Homicide in discharged patients with schizophrenia and other psychoses: A national case-control study

https://doi.org/10.1016/j.schres.2010.08.019Get rights and content

Abstract

Objective

To investigate factors associated with homicide after discharge from hospital in patients with schizophrenia and other psychoses.

Design

All homicides committed by patients with psychosis within 6 months of hospital discharge were identified in Sweden from 1988–2001 and compared with patients with psychoses discharged over the same time period who did not subsequently commit any violent offences. Medical records were then collected, and data extracted using a validated protocol. Interrater reliability tests were performed on a subsample, and variables with poor reliability excluded from subsequent analyses.

Results

We identified 47 cases who committed a homicide within 6 months of discharge, and 105 controls who did not commit any violent offence after discharge. On univariate analyses, clinical factors on admission associated with homicide included evidence of poor self-care, substance misuse, and being previously hospitalized for a violent episode. Inpatient characteristics included having a severe mental illness for one year prior to admission. After-care factors associated with homicide were evidence of medication non-compliance and substance misuse. The predictive validity of combining two or three of these factors was not high. Depression appeared to be inversely associated with homicide, and there was no relationship with the presence of delusions or hallucinations.

Conclusions

There are a number of potentially treatable factors that are associated with homicide in schizophrenia and other psychoses. Associations with substance misuse and treatment compliance could be the focus of therapeutic interventions if validated in other samples. However, their clinical utility in violence risk assessment remains uncertain.

Introduction

Between 10–20% of homicide perpetrators have psychosis (Eronen et al., 1996a, Eronen et al., 1996b, Fazel & Grann, 2004), an increased risk for homicide of around 20 times in individuals with schizophrenia and other psychoses compared with general population controls (Fazel et al., 2009a, Fazel et al., 2009b). Although risk factors for violence in individuals with schizophrenia have been studied (Fazel et al., 2009a, Walsh et al., 2004, Soyka et al., 2007), little is known about what factors are associated with homicide in psychosis. Clinical studies investigating homicide perpetrators have focused on clinical characteristics, but without appropriate comparison groups, they are not able to explore potential risk factors. Within-group designs have demonstrated comorbid substance abuse significantly increases the risk of homicide in individuals with schizophrenia (Eronen et al., 1996a, Eronen et al., 1996b, Schanda et al., 2004). In addition, there is evidence that treatment mediates risk. Those with first episode psychosis have higher rates of homicide than individuals after their first episode (Nielssen and Large, 2008).

Therefore, we aimed to undertake a case control study to investigate factors associated with homicide after discharge from psychiatric hospital in individuals with psychosis. This design has two advantages. First, it allows identification of factors that distinguish patients who were at future risk at committing homicide from patients who were not. Second, by investigating homicide only after discharge from hospital, it has the potential to inform clinical risk assessment. To our knowledge, this is the first such study that uses a control population with psychosis. We hypothesized that important factors associated with homicide in schizophrenia would be potentially treatable.

Section snippets

Methods

In Sweden, all residents are given a ten-digit unique identification number making linking of data in national registers possible. We linked two nationwide population-based registries in Sweden: The Hospital Discharge Registry (HDR; held at the National Board of Health and Welfare), and the National Crime Register (National Council for Crime Prevention).

The Hospital Discharge Register monitors all psychiatric hospitals and is the largest inpatient register worldwide. Reporting to the register

Results

A total of 2,005 individuals committed 2,093 homicides 1988–2001. The average population in Sweden for those aged 15 and over was 7,176,361, which equates to an overall rate of 28 homicide perpetrators per 100,000 persons in the Swedish criminally responsible population. Overall, 409 homicide offenders had a psychosis diagnosis out of a total number of 98,092 patients discharged with schizophrenia and other psychoses (equivalent to a rate of 417 homicide perpetrators per 100,000 patients with

Discussion

We report a national case-control study of patients with psychosis, of whom 47 committed a homicide within 6 months of discharge. These were compared with 105 patients with psychosis who were not convicted of any violent offence after discharge. This design enabled exploration of factors associated with homicide in patients leaving hospital that could potentially inform violence risk assessment and management. As the number of cases was small, our findings should be interpreted primarily as

Conclusion

Homicide by psychiatric patients is rare but an event of considerable consequences. In addition to the effects on the family and friends of victims, the impact on services and professionals is also noteworthy as questions of accountability and responsibility are inevitably raised. The public image of the psychiatry, stigma suffered by patients (Corrigan et al., 2002, Van Dorn et al., 2005), and even recruitment into the profession (Lambert et al., 2006) are likely longer term consequences of

Role of the funding source

This project was funded by the Stanley Medical Research Institute (grant#04R-780).

Contributors

SF and MG designed the study and wrote the protocol. PB and VR assisted with the administration of the project and supervision of research assistants, and PB with the data analysis. SF did the data analysis and prepared the first draft. All authors contributed to revisions and a final draft.

Conflict of interest

The authors report no conflict of interest.

Acknowledgements

We are grateful to Ms Andrea Mogues for the data extraction, to Dr John Powell and Dr Julia Sinclair for sharing with us data extraction sheets that we adapted for this study, and Ms Lena Andersson for extracting data for the interrater reliability study.

References (32)

  • B. Ekholm et al.

    Evaluation of diagnostic procedures in Swedish patients with schizophrenia and related psychoses

    Nord. J. Psychiatry

    (2005)
  • M. Eronen et al.

    Mental disorders and homicidal behavior in Finland

    Arch. Gen. Psychiatry

    (1996)
  • M. Eronen et al.

    Schizophrenia and homicidal behavior

    Schizophr. Bull.

    (1996)
  • S. Fazel et al.

    Psychiatric morbidity among homicide offenders: a Swedish population study

    Am. J. Psychiatry

    (2004)
  • S. Fazel et al.

    The population impact of severe mental illness on violent crime

    Am. J. Psychiatry

    (2006)
  • S. Fazel et al.

    Risk factors for violent crime in schizophrenia: a national cohort study of 13, 806 patients

    J. Clin. Psychiatry

    (2009)
  • Cited by (69)

    • The phenomenon of overkill in northern Greece: A descriptive forensic psychiatric study between 2015 and 2020 on criminal offenders found not guilty by reason of insanity

      2022, Journal of Forensic and Legal Medicine
      Citation Excerpt :

      As regards the Positive and Negative Syndrome Scale (PANSS) at the time the crime was committed, offenders showed significantly higher scores for symptoms of schizophrenia, with seven-fold more offenders scoring above average (p < 0.001) [overall mean value 33.826–45]. Regarding the Negative scale, moderately higher levels were exhibited by female offenders (p = 0.087) [overall mean value 25.412–40], while the General Psychopathology scale disclosed significantly more (2.4-times) offenders scoring below average (p = 0.00386) [overall mean value 57.639–78]. The overall PANSS Total score mean value was estimated at 116.8 (92–150), while comparative scores for other subcategories were unremarkable.

    • Impulsive and aggressive traits and increased peripheral inflammatory status as psychobiological substrates of homicide behavior in schizophrenia

      2022, European Journal of Psychiatry
      Citation Excerpt :

      The majority of individuals with schizophrenia are not considered to be dangerous to society, and their contribution to the overall homicide rate is small. However, an association between schizophrenia and homicide has not only been established in the public mind and law, but has also been demonstrated in epidemiological research.1–5 This association appears to be complex and is largely mediated by factors distinct from the diagnosis itself.

    View all citing articles on Scopus
    View full text