense, the population studied, substance use, previous violence, environmental circumstances, victimization, and the rules through which criminal justice systems record behavior.
This evidence review answers the question “mental illness and crime” using population studies, longitudinal research, systematic reviews, justice statistics, and professional guidance. It distinguishes psychological constructs, clinical disorders, behavioral events, police contacts, criminal convictions, statistical risk estimates, forensic opinions, and legal decisions. The purpose is to explain what research supports, why certain associations appear, where estimates fail, and what follows for public safety and humane mental health policy.
What the research shows at a glance
The scientific evidence supports several connected conclusions. Mental illness is a broad group of conditions whose associations with offending differ greatly. Some studies find higher recorded rates of violent offending in defined clinical subgroups, especially when substance use disorders also occur. Those group differences never imply that most affected people commit violence or that a diagnosis proves a crime. Many measured associations also involve prior violence, exposure to dangerous environments, social disadvantage, and justice-system selection.
In a prospective US survey of 34,653 adults, schizophrenia, major depression, and bipolar disorder considered alone did not independently predict later self-reported violence after relevant historical and contextual variables were considered. Substance use, past violence, and other factors carried important predictive information. The analysis concerns specific disorders, outcomes, and survey waves, rather than every possible clinical presentation. Elbogen and Johnson (2009).
A Swedish population study found that 13.2% of people with schizophrenia had at least one conviction for a defined violent offense during follow-up, compared with 5.3% of matched controls. The rate among people with schizophrenia and substance abuse comorbidity was 27.6%, versus 8.5% among those without recorded substance abuse. The period, setting, conviction definitions, and selection criteria are essential to interpreting those figures. They are not contemporary US prevalence estimates or individual predictions. Fazel and colleagues (2009).
Another essential finding is that people with mental disorders can be particularly vulnerable to being subjected to crime. Research into victimization is necessary alongside research into offending. Dean and colleagues (2018).
Defining mental illness, crime, violence, and justice-system contact
A psychiatric diagnosis is a clinical classification
Mental illness includes heterogeneous conditions such as depression, bipolar disorder, psychotic disorders, anxiety disorders, substance use disorders, and personality disorders. Clinical diagnosis, symptom severity, a past treatment record, a self-reported history, and a brief screening score are distinct kinds of evidence. Studies that define a group using one of these measures should not be represented as though they measured the others.
A diagnostic label concerns a health condition. A symptom describes an experience or behavior; a screening result suggests whether further assessment may be warranted; and a forensic opinion responds to a specific psycholegal question. None of these automatically establishes criminal intent, guilt, legal incapacity, or future dangerousness. The American Psychological Association's Specialty Guidelines require accuracy, impartiality, and attention to the limits of forensic evidence.
Crime is determined by the relevant jurisdiction's law
Crime refers to conduct classified as an offense under applicable law, which varies across jurisdictions and changes over time. Researchers may measure self-reported behavior, police contact, arrests, charges, convictions, incarceration, or technical violations of supervision. These outcomes are related but cannot be treated as interchangeable. An arrest is not a conviction; a conviction is not a complete record of all offending; and administrative records also reflect reporting, surveillance, policing, and judicial practices.
Violent crime is more specific than crime in general. Different studies include different combinations of assault, homicide, robbery, sexual offenses, arson, threats, and intimidation. Evidence on violent convictions should not automatically be used to explain property crime or public-order offenses. Homicide is a particularly rare outcome, and data on ordinary aggression do not supply a reliable homicide forecast.
Why measurement changes the answer
A clinical sample of hospitalized patients, a national community survey, and a cohort of previously convicted prisoners answer different research questions. Their base rates, histories, social contexts, and selection rules differ. A study of people discharged from forensic hospitals is especially unsuited to representing everyone who has the same diagnostic condition in the general community.
The same caution applies to place and time. Findings from Swedish criminal registers during previous decades provide valid evidence for that population and design, but legal definitions, treatment systems, and social conditions may differ in the United States in 2026.
Is mental illness associated with violent crime?
Relative risk is not the probability that an individual will offend
A relative-risk estimate compares the frequency of an outcome across groups; an absolute rate describes how frequently the outcome occurred in a specified group during a defined period. A group can have elevated relative risk while the large majority of its members never commit the measured offense. Interpretation also requires the reference group, duration, and precise behavioral endpoint.
A systematic review of schizophrenia and violence covering 20 studies and 18,423 people with schizophrenia or other psychoses found substantial differences between study estimates. Co-occurring substance abuse was a major factor. In the subset of studies where population-attributable fractions were calculable, the estimated fractions were at or below 10%; these are model-dependent population estimates, not proof that particular crimes were caused by illness and not an estimate for all psychiatric disorders.
An influential prospective study compared violence in individuals discharged from acute psychiatric facilities with violence among people living in the same neighborhoods. Substance-related symptoms were important, showing why a clinical sample compared only with a broad national average can misrepresent the relationship. Steadman and colleagues (1998).
Why a simple yes-or-no answer fails
Mental health conditions can differ by symptom presentation, severity, comorbidity, access to care, and social circumstances. Research designs also differ in statistical adjustment, whether outcomes are measured prospectively, and the kinds of offenses included. These differences produce heterogeneous results rather than one universal estimate.
A responsible conclusion preserves both facts: associations can be important for specified outcomes and populations, and most people with psychiatric disorders do not engage in violence. Diagnoses alone do not identify perpetrators.
Schizophrenia, psychosis, and criminal behavior
Some population studies find elevated rates of recorded violent offenses in groups with schizophrenia. The association is sensitive to substance use and other shared factors. In the Swedish study, the relationship was much greater in the group with substance abuse comorbidity; analyses involving unaffected siblings suggested familial and environmental confounding. Treatment information was unavailable, limiting conclusions about whether particular interventions explained the results. Fazel and colleagues (2009).
A person can experience delusions, hallucinations, disorganization, or fear without committing any crime. In a specific legal case, whether a symptom was present when an act occurred requires contemporaneous evidence. Diagnosis alone cannot reconstruct a defendant's thoughts, motivation, understanding, or intentions at that time.
The English Hub addresses the clinical literature in greater depth in Schizophrenia and Violence: What Research Really Shows. The current article maintains the broader, cross-diagnostic forensic question of mental illness and crime.
Bipolar disorder and offending
A population study and systematic review found elevated recorded violent convictions among individuals with bipolar disorder in a Swedish cohort. The observed association was largely concentrated in those with substance abuse comorbidity. After comparing individuals without that comorbidity against unaffected siblings, the estimated excess was further attenuated. The research did not support treating the diagnosis or one episode category as a simple individual predictor. Fazel and colleagues (2010).
Agitation, reduced sleep, impulsivity, or psychotic symptoms may be clinically relevant under certain circumstances. Their presence, intensity, timing, and actual relationship to an event have to be established. A diagnostic history does not answer a question about guilt, legal responsibility, or a particular future offense.
Depression and criminal behavior
Depression varies widely in clinical presentation and associated difficulties. A Swedish population study identified an association between an outpatient depression diagnosis and later recorded violent crime and also examined symptoms in a twin cohort. The findings require attention to possible confounding, subgroup variation, and the outcome's legal definition. Fazel and colleagues (2015).
Depression should be assessed for its effects on suffering, functioning, treatment needs, and individual safety. If a specific crime is under consideration, additional evidence about the circumstances is essential. Neither the presence nor the absence of depression can resolve that forensic question.
Substance use disorders and co-occurring conditions
Across several major studies, substance use emerges as a relevant component of the association between severe mental illness and recorded violence. That association can involve intoxication, withdrawal, judgment in particular situations, social and environmental exposures, and accumulated disadvantage. The relative importance of these mechanisms varies by person, substance, context, and outcome.
The prospective US analysis by Elbogen and Johnson linked subsequent violence to past violence, substance abuse or dependence, and social factors rather than finding an independent predictive effect of severe mental illness alone in multivariable models. In the Swedish schizophrenia study, substance abuse comorbidity considerably changed the observed comparison.
A substance use disorder is a clinical diagnosis; it is not a legal finding that someone committed a drug offense or a violent act. Findings about one drug, type of use, or justice-system group should not be generalized to all substance-related conditions.
Personality disorders, trauma-related conditions, and behavioral traits
Research sometimes reports differences among people diagnosed with certain personality disorders or with histories of trauma. These broad categories contain heterogeneous experiences and presentations. Social adversity, substance use, previous victimization, prior justice involvement, and behavioral histories may explain part of an association. Researchers must separate these components and avoid inferring an inevitable pathway from psychiatric traits to crime.
Antisocial behavior is a behavioral description; antisocial personality disorder is a clinical diagnosis; psychopathy is a particular assessment construct; and a criminal conviction is a legal finding. A psychometric score is not a verdict or proof of guilt. Likewise, a history of trauma describes an experience, not an inevitable cause of later offending.
Mechanisms and pathways: what may explain the association?
Previous violence and justice-system history
Past behavior can contribute useful information to appropriately defined group-level recurrence estimates, particularly when the outcome is similar to the behavior observed previously. Prior arrests or convictions also embody decisions made by legal institutions and may reflect unequal detection or enforcement. A prior history cannot establish that a new offense took place.
The US prospective survey found historical factors such as prior violence and juvenile detention relevant to later self-reported violence. These should be recorded separately from diagnostic status and from any current risk-management decision.
Social context and disadvantage
Unstable housing, poverty, unemployment, lack of accessible care, unsafe environments, exposure to violence, and weak social support can interact with mental health difficulties and justice contact. These conditions have complex, potentially reciprocal relationships. A correlation between unemployment and offending, for example, does not establish that unemployment caused a particular crime.
Strong causal explanations require temporal ordering, appropriate comparison groups, careful measurement, and consideration of confounders. Even longitudinal observational studies can leave important factors unmeasured.
Acute symptoms and event-specific circumstances
A psychiatric crisis may require urgent assessment or treatment. In a forensic case, the meaningful question is which symptoms were actually present during the relevant period, what behavior occurred, what evidence supports the account, and how the applicable legal standard treats those facts. Unusual conduct or shocking allegations alone cannot establish psychosis, lack of capacity, or a causal psychiatric explanation.
Treatment, continuity of care, and support
Treatment can improve symptoms, functioning, and quality of life. Whether it reduces a given criminal outcome is a separate empirical question. Effective services may need to address substance use, housing, conflict, financial strain, and transition from correctional settings alongside psychiatric care. No single treatment effect should be claimed without an appropriate study of that outcome.
Victimization: the other side of mental illness and crime
One of the clearest limitations of public discussions is their narrow focus on offending. People living with psychiatric conditions can also experience elevated exposure to assault, exploitation, and other crimes. That vulnerability matters for clinical practice, policing, social policy, and legal protection.
A Danish national cohort study followed more than two million people using psychiatric and police records. After recorded onset of mental illness, the adjusted incidence rate ratio for violent crime victimization was 1.76 for men and 2.72 for women, compared with those without recorded mental disorders. These ratios are not absolute probabilities, and the outcomes were police-reported. The analyses adjusted for participants' own offending and other factors without eliminating every possible confounder. Dean and colleagues (2018).
Victimization prevention, safe accommodation, trauma-responsive services, and accessible reporting mechanisms are therefore central to responsible justice policy. Mental health and safety services should be attentive to the person's protection and rights as well as their clinical needs.
Mental health problems in US jails and prisons
People entering correctional facilities may already have significant mental health needs. Confinement can also affect psychological functioning, while the screening systems, demographics, and service access of incarcerated populations differ from the general community. Consequently, a high prevalence of distress in custody is evidence of service need, not proof of psychiatric causation of prior offenses.
The US Bureau of Justice Statistics reported that 14% of state and federal prisoners and 26% of jail inmates met a serious psychological distress threshold in the 2011–2012 National Inmate Survey, compared with a standardized general-population figure of 5%. Serious psychological distress is a survey measure, not equivalent to a diagnosis, and these figures describe the survey period rather than a new count for 2026.
The pathway to incarceration can include policing, charges, bail decisions, access to lawyers, sentencing, support networks, and treatment availability. Correctional health indicators alone cannot identify which factor brought an individual into custody.
Can mental illness predict future crime?
A forecast requires a specified behavior, target population, time horizon, and decision context. Risk assessments typically combine historical information, current circumstances, relevant clinical factors, and protective conditions rather than relying on diagnosis alone. A result is an estimated risk under assumptions, not certainty and not a legal finding.
A 2023 systematic review and meta-analysis of violence-risk instruments in people discharged from forensic psychiatric hospitals included 50 publications, 36 tools, 10,460 participants, and 12 countries. For selected instruments, pooled discrimination estimates for violent reoffending ranged roughly from an area under the curve of 0.64 to 0.72. Almost all studies were assessed as having high risk of bias, while calibration and other important performance information were often missing. The evidence base was limited for women and did not establish generalizability to lower-income countries.
Discrimination and calibration
Discrimination indicates how well scores rank individuals who later experience an outcome relative to those who do not. Calibration asks whether a predicted probability corresponds to an observed frequency in a suitable validation population. A tool can rank some outcomes better than chance yet provide poorly calibrated probabilities in a different setting.
False positives, false negatives, and base rates
A false positive is a prediction of an outcome that does not occur; a false negative is an outcome that occurs after a low-risk prediction. The rates depend on the cutoff, event prevalence, and assessment population. For rare outcomes, false positives can be numerous even when discrimination seems encouraging. Decision-makers must understand the harms of both error types before using a tool.
Population limits, culture, and fairness
Tools developed from forensic inpatients, previous offenders, or predominantly male cohorts cannot automatically represent women, young people, clinical outpatients, or people from different cultural and legal systems. Fairness requires attention to measurement, historical disadvantage, local calibration, and subgroup performance. Overall predictive accuracy alone does not establish equitable use.
A protective factor, such as stable housing or supportive care, can inform risk-management planning. It is still a conditional correlate in a specific research context. Neither risk nor protection scores identify a criminal act as proven.
Mental illness and the law: responsibility, competence, and verdicts
Psychology investigates mental functioning, behavior, cognition, and evidence. Criminal law specifies offenses, required mental elements, defenses, procedures, and standards of proof. Diagnosis and legal responsibility meet in some cases, but a diagnostic category does not decide criminal culpability.
The insanity defense concerns the law governing responsibility at the relevant time of an alleged act. Competency to stand trial concerns a defendant's present ability to participate in legal proceedings under the applicable legal test. These concepts differ from ordinary clinical functioning and from prediction of future behavior. In Kahler v. Kansas (2020), the US Supreme Court considered state variation in approaches to insanity defenses, illustrating that no single national clinical definition supplies the legal answer.
A forensic psychologist may provide a qualified opinion about mental state, functioning, or assessment findings. The court or other authorized tribunal decides the legal question. Forensic reports should keep behavioral observations, symptoms, test results, diagnostic conclusions, expert inferences, and legal determinations separate, as recommended in the APA Specialty Guidelines.
Nonviolent offending, arrests, and public-order enforcement
Nonviolent crime includes distinct offense classes, such as property offenses, drug-law violations, and certain public-order offenses. Local laws and enforcement priorities influence how frequently conduct is recorded. Some encounters with law enforcement occur during crises or in public spaces when health and housing resources are scarce. An arrest in such circumstances should not automatically be interpreted as direct evidence of psychiatric symptoms causing criminal behavior.
A claim that mental illness predicts “crime” needs to specify whether the research measured charges, convictions, actual offending, or police contacts. When studies use composite outcomes, conclusions should preserve the components rather than creating a universal criminal-risk profile.
Recidivism and justice-involved populations
Recidivism is a defined form of new justice-system contact after an earlier qualifying event. Studies may count rearrest, reconviction, reincarceration, or a specific violent offense. These definitions and follow-up periods matter. The characteristics of people already convicted are not representative of everyone with a mental health diagnosis.
A Swedish cohort study examined diagnosed psychiatric disorders and violent reoffending among previously convicted prisoners. It is informative for corrections and the design of follow-up studies, but it cannot be used to forecast offending by someone who has never entered the justice system. Forensic recidivism assessment also requires attention to the base rate and to interventions and circumstances that can change after release.
Does treatment or diversion reduce offending?
Mental health services can provide clinical benefits regardless of their measurable effect on recidivism. Public safety programs should also be evaluated for court outcomes, repeat police contact, victimization, housing stability, treatment access, and human rights. The goal and the outcome measure must be explicit.
A meta-analysis of mental health courts combining 17 studies involving 16,129 participants found a small average reduction in recidivism relative to traditional criminal processing, with substantial heterogeneity and quality-related limitations. A 2024 systematic review and meta-analysis reported favorable results across 15 studies; study designs and outcomes require careful interpretation before treating the results as universal effects.
Police-based prebooking diversion uses crisis response and referral to services before conventional criminal processing. A systematic review by Dewa and colleagues found limited and methodologically weak evidence that such programs reduce arrests, alongside more encouraging evidence for increased linkage to mental health services. Outcomes vary by local services and implementation.
Effective policy combines proper clinical care with fair legal processes. Treatment, housing, substance-use support, safe community integration, and proportionate justice decisions each address different contributors to well-being and public safety.
How to evaluate alarming claims about mental illness and crime
Start with the denominator. Ask what proportion of people in the studied group actually experienced the recorded outcome and over what period. Check whether a headline substitutes a relative risk ratio for an absolute probability, or a rare offense for all crime. Identify whether the outcome was self-reported behavior, arrest, conviction, or hospitalization.
Next examine the sample. A comparison of forensic hospital patients with a national population differs from a well-matched community comparison. Determine whether substance use, previous violence, age, neighborhood context, and economic circumstances were measured and appropriately considered.
Finally, examine what the evidence establishes. An association does not identify a necessary or sufficient cause of an offense. A risk instrument does not establish guilt. Clinical documentation does not settle a legal standard, and media speculation cannot substitute for direct professional assessment.
Myths about mental illness and crime
“People with mental illness are generally dangerous.”
The statement ignores heterogeneity and base rates. Most people living with mental illness are not violent offenders. Some specific populations show elevated statistical associations with defined violent outcomes; these are conditioned by substance use, history, circumstances, and study design.
“A psychiatric diagnosis explains why a crime happened.”
Event-specific explanation requires evidence of the person's behavior, clinical state at the relevant time, possible motives, situational context, and alternative mechanisms. Diagnostic status alone cannot supply that explanation.
“Risk tools can identify future criminals.”
Even structured instruments have false positives and false negatives. Their predictive properties are conditional on population, outcome, follow-up period, and validation quality. A risk score is neither a verdict nor a test of moral character.
“Treatment alone eliminates criminal risk.”
Treatment can address psychiatric symptoms and improve functioning, but reducing offenses may require attention to other needs and circumstances. Program impacts have to be established through appropriate comparisons.
“Mental illness and crime means offending, not victimization.”
People with psychiatric disorders are also subjected to crime, with elevated victimization rates reported in population data. Good safety and justice policy must account for that evidence. Dean and colleagues (2018).
Practical implications
Clinicians should assess the specific person's needs and relevant history while maintaining confidentiality and avoiding stereotype-driven assumptions. Event-specific threats and behavior require attention in their own right. Mental health support must also address victimization, exploitation, and access to safe environments.
Families and communities can respond more effectively to concrete warning behaviors and acute needs than to diagnostic labels. For journalists, responsible reporting means distinguishing verified clinical findings from rumor and ensuring that exceptional criminal events are not portrayed as typical of a psychiatric population.
For justice professionals, relevant psycholegal standards, validated assessment practice, independent fact-finding, due process, and procedural fairness must remain distinct. Policymakers should compare outcomes of community services, diversion programs, correctional care, and reentry support without equating clinical need with criminal propensity.
Frequently asked questions
Are most crimes committed by people with mental illness?
There is no defensible universal percentage. Studies use different definitions of mental illness and different measures of crime. The evidence does not support using psychiatric diagnosis as a general explanation of offending. Population-attributable estimates in selected psychosis studies cannot be generalized to all disorders or all crime. Fazel and colleagues (2009).
Are people with schizophrenia likely to commit violence?
Most do not. Relative rates of certain violent outcomes are higher in some cohorts, especially with substance-related comorbidity. A diagnosis cannot predict what any particular person will do. Fazel and colleagues (2009).
Does depression increase criminal behavior?
A population study observed associations with recorded violent offenses, but clinical presentation, context, confounding, and the outcome definition matter. A depression diagnosis is not an individualized prediction or explanation. Fazel and colleagues (2015).
Is bipolar disorder linked to crime?
Some research finds an association with violent convictions that is strongly influenced by substance use and shared factors. The diagnosis alone does not determine legal responsibility or individual risk. Fazel and colleagues (2010).
Are people with mental illness also at greater risk of being victims?
Large registry research has found elevated rates of recorded crime victimization in several psychiatric groups, including violent victimization. Estimates differ by population and circumstances. Dean and colleagues (2018).
Why do US prisoners have high levels of psychological distress?
Correctional populations have significant health needs, and incarceration interacts with social disadvantage and access to care. Bureau of Justice Statistics survey findings describe distress in 2011–2012 custody populations; they do not demonstrate that psychiatric diagnoses caused the prior offenses.
Does mental illness mean someone is legally insane?
No. Clinical diagnosis and criminal responsibility are distinct. US legal defenses vary between jurisdictions; the Kahler decision illustrates that legal diversity.
Can a psychologist predict violent crime?
Structured approaches can sometimes provide useful discrimination for defined outcomes, but accuracy is limited, errors occur, and calibration often lacks adequate evidence. The 2023 meta-analysis emphasizes major limitations in forensic risk-tool studies.
Can mental health courts reduce recidivism?
Meta-analyses report average benefits in some comparisons, but effects vary with study quality, court design, participant selection, and outcome definition. 2017 meta-analysis; 2024 review.
Conclusion
The evidence on mental illness and crime points to heterogeneous, context-dependent associations rather than a diagnostic explanation of criminality. Some measured differences in violent offending exist in particular cohorts, and substance use, previous violence, environmental adversity, and selection into justice systems are important. Most people with psychiatric conditions are not violent offenders, while victimization is also a major, sometimes neglected, dimension of the relationship.
A useful forensic framework keeps clinical diagnosis, observed conduct, screening or psychometric results, risk and protective factors, expert opinions, and legal findings in their proper places. Public safety and fairness both benefit from precise evidence, accessible treatment, credible victim protection, and justice-system interventions evaluated for their real outcomes.
Related articles
Schizophrenia and Violence: What Research Really Shows provides the dedicated clinical evidence review of psychotic disorders and violence.
Related Articles
References
American Psychological Association. (2013). Specialty guidelines for forensic psychology. American Psychologist, 68(1), 7–19. DOI.
Bronson, J., & Berzofsky, M. (2017). Indicators of mental health problems reported by prisoners and jail inmates, 2011–2012. Bureau of Justice Statistics. Official report.
Chang, Z., Larsson, H., Lichtenstein, P., & Fazel, S. (2015). Psychiatric disorders and violent reoffending: A national cohort study of convicted prisoners in Sweden. The Lancet Psychiatry, 2, 891–900. PubMed.
Dean, K., Laursen, T. M., Pedersen, C. B., Webb, R. T., Mortensen, P. B., & Agerbo, E. (2018). Risk of being subjected to crime, including violent crime, after onset of mental illness. JAMA Psychiatry, 75, 689–696. DOI.
Dewa, C. S., Loong, D., Trujillo, A., & Bonato, S. (2018). Evidence for the effectiveness of police-based pre-booking diversion programs in decriminalizing mental illness. PLOS ONE, 13, e0199368. DOI.
Elbogen, E. B., & Johnson, S. C. (2009). The intricate link between violence and mental disorder. Archives of General Psychiatry, 66, 152–161. DOI.
Fazel, S., Gulati, G., Linsell, L., Geddes, J. R., & Grann, M. (2009). Schizophrenia and violence: Systematic review and meta-analysis. PLOS Medicine, 6, e1000120. DOI.
Fazel, S., Långström, N., Hjern, A., Grann, M., & Lichtenstein, P. (2009). Schizophrenia, substance abuse, and violent crime. JAMA, 301, 2016–2023. DOI.
Fazel, S., Lichtenstein, P., Grann, M., Goodwin, G. M., & Långström, N. (2010). Bipolar disorder and violent crime. Archives of General Psychiatry, 67, 931–938. PubMed.
Fazel, S., Wolf, A., Chang, Z., Larsson, H., Goodwin, G. M., & Lichtenstein, P. (2015). Depression and violence: A Swedish population study. The Lancet Psychiatry, 2, 224–232. PubMed.
Jalain, C. I., Lucas, P. A., & Higgins, G. E. (2024). Assessing the effectiveness of mental health courts in reducing recidivism: A systematic review with meta-analysis. Justice Evaluation Journal, 7, 212–230. DOI.
Kahler v. Kansas, 589 U.S. 271 (2020). US Supreme Court opinion.
Lowder, E. M., Desmarais, S. L., & Baucom, D. J. (2017). Effectiveness of mental health courts in reducing recidivism: A meta-analysis. Psychiatric Services. PubMed.
Ogonah, M. G. T., Seyedsalehi, A., Whiting, D., & Fazel, S. (2023). Violence risk assessment instruments in forensic psychiatric populations: A systematic review and meta-analysis. The Lancet Psychiatry, 10, 780–789. PubMed.
Steadman, H. J., Mulvey, E. P., Monahan, J., Robbins, P. C., Appelbaum, P. S., Grisso, T., Roth, L. H., & Silver, E. (1998). Violence by people discharged from acute psychiatric inpatient facilities and by others in the same neighborhoods. Archives of General Psychiatry, 55, 393–401. PubMed.