The link between violent acts and severe mental disorder has been proposed for a long time (Lindqvist & Allebeck, 1990). Quantitative methods have largely confirmed this relationship. For instance, most studies show that violent crime is around 2-10 times more common among individuals diagnosed with schizophrenia than compared populations (Fazel et al, 2009).

Moreover, a recently published study also indicates that individuals with schizophrenia were much more likely to be convicted of a violent crime in comparison to those who were not.

So…what are we to make of all this?

Violence is common. There were 1.3 million incidents of violence reported in England and Wales between July 2013 and June 2014 according to the Crime Survey of England and Wales (ONS, 2013).

Schizophrenia, although rare, with a median incidence of just around 15 cases per hundred thousand of the population per year, is as yet incurable and persists in people, resulting in a relatively high prevalence (McGrath et al, 2008).

There has been a consensus that violence risk is increased in schizophrenia. Much of this research was summarised in a meta-analysis in 2009 (Fazel et al, 2009). It reported an average four-fold increase in violent crime in men with schizophrenia, compared to men without schizophrenia.

However, many older studies were considered small and methodologically limited. They were also seen as unable to adequately assess the temporal relationship between schizophrenia and violence.

Last October saw the publication of another study in this area, this time in the journal Psychological Medicine (Fleischman, 2014). The investigators identified hospitalisations for schizophrenia using an Israeli case registry.

These cases, numbering more than three thousand, were compared to an age- and gender- matched comparison group on their exposure of interest, which was having committed a violent crime. Information on this was gathered from police records, to which the case registry data was linked.

Logistic regression was used to assess the relative difference in the prevalence of violent offences in cases compared to controls, after accounting of alternative explanations for the difference, including age, gender, marital status, family structure, and substance misuse.

Although the full statistical model is not presented, the report states that schizophrenia still had a significant effect on violent crimes, after accounting for confounding by the other variables.

Methods

• The study linked the Israeli population registry with a separate register of psychiatric hospitalisations. Within the case register, the case series for analysis were defined as those individuals admitted to any psychiatric ward born after 1970.
• For each case identified, the population register was used to identify their parents and all their siblings. For each case of sibling and parent identified, four controls were identified from the population register at random.
• Sociodemographic information and details of crimes committed was gathered for both cases and controls. Violent crimes were defined as murder, attempted murder, assault, rape, threats and robbery.
• Half-siblings were excluded. Where there were families with more than one case eligible for inclusion, one was selected at random.
• Differences in the prevalence of criminal conviction were assessed between cases and controls.
• Logistic regression was used to adjust for age, gender, marital status, parity, substance misuse, and suicide attempts. This accounted for differences in conviction prevalence (other than the diagnosis).

Results


• There was an increased risk of criminal activity with schizophrenia, peaking at around the time of the first admission.
• Overall, violent crimes were 4.3 times more common in cases compared to controls, and 2.8 times more common in cases compared to unaffected siblings.
• There were also associations with sex-related and drug-related crimes.
• Among patients/cases, violent crime conviction was predicted significantly by:
  • male gender
  • substance misuse
  • multiple hospitalizations (more than four)
  • having violent parents
  • having a suicide attempt
  • having a sick family member.

Conclusions

The study found that individuals with schizophrenia were more likely to have committed an offence, an association which was strongest for violent crimes, but which pertained also for sex- and drug- related crimes.

This was consistent with previous evidence, although the authors pointed out that there are relatively few previous studies. The authors also pointed out that the association with sex-related crimes goes contra to some of the literature. They do not address possible explanations for this phenomenon in their data, e.g. confounding or measurement bias.

This study demonstrates an association not only between schizophrenia and subsequent crime, but also with crimes committed prior to the index admission, bolstering the idea that individuals who go on to be diagnosed with illness later in life exhibit behavioural differences to the rest of the population.

As expected, individuals who had a large number of admissions were at greater risk of being convicted for a violent crime – implying either more severe illness resulting in more severe behavioural disturbance, or more behavioural disturbance leading to more recent admissions.

Strengths and limitations

The authors, in their discussion of the limitations of the work, refer to:

• the inability to calculate the impact of schizophrenia on crime rates,
• the lack of socioeconomic information in their analysis,
• the possibility that the measures of substance misuse were vulnerable to reporting bias, because they were based on self-report information.

As a reader, the analysis is well conducted and based on an enormous dataset- the effect of chance on this dataset is likely to be limited.

However, the absence of any socioeconomic information on the cases or controls is a big problem. We know that there is probably a causal relationship between low socioeconomic status and criminal activity – if anything it is the most widely agreed upon cause!

And we, of course, also know that a very strong correlation exists between social position and schizophrenia (Wicks et al, 2005). The differences observed in this study between convicted and non-convicted subjects may lie not in whether or not they have been diagnosed with schizophrenia, but rather in the fact that those convicted are likely to be poorer, less well educated and less employed.

Thus, the authors leave an important competing explanation for their results unexplored. This is unfortunate.

The authors point out that there is a consensus that schizophrenia is associated with offending, and that therefore this should involve practice in the assessment, treatment and monitoring of patients.

I think, however, that this work reflects a traditional bias in the issues that psychiatric research has concerned itself with, focusing on the ill as perpetrators rather than victims. I think the changes that should be made as a result of this research should be a keener focus on individuals with the diagnosis of schizophrenia as victims, including prior to the onset of illness.

Links

Fleischman A, Werbeloff N, Yoffe R, et al. Schizophrenia and violent crime: a population-based study. Psychological medicine 2014;44(14):3051-57. [Abstract]

Lindqvist P, Allebeck P. (1990) Schizophrenia and crime. A longitudinal follow-up of 644 schizophrenics in Stockholm. The British Journal of Psychiatry 1990;157(3):345-50 [Abstract]

Fazel S, Langstrom N, Hjern A, et al. (2009) Schizophrenia, substance abuse, and violent crime. JAMA – Journal of the American Medical Association 2009;301(19):2016-23.

ONS (2013). Crime in England and Wales, Year Ending June 2013. Office for National Statistics.

McGrath J, Saha S, Chant D, et al. Schizophrenia: a concise overview of incidence, prevalence, and mortality. Epidemiologic Reviews 2008;30(1):67-76

Wicks S, Hjern A, Gunnell D, et al. (2005) Social adversity in childhood and the risk of developing psychosis: a national cohort study. Am J Psychiatry. 2005 Sep;162(9):1652-7. [PubMed abstract]

maratr / Shutterstock.com