Mental illnesses are associated with a range of adverse outcomes that can seem to jostle for attention from researchers, clinicians, elves and the public. Criminal offending, especially violence, is one that’s been much debated. This has driven research, and a more measured understanding: a range of diagnoses are associated with modest increases in relative risk, more so if substance use disorder is also present (Stevens H et al, 2015; Webb R et al, 2014; Fazel S et al, 2014), but most individuals with mental illness aren’t violent. These subtleties need responsible communication, facing up to the evidence (Mullen P, 2009), but with care not to overstate links, as perceived dangerousness remains a major part of stigma.

Whenever media outlets fail to convey these nuances, important perspective can be provided by highlighting another outcome: people with mental illnesses are as likely to be victims of crime as offenders. Most clinicians would agree this rings true in general psychiatric practice, but as pointed out here by Dean and colleagues, this phenomenon hasn’t received the same epidemiological attention as perpetration research. Cross-sectional data has though demonstrated high rates of self-reported crime victimisation, and proneness to psychological sequelae from it (Khalifeh H. et al, 2015).

In this new paper, researchers from Australia, Denmark and the UK seek to better understand associations between mental illness and vulnerability to crime by using techniques that have helped unpick risks of violence perpetration. That is, they work longitudinally with big registry data; an approach several of the authors previously applied to violent offending (Stevens H et al, 2015).

Methods

This national cohort study included all persons born in Denmark 1965-1998 alive on their 15^th birthday (2,058,063 individuals).

Data was collected from 2001 (or from 15^th birthday, whichever later) to either the end of 2013, first event of reporting crime victimisation, death or emigration.

Everyone in Denmark is routinely assigned a personal identification number, allowing linkage between national registers. Key parameters extracted were:

• Reports to police that an individual had been the victim of a crime, whether pursued after report or not.
• ICD codes for mental disorders diagnosed by a treating clinician, grouped according to ICD-10 diagnostic category.
• Date of onset taken as first inpatient, outpatient or emergency department contact.

A time-to-event survival analysis was undertaken, calculating an incident-rate ratio for first crime victimisation event per 1,000 person-years at risk (i.e. following onset of mental illness), compared to no mental disorder.

Several covariates were included: individuals’ own offending (guilty verdicts from the National Crime Register); paternal income; highest maternal education level; age; calendar year and unknown paternal identity.

Results

• Among individuals with any mental disorder, incident rate ratios (IRRs) of being subjected to any crime compared to individuals without mental disorder were
  • 1.68 (95% CI 1.65 to 1.71) for men
  • 1.71 (95% CI 1.68 to 1.73) for women
• This was attenuated but remained significant when adjusted for confounders, with individuals’ own criminal offending having the strongest effect, adjusted IRRs:
  • 1.49 (95% CI 1.46 to 1.51) for men
  • 1.64 (95% CI 1.61 to 1.66) for women
• The association was seen across diagnostic categories
  • Except intellectual disability and developmental disorders (where for the latter there was a negative one)
  • Strongest associations were for:
    • substance use-disorders (adjusted IRRs 1.95 in men and 2.65 in women)
    • personality disorders (adjusted IRRs 1.60 in men and 1.70 in women)
• For being subject to violent crime, patterns didn’t materially change but magnitude increased, particularly for women. Adjusted IRRs for any mental disorder were:
  • 1.76 (95% CI 1.72 to 1.80) for men
  • 2.72 (95% CI 2.65 to 2.79) for women
  • Strongest associations were again with substance-use disorders
• The effect of an individual’s own offending was most pronounced for substance-use disorders, where adjustment reduced IRR by 20-30% for any offence and 35% for violence offences.

Conclusions

The authors summarise:

Our findings augment the existing literature by demonstrating that the heightened vulnerability to being subjected to crime with any offense, and violent crime in particular, extends to those with a wide range of mental illnesses, is true of those reporting to police, occurs after mental illness onset, and is not confined to individuals with mental disorders in the population who are treated as inpatients.

Regarding the effects of an individual’s own offending, they suggest this is complex, and that overlap likely arises from shared risk factors for the two outcomes among individuals with and without mental illness.

Strengths and limitations

This study adds value in an important area with a large dataset and temporal ordering of illness and outcome (rather than using lifetime diagnosis). Examination of diagnostic categories and offence perpetration within the same study are also helpful additions.

Choice of outcome measure is important. Police-report data on crime victimisation is held up as an improvement on survey data. It avoids issues such as recall bias, and allows national-scale extraction. However, we are not presented with much justification of the validity and accuracy of this pretty specific piece of Danish police data.

The exception of developmental disorders is an interesting signal, but a broad category including both autism spectrum disorders (ASD) and attention deficit hyperactivity disorder (ADHD) perhaps limits interpretation. Register-based ASD diagnoses were found to not be associated with increased risk of perpetrating violence in a Swedish population study (Lundstrom S. et al, 2013), whereas increased risk was found for ADHD. It would have been informative to know whether such separation was seen here for victimisation.

The overlap between perpetration and victimisation is for me one of the most interesting aspects, though the types of offences perpetrated are unspecified. It adds to work that for example demonstrated exposure to violence as a trigger for perpetration (Sariaslan A. et al, 2016), raising further the idea that targeting certain mechanisms may be of shared benefit. In its conclusions, the paper pits victimisation against offending as an unfairly ignored competing outcome. For me though, the point seems more that our patients may be both victims and at increased risk of offending, and we should strive to address both.

Implications for practice

This study adds epidemiological weight to support the routine clinical assessment of risk of crime victimisation. My own experience is that clinicians are aware of such vulnerabilities, and in my trust ‘risk from others’ is already part of the mandated risk assessment. So maybe a more pertinent need is for better monitoring of crime receipt as an adverse outcome (it is not for example captured by Health of the Nation Outcome Scales). And as alluded to by the authors, perhaps efforts to identify unmet mental health needs should consider victims in contact with the justice system, rather than just offenders.

The whole point of assessing the risk of an outcome is to try to do something helpful about it. Interventions for crime victimisation in psychiatry therefore need development. To target treatment at those most likely to benefit, we also need clinically useful ways of assessing and stratifying risk (Fazel et al, 2017). Despite violence perpetration being more studied epidemiologically, this still hasn’t translated into much in the way of evidence-based interventions (Wolf A. et al, 2017). The further indication in this study of overlap between offending and victimisation suggests evaluation of therapeutic approaches may do well to consider both.

Conflicts of interest

None.

Links

Primary paper

Dean K, Laursen TM, Pendersen CB et al (2018) Risk of being subjected to crime, including violent crime, after onset of mental illness. A Danish national registry study using police data. JAMA Psychiatry 2018 75(7):689-696. [PubMed Abstract]

Other references

Mullen PE (2009) Facing up to unpalatable evidence for the sake of our patients. PLoS Med 6(8): e1000112.

Stevens H, Laursen TM, Mortensen PB et al (2015) Post-illness-onset risk of offending across the full spectrum of psychiatric disorders. Psychol Med 45(11):2447-2457. [PubMed Abstract]

Webb RT, Lichtenstein P, Larsson H et al (2014) Suicide, hospital-presenting suicide attempts, and criminality in bipolar disorder: examination of risk for multiple adverse outcomes. J Clin psychiatry. 75(8):e809-16.

Fazel S, Wolf A, Palm C et al (2014) Violent crime, suicide, and premature mortality in patients with schizophrenia and related disorders: a 38-year total population study in Sweden. Lancet Psychiatry. 1(1):44-54.

Khalifeh H, Johnson S, Howard LM et al (2015) Violent and non-violent crime against adults with severe mental illness. Br J Psychiatry 206(4):275-282.

Lundstrom S, Forsman M, Larsson H et al (2014) Childhood neurodevelopmental disorders and violent criminality: a sibling control study. J Autism Dev Disord 44(11):2707-16. [PubMed Abstract]

Sariaslan A, Lichtenstein P, Larsson H et al (2016) Triggers for violent criminality in patients with psychotic disorders. JAMA Psychiatry 73(8):796-803.

Fazel S, Wolf A, Larsson H et al (2017) Identification of low risk of violent crime in severe mental illness with a clinical prediction tool (Oxford Mental Illness and Violence Tool [OxMIV]): a derivation and validation study. Lancet Psychiatry 4(6):461-8.

Wolf A, Whiting D, Fazel S (2017) Violence prevention in psychiatry: an umbrella review of interventions in general and forensic psychiatry. J Forensic Psychiatry Psychology [Abstract]

Photo credits

Photo by Samuel Zeller on Unsplash