23 May 2009. Not many studies look at the whole population of a nation, but a study that did so in Sweden clarified the likelihood that people with schizophrenia will commit violence. In the May 20 JAMA, Seena Fazel of the University of Oxford, Warneford Hospital in Oxford, England, and colleagues offer evidence that substance abuse spikes the risk of violence in people with schizophrenia. Indeed, they conclude that, in the absence of substance abuse, schizophrenia raises the risk of violence only a little. The researchers further find that family background confounds the relationship between violence and schizophrenia, based on comparisons of siblings with and without the illness.
To put the findings into context, SRF spoke with two researchers not involved in the study, Jeffrey Swanson of the Duke University School of Medicine, in Durham, North Carolina, and E. Fuller Torrey of the Stanley Medical Research Institute in Chevy Chase, Maryland. Both lauded the new study as important for advancing the understanding of violence in schizophrenia despite its lack of novel findings. They found the study compelling due to its large number of subjects and its avoidance of the selection bias that plagues studies done in places like the United States.
Everything under the Swedish sun
The study’s strengths arise from its use of several nationwide Swedish registries, which the researchers linked. These included the 1970 and 1990 National Census, as well as registries of hospital discharges, crimes, and family ties. Using these databases, the researchers identified subjects who had been hospitalized with schizophrenia or an alcohol or drug abuse disorder. For added confidence in the schizophrenia diagnosis, they only counted as cases subjects with at least two hospitalizations for the illness.
As an outcome measure, the researchers tallied violent crimes such as homicide, assault, sexual offenses, illegal threats, intimidation, robbery, and arson. They were able to rely on conviction data to measure crime because, under Swedish law, people can be found guilty even if they are mentally ill and sent for treatment.
According to Swanson, “When you ask the question, ‘Is someone more violent?’ the first thing you have to ask is, ‘more violent than whom?’” The study answered that question two ways. The first set of analyses compared the 8003 subjects who had been hospitalized for schizophrenia with 80,025 age- and sex-matched control subjects from the general population.
The second set of analyses used a different approach: It compared a subset of the subjects with schizophrenia to their unaffected full siblings, who would presumably share heritable and environmental influences. These analyses included 4674 subjects with schizophrenia and 7,780 siblings.
While 5 percent of those in the general population had been convicted of at least one violent crime, 13 percent of subjects with schizophrenia had such a record, yielding an adjusted odds ratio of 2.0 (95 percent CI = 1.8-2.2). “What was surprising is the fact that these results are as similar as they are to the pattern found in the ECA study (Swanson et al., 1990) and the NIMH CATIE study (Swanson et al., 2008) in patients with schizophrenia,” Swanson said, referring to two studies sponsored by the National Institute of Mental Health: the Epidemiological Catchment Area Study and the Clinical Antipsychotic Trials of Treatment Effectiveness.
Even so, what looked like one picture became two when the researchers focused on the role of substance abuse. A whopping 28 percent of subjects who had both schizophrenia and substance abuse disorder, versus 8 percent of those with schizophrenia only, had racked up a violent record. In analyses that used the general public as the control group, Fazel and colleagues found that the dual diagnoses quadrupled the odds of engaging in violence (adjusted odds ratio = 4.4, 95 percent confidence interval = 3.9-5.0).
In contrast, schizophrenia alone raised the odds of violence relatively little (odds ratio = 1.2, 95 percent confidence interval = 1.1-1.4). Based on these findings, the researchers write, “The association between schizophrenia and violent crime is minimal unless the patient is also diagnosed as having substance abuse comorbidity.”
While substance abuse seems to mediate the relationship between schizophrenia and violence, familial factors may also play a role. In the sibling analyses, 28 percent of subjects with schizophrenia coupled with substance abuse and 18 percent of their unaffected siblings had engaged in violent crime (adjusted odds ratio = 1.8, 95 percent confidence interval = 1.4-2.4). However, in the absence of substance abuse disorder, this gap narrowed, with only 7 percent of those with schizophrenia and 5 percent of their siblings showing a violent past (adjusted odds ratio = 1.3, 95 percent confidence interval = 1.0-1.4).
“This is provocative and intriguing evidence that violence probably is multi-determined,” Swanson said, although it raises the question, “What is it about the family?” Unfortunately, the researchers lacked the kind of in-depth data that would have allowed them to tease apart heritable and environmental influences.
Finding the right treatment regimen
Torrey, who called the study “marvelous,” nonetheless wished that it had included treatment data. “That’s a critical issue because, in fact, we know that people with substance abuse and schizophrenia are considerably less likely to take their medication than people who do not have substance abuse,” he said. Furthermore, treatment seems to curb violent behavior in people with schizophrenia.
“Regardless of the nature of the mechanism, adequate substance abuse treatment for individuals with schizophrenia is likely to reduce the risk of violence and should be part of the routine assessment and management of all such patients,” write Fazel and colleagues. Torrey quibbled with that conclusion by stressing the need to adequately treat the schizophrenia, too: “In fact, the treatment of substance abuse in people who are not being treated for their schizophrenia is a dismal failure.”
Any increase in violence takes a tragic toll on human beings, but Swanson said, that contrary to perceptions of the general public, most patients with schizophrenia do not resort to violence. “If someone with schizophrenia commits a violent act, the immediate assumption is that the reason they did it is because of their disease,” he said. However, they may behave violently for the same reasons as anyone else, and those reasons could include drug addiction or family background.—Victoria L. Wilcox.
Fazel S, Långström N, Hjern A, Grann M, Lichtenstein P. Schizophrenia, substance abuse, and violent crime. JAMA. 2009, May 20; 301(19):2016-2023. Abstract
Q&A With Sheilagh Hodgins. Questions by Victoria L. Wilcox.
Q: Fazel and colleagues cite your 2001 paper (Hodgins, 2001) as an example of calls to move beyond further study of violence in the mentally ill to focus on treatment. Have your views changed or stayed the same since you wrote that paper?
A: No, my views have not changed. We know that schizophrenia increases the risk of aggressive behavior and crime; the evidence is robust. We need to move to identifying strategies that effectively reduce these behaviors.
Do we know enough about the causes to focus on treatment? We certainly know enough to begin testing different management strategies and treatments. For example, we are currently comparing aggressive behavior among patients with severe mental illness who are cared for by community teams, one of which has been trained to use the Historical, Clinical, and Risk Management-20 (Webster et al., 1997) to assess risk and to identify factors that need to be changed in order to reduce aggressive behavior, while the other community team is proceeding as usual. We and others have proposed testing in patients with severe mental illness the effectiveness of learning-based programs that have been shown to be effective in reducing offending among non-mentally ill offenders. The results of such trials are beginning to be published (see, for example, Ashford et al., 2008). Notably, we have also shown that outcomes (symptoms, aggressive behavior, substance misuse) among men with schizophrenia treated by forensic services are much more positive than among those treated by general adult psychiatric services (Hodgins et al., 2007a).
The evidence suggests, however, that patients with schizophrenia who engage in aggressive behavior and/or criminality constitute a heterogeneous population with respect to treatment needs. These subtypes of patients (Hodgins, 2008) present different needs for treatment, and we hypothesize, distinctive etiologies.
Q: In a 2008 paper in Philosophical Transactions of the Royal Society (Hodgins, 2008), you described the evidence that people with schizophrenia are at increased risk of being convicted for committing violent acts as "robust." With that in mind, does the JAMA study add anything new to what was already known about violence and schizophrenia? If yes, what? If no, why not?
A: The JAMA paper confirms previous findings but fails to examine sex differences. This is important because the elevation in risk for violence with severe mental illness in previous studies has been shown to be much greater among women than men (see, for instance, Brennan et al., 2000). The paper states that this sample is larger than all other combined samples that have been studied, which is inaccurate. The paper extends knowledge by showing that the rates of violent crime and substance misuse among individuals with schizophrenia and their unaffected siblings are similar. This is important, but I do not agree with the authors' interpretation of their findings.
The paper failed to examine childhood conduct problems. Yet, it is known that childhood conduct problems are antecedents of schizophrenia in a minority of cases (between 20 percent and 40 percent) and that the conduct problems prior to age 15 are associated with violent crime and aggressive behaviors that do not lead to prosecution through middle age. It is also known that conduct problems that start in childhood and that remain stable across the lifespan are hereditary, and that this pattern of stable antisocial behavior that is inherited includes a vulnerability for substance misuse (Krueger et al., 2002; Moffitt, 2005). Therefore, while reducing substance abuse among patients with schizophrenia may have a positive impact on their general psychosocial functioning and symptoms, among those with a history of conduct problems in childhood, such interventions would not likely reduce aggressive behaviors. Patients with a childhood history of conduct problems, we have hypothesized, require learning-based treatments that teach them not to behave aggressively and teach them alternate behaviors for use when they are angry, frustrated, or afraid. By contrast, among those whose violent behaviors begin as the illness does, reducing substance misuse may have a more dramatic effect on reducing violent behaviors.
Finally, the evidence presented in the JAMA paper does not reveal the reasons for the observed association between the behavior of the patients and their unaffected siblings. It simply shows that individuals who grow up in the same families show similar behaviors; it does not address whether that similarity results from environmental or genetic factors or combinations of both.
Q: In the same 2008 paper mentioned above, you wrote that violent offenders with schizophrenia should be viewed as heterogeneous, and you described three subtypes of offenders. How do Fazel's findings compare or contrast with your thinking in this regard?
A: The JAMA paper does not separate out the offenders with schizophrenia who had exhibited aggressive and antisocial behaviors since childhood. By lumping all the patients together (even males and females, those with offences prior to illness onset and those with offences long after), the paper failed to provide information that could be relevant to improving treatments and services to reduce violence in this population.
Q: What does the new study imply about how to prevent violent acts that might be committed by people with schizophrenia? Do the findings suggest any change in current treatment approaches?
A: The paper again emphasizes the need for general (civil) psychiatric services to assess and manage the risk of violence among patients with schizophrenia and to use treatments that have been proven to be effective to reduce substance misuse in this population. The paper also confirms previous findings that the families of some patients with schizophrenia may not be helpful resources for the patients as they themselves may be engaging in criminality and substance misuse.
Q: What do you think of the methods used by the researchers?
A: The analyses are done well. My criticisms would be with the reasoning behind some of the analyses and with the interpretation of the findings.
Q: You have written that stigmatization and lack of resources present barriers to preventing and treating individuals with mental illness who might be prone to violence. Do you think the results by Fazel and colleagues will help or hurt in that regard, or is it likely to have no effect at all?
A: People with schizophrenia who engage in violence not only hurt their victims; they themselves suffer. In some countries, they are incarcerated in prisons, where they often are abused. Several studies have shown that people with schizophrenia are more likely than the general population to be victims of crime and particularly to be victims of physical assaults (Hodgins, et al., 2007b; Maniglio, 2009). Their own aggressive behavior is the strongest predictor that we have found of physical victimization (Hodgins et al, 2007b). If we successfully reduced violent behavior by persons with schizophrenia, I think that we would reduce stigmatization and increase public acceptance of adequate funding for community treatment programs.
Q: Is there anything else you would like to say about this paper or the wider topic of violence and schizophrenia?
A: I am currently preparing a special issue of the journal European Psychiatry with papers from several countries showing high levels of aggressive behavior and/or crime among patients with schizophrenia being cared for by general (civil) psychiatric services in several different countries. The evidence that schizophrenia increases the risk of aggressive behavior and victimization continues to accumulate, and it is urgent that general psychiatric services begin to assess and manage these problems. Forensic mental health professionals have expertise that could be shared with their colleagues in general psychiatric services to improve patient care and outcomes.