Lately, I've been busy wading through a heavily publicized study that was published this month in the Lancet. In their paper, "Antipsychotics, mood stabilisers and risk of violent crime," Dr. Seena Fazel and his associates linked Swedish national registers to compare rates of violent crime among 82,647 male and female psychiatric patients to assess the effect of medication on this outcome.
The study made quite a splash in the news, because the outcome was almost too good to be true. There was a 64% reduction in violent crime among patients who had been prescribed any antipsychotic or mood stabilizer, compared with those taking other psychotropics. The reduction in violence for those taking neuroleptics and mood stabilizers was 45% and 24%, respectively. Selective serotonin reuptake inhibitors (SSRIs) had no apparent effect on crime (Lancet 2014 [doi:10.1016/S0140-6736(14)60379-2]).
Given our American anxiety over spree shooters and other high-profile crimes allegedly committed by untreated psychiatric patients, this study clearly deserves some scrutiny to thoroughly understand the findings, limitations, and other factors that could limit generalizability to the United States.
The authors compared mental health treatment registries with the national criminal history database. They looked at the rate and types of crimes committed by psychiatric patients when they were in and out of treatment. The "in-treatment" time interval was defined as the time between two or more prescriptions, as long as the prescriptions were no more than 4 months apart. Individuals who had only been given one script [prescription] were excluded. The outcome measure was any criminal conviction. The conviction outcome was based upon the date the offense took place, not the date of conviction. Individuals were excluded if the offense date was unknown.
A within-individual analysis showed significant reduction in all crimes, including violent crime, drug-related crime, and less severe crimes, during times when patients were prescribed medication, compared with medication-free intervals. When medicated, the rate of violent crime did not differ between patients with and without a history of violent offenses when diagnosis was not considered. When the analysis was limited to people with schizophrenia, bipolar disorder, or other psychotic disorders, the prescription of neuroleptics significantly reduced violent crime for both men and women.
For bipolar disorder, mood stabilizing medication reduced violent crime for men but not for women. The SSRI-medicated group was used as a control, to account for the general effect of contact with the mental health system and non-medication interventions related to this, and there was no effect on violent crime with this class of medication.
Now on to the limitations. Medication adherence was not assessed and could not be verified apart from patients given depot neuroleptics. The overall rate of violent crime was low, as would be expected. Only 6% of men and 1% of the women committed a violent crime. The numbers were so low that the study could not statistically assess the impact of violent crime history among patients diagnosed with psychosis. This is a small but crucial finding that did not make the traditional media coverage of this study.
Also, only 40% of those patients taking antipsychotics and mood stabilizers had a diagnosis of schizophrenia, other psychosis, or bipolar disorder, suggesting that, in Sweden, these medications might be prescribed for other indications such as characterologic low frustration tolerance or irritability. The analysis did not look at impact on violent crime by personality disorder diagnosis.
The authors acknowledged that their research could not prove a causal link between psychiatric illness and violence, another important conclusion that was not mentioned in traditional media coverage. In Sweden, mental illness cannot be used to prevent or mitigate a criminal conviction, so any connection between psychiatric symptoms and crime in this population can't be determined. The study also did not consider which subjects, if any, were taking medication or in treatment under court-mandated conditions.
As legislators and advocacy groups push to strengthen involuntary treatment laws, there is a risk that "bottom line" media coverage of research like this may inappropriately sway public opinion. Psychiatrists should be prepared to respond to proposed policies based on inaccurate interpretation of research.
Dr. Hanson is a forensic psychiatrist and coauthor of "Shrink Rap: Three Psychiatrists Explain Their Work" (Baltimore: The Johns Hopkins University Press, 2011). The opinions expressed are those of the author only, and do not represent those of any of Dr. Hanson's employers or consultees, including the Maryland Department of Health and Mental Hygiene or the Maryland Division of Correction.