The relationship between the rate of homicide by those with schizophrenia and the overall homicide rate: A systematic review and meta-analysis
Introduction
Schipkowensky (1973) proposed that rates of homicide by the mentally ill are similar across countries and are unrelated to total homicide rates. Coid (1983) reached a similar conclusion after reviewing some of the available evidence, and proposed the existence of an ‘epidemiological law’, that “the higher the rate of homicide in a population, the lower the percentage of offenders who are found to be mentally abnormal” (page 857). The belief that rates of homicides by the mentally ill are fixed and are unrelated to total homicide rates has gained wide acceptance (Shaw, 1999, Taylor and Gunn, 1999, Simpson et al., 2004, Coid et al., 2006) and has been extended to rates of non-lethal violence by the mentally ill (Appelbaum, 2006, Buchanan, 2008). Coid's (1983) paper did not report the criteria used to exclude some studies, lacked a statistical analysis to support its findings (Large et al., 2009, Coid, 2009a) and was confounded by heterogeneous definitions of mental illness (Coid, 2009b). However, no subsequent study has examined the relationship between rates of homicide by the mentally ill and all homicides, despite the implications that these findings might have for homicide prevention strategies.
Constant rates of homicide by the mentally ill between regions and over time could be regarded as evidence that most of these homicides are due to aspects of the illness itself. This would suggest that measures to improve treatment, including the containment of some patients, would reduce this form of homicide. However, a contrary finding that the rates by the mentally ill are associated with overall homicide rates raises other possibilities, for example, that the mentally ill are particularly vulnerable to the social factors that have been shown to be associated with the incidence of other homicides, such as social disadvantage and substance abuse.
The aim of this study was to test the null hypothesis that the proportion of homicides by people with schizophrenia is not associated with the rate of other homicides.
Section snippets
Methods
We examined studies reporting the number of homicides by persons diagnosed with schizophrenia and the total number of homicides in defined populations. We only considered studies from developed countries because the judicial procedures and official statistics were more likely to be reliable in those countries. Because legal definitions of mental illness vary between jurisdictions, we excluded studies that did not report a psychiatric diagnosis and instead reported the legal verdict that the
Results
The search results are shown in Fig. 1. We found 25 studies conducted in high-income countries which reported rates of homicide by the seriously mentally ill, of which 18 reported the number of offenders diagnosed with schizophrenia (Table 1).
Although we excluded studies describing subjects that had already been reported in other papers, two Danish studies reporting homicides from overlapping periods were included (Gottlieb et al., 1987, Hart Hansen, 1977), because the period of overlap was
Discussion
We found that the number of homicides per capita committed by those diagnosed with schizophrenia was strongly associated with the rates of other homicides. We also found no evidence that the proportion of homicides by those with schizophrenia is lower in places with higher homicide rates. This finding is the opposite of the conclusion of Coid (1983) and challenges the widely held belief that rates of homicide and serious violence by the mentally ill are constant and are associated with the
Conclusion
This systematic review and meta-analysis found that the rates of homicide by those diagnosed with schizophrenia are correlated with the total homicides rates. It found that the proportion of homicides by patients with schizophrenia is not associated with the total homicide rate and casts doubt on the existence of an epidemiological law to this effect. This finding suggests that measures to prevent homicide by those diagnosed with schizophrenia should include not only an attempt to provide
Role of funding source
The study was not funded.
Contributors
Dr Large formulated the hypothesis, performed the searches, extracted and analysed the data and participated in the writing of the initial and final drafts.
Dr Smith extracted and analysed the data and participated in the writing of the initial and final drafts.
Dr Nielssen participated in formulating the study, performed the searches, and participated in the writing of the initial and final drafts.
Conflict of interest
We have no conflict of interest.
Acknowledgments
We would like to thank E. Fuller Torrey MD, for his helpful comments on an early draft of the manuscript, and Dr Peter Arnold, for his assistance in its preparation for publication.
References (60)
Schizophrenia and violence
Acta Psychiatr. Scand., Suppl.
(2000)Violence and mental disorders: data and public policy
Am. J. Psychiatry
(2006)- et al.
Five Year Report of the National Confidential Inquiry into Suicide and Homicide by People with Mental Illness
(2006) - et al.
Mental disorders and violence in a total birth cohort: results from the Dunedin Study
Arch. Gen. Psychiatry
(2000) Homicide in Western Nigeria
Br. J. Psychiatry
(1969)- et al.
Developing world perspective on homicide and mental disorder: an eighteen-year retrospective study (1980–1998) at Jos, Nigeria
Int. J. Forensic Ment. Health
(2004) - et al.
The prediction of criminal and violent recidivism among mentally disordered offenders
Psychol. Bull.
(1998) - et al.
Comprehensive Meta-Analysis Version 2
(2005) Risk of violence by psychiatric patients: beyond the “actuarial versus clinical” assessment debate
Psychiatr. Serv.
(2008)- et al.
Psychiatric comorbidities and schizophrenia
Schizophr. Bull.
(2009)