Is Arson the Crime Most Strongly Associated With Psychosis?—A National Case-Control Study of Arson Risk in Schizophrenia and Other Psychoses Sophia Anwar 3 , Niklas La ˚ngstro ¨m 2 , Martin Grann 2 , and Seena Fazel 1,3 2 Centre for Violence Prevention, Karolinska Institute, Stockholm, Sweden; 3 Department of Psychiatry, University of Oxford, Warneford Hospital, Oxford, UK Background: The association of psychosis with certain se- rious crimes, such as homicide, has been clearly demon- strated, but it is uncertain to what extent psychotic disorders are associated with arson. Methods: We used a case-control design to investigate the association of being diagnosed with schizophrenia and other psychoses and com- mitting arson. Data were obtained from Swedish national registers for criminal convictions, hospital discharge diag- noses (International Classification of Diseases, Ninth Re- vision [ICD-9], and International Classification of Diseases, Tenth Revision [ICD-10]), and sociodemographic factors for 1988–2000. We included all convicted arson offenders of both sexes in Sweden (N 5 1689) and com- pared them with a random sample of general population control subjects (N 5 40 560). Results: After adjustment for sociodemographic confounders, arson offenders were more likely to be diagnosed with schizophrenia (in men, adjusted odds ratio [OR] 5 22.6, 95% confidence interval [CI] 5 14.8–34.4; in women, adjusted OR 5 38.7, 95% CI 5 20.4–73.5) or other psychoses (in men, adjusted OR 5 17.4, 95% CI 5 11.1–27.5; in women, adjusted OR 5 30.8, 95% CI 5 18.8–50.6). Conclusions: Individ- uals with schizophrenia and other psychoses have signifi- cantly increased risks of an arson conviction. These risk estimates are higher than those reported for other violent crimes and place arson in the same category as homicide as crimes that are most strongly associated with psychotic disorders. Key words: schizophrenia/violence/crime/fire-setting behavior/case-control studies Introduction Considerable attention has been directed to the associa- tion between homicide and schizophrenia, and it is esti- mated that the risk of homicide in individuals with psychosis is 10–20 times higher than in the general pop- ulation. 1–3 Supporting this strong association, a recent systematic review of 10 studies demonstrated that treat- ment could reduce homicide rates in first-episode psycho- sis by 15 times. 4 Risk increases for other crimes, including property, 5 drug, 5 sexual, 6 and other violent offences, 7,8 have also been demonstrated, but the associations are not as strong as they are for homicide. 9 Clinical studies have identified high rates of arsonists in secure hospitals, and one study found high rates of schizophrenia in a se- lected sample of severe cases of arson 10 ; however, the as- sociation of psychotic disorders with arson remains uncertain. Around 10% of all individuals admitted to forensic psychiatric services in United Kingdom, Sweden, and Finland have been convicted for arson. 11–13 These rates appear to be higher for women in high-security hos- pitals; 36% of women admitted to English high-security hospitals had committed property offences, mostly arson. 14 In contrast, studies comparing prisoners and fo- rensic patients that have used other offenders as control subjects did not find associations between arson and severe mental illness, 15,16 nor did a population-based study, 17 although the latter may have been underpowered. Although its effects on psychiatric services are high, the wider impact of arson is notable. In the United States, there were a reported 64 000 offences in 2007, corre- sponding to 25 offences per 100 000 inhabitants. 18 In the United Kingdom, the conviction rate for arson was higher at 65 per 100 000 persons 19 and was estimated to cost £2.8 billion annually, with around 40 000 offences recorded by the police, and an estimated 3600 fires started deliberately every week in 2007. 20 In Australia, arson is estimated to cost more than $1 billion per year, and recent widespread fires are thought to have been started deliberately. 21 Thus, if risk factors for arson were modifiable, this could make a significant impact on public health and safety. As the association of arson with schizophrenia and other psychoses is uncertain, and few population- based studies have been conducted, we carried out 1 To whom correspondence should be addressed; fax: þ44-1865- 793101; e-mail: seena.fazel@psych.ox.ac.uk Schizophrenia Bulletin vol. 37 no. 3 pp. 580–586, 2011 doi:10.1093/schbul/sbp098 Advance Access publication on October 22, 2009 Ó The Author 2009. Published by Oxford University Press on behalf of the Maryland Psychiatric Research Center. All rights reserved. For permissions, please email: journals.permissions@oxfordjournals.org. 580 at Karolinska Institutet University Library on May 10, 2013 http://schizophreniabulletin.oxfordjournals.org/ Downloaded from