This study compared officially recorded rates of crime victimisation between a large sample of persons with schizophrenia-spectrum disorders and a randomly selected community sample that had never been diagnosed with a schizophrenic illness. Persons with schizophrenia-spectrum disorders were significantly less likely to have an official record of non-violent victimisation, but more likely to have a record of violent crime victimisation. Furthermore, when victimised, they were more likely than the community sample to experience repeated instances of victimisation. Although co-morbid substance-use disorders were associated with more frequent victimisation records, even schizophrenia-spectrum patients without a substance-use disorder were more likely than controls to have a record of violent victimisation.
This was also the first study to compare rates of crime victimisation over a period of deinstitutionalisation. Alarmingly, the rates of victimisation amongst schizophrenia-spectrum patients were more than double in those first diagnosed in 1995 compared to those first diagnosed in 1975, whilst rates in the community sample remained constant. These findings provide further indication that people with severe mental illness are vulnerable to harm and victimisation in the community, and bring into question whether the current community-based mental health care model is able to manage and protect this vulnerable population.
Limitations of the study
An inevitable degree of error is incurred when using case-linkage procedures. However, the chance of these results being impacted by such error is minimised by the large sample size (N=8,809) and exclusion of cases where there was some indication of linkage inaccuracy. An additional limitation of case-linkage is that data interpretation is restricted by the accuracy and completeness of data available. Information contained in the databases was not collected for research purposes and is limited in scope, precluding the consideration of certain variables (such as socioeconomic status) in this study. It is also possible that relatively minor instances of crime victimisation which occur in institutional settings (such as prisons) may be dealt with internally and not reported to police, therefore excluding these incidents from the source database. However, if this is the case, then rates of victimisation would be expected to be even greater in the schizophrenia-spectrum sample, as this group are more likely to be institutionalised.
It is well-recognised that official records are an under-estimate of the true crime rate , and the rates reported here will under-estimate the true levels of victimisation in the community. However, the aim of this study was not to identify absolute rates of community victimisation, but rather to ascertain the level of officially recorded victimisation amongst schizophrenia-spectrum patients, and compare this with that recorded by the general population (using identical methods). Although there is currently no research comparing crime reporting by people with and without mental illness, there is reason to suspect that people with severe mental illness may be even less likely than others in the community to report victimisation experiences to the police . As such, any significant findings of more frequent victimisation in the schizophrenia-spectrum sample can be interpreted with confidence, as they are likely to only underestimate the true difference between these groups.
In addition to crime victimisation, the rate of substance-use disorders in both groups is likely to be underestimated here, due to less than perfect detection and recording of substance problems on the VPCR. It may further be the case that substance-use disorders are even less likely to be identified in the community sample, as this group were typically not in regular contact with the public mental health system and arguably had less opportunity for diagnosis and detection of substance problems. However, it is well-recognised that people with major mental illness engage in more prevalent substance misuse than the general community, and this general pattern was reflected in the current study [31, 32].
Finally, it must be noted that the current study was conducted in a country with relatively low rates of crime, which now operates under a community-based mental health system. These results therefore may not be directly generalisable to other jurisdictions with different base rates of crime, or significant variations in policing and mental health systems. We note also that the base rate of violent victimisation is low, and that the practical significance of statistically significant differences in odds ratios must be considered in this context. Nevertheless, given the serious consequences of violence victimisation, even relatively small numerical differences may have a powerful affect on individuals and communities.
Crime victimisation in people with schizophrenia-spectrum disorders
In contrast with previous research , the current study found that people with schizophrenia-spectrum disorders are less likely than the general community to have an official record of victimisation, and also have a lower rate of recorded non-violent victimisation. This finding appears to be at odds with the general hypothesis that people with mental illness are more vulnerable to both violent and non-violent forms of crime [1, 15]. One possible explanation for these results is that people with schizophrenia-spectrum disorders are less likely to report non-violent or less serious victimisation to police than other persons in the community. Contemporary research also shows that offenders or persons who have had prior negative experiences with police are less likely to report their own victimisation experiences , a finding which may be particularly pertinent for schizophrenia-spectrum patients given the increased rates of offending and police contact found among this population [7–9]. Further, it is plausible that certain psychotic symptoms (such as paranoia or persecutory delusions) will further discourage schizophrenia-spectrum patients from reporting victimisation experiences to authority figures. It is also recognised that many people are motivated to report non-violent property crimes by requirements of their insurance policies ; if persons with severe mental illness are less likely to have comprehensive insurance (or cannot afford to pay excess costs), it follows that they would be less likely to report such crimes. Thus, it is perhaps reasonable that recorded rates of victimisation overall are lower in people with schizophrenia-spectrum illnesses. Further research could consider a more nuanced comparison of different types of non-violent crime to tease out where the differences lie and therefore to provide a focus for enhanced criminal justice supports and interventions.
On the other hand, the recorded rates of violent victimisation were significantly higher amongst the mentally ill group. Overall, one in ten people with a schizophrenia-spectrum disorder had been recorded as a victim of violent crime, and this group were more likely to have a record of victimisation than the community controls. Of concern, the risk of victimisation seems to be particularly high for sexual offences; patients with schizophrenia-spectrum disorders were nearly three times more likely to have a record of sexual violence victimisation. Moreover, victims with schizophrenia-spectrum disorders had more frequent and repeated victimisation incidents than community controls.
The lifetime prevalence of recorded violent victimisation for persons with schizophrenia-spectrum disorders was 10%, a rate markedly lower than that reported in most self-report studies [1, 14, 15]. The notable disparity here confirms that there is a significant ‘dark figure’ of violent victimisation, and suggests that people with serious mental illness may be less inclined to report even serious violent offences to the authorities.
Relationship between victimisation, prior offending, and substance-use disorders
Confirming previous research, the current study demonstrated that the presence of a substance-use disorder significantly increases the risk of victimisation [15, 16]. Indeed, patients with both a schizophrenia-spectrum and substance-use disorder were nearly four times more likely than community controls to have a record of violent crime victimisation, and over six times more likely to have a sexually violent victimisation record. They were also nearly twice as likely to have a victimisation record when compared to schizophrenia-spectrum patients without co-morbidsubstance-use disorders. However, substance misuse alone cannot explain this association, since the risk of victimisation was significantly higher among those with schizophrenia-spectrum disorders even when the presence of a co-morbidsubstance-use disorder was taken into account. Moreover, violent victimisation records were still significantly more common in schizophrenia-spectrum patients without co-morbid substance-use disorders than in controls.
A history of offending also increased the risk of victimisation in both the community and schizophrenia-spectrum groups. Community members who had offended were three times more likely than other community members to have a victimisation record, whilst the rates of official victimisation amongst offending schizophrenia-spectrum patients was more than 3.5 times greater than that of patients with no offence history. This finding supports previous research which shows that offending and victimisation are important reciprocal risk factors , and emphasises that the two categories are not mutually exclusive .
Victimisation over time: the impact of deinstitutionalisation
One of the most alarming findings in this research is that official rates of victimisation in schizophrenia-spectrum patients appear to have risen dramatically over the period of deinstitutionalisation, an increase not paralleled in the general community. Indeed, whilst victimisation rates in the community have remained fairly constant over the past thirty years, the number of recorded victimisation incidents in the schizophrenia-spectrum sample has more than doubled.
There are three plausible explanations for this apparent increase. Firstly, one can hypothesise that the deinstitutionalisation process which occurred between the 1970s to mid -1990s has resulted in increasing numbers of mentally ill persons being exposed to risks within the community that they had previously been ‘sheltered’ from in institutional care, thus leading to a rise in reported victimisation rates. Indeed, a similar argument has been proposed to account for rising levels of homelessness [23, 35]. In support of this theory, the completion of deinstitutionalisation in Australia in the mid-1990s  corresponds closely with a stabilisation in the rates of victimisation in schizophrenia-spectrum patients; this rate has now remained relatively constant for 15 years. A second related consideration relates to the decreases in the length of inpatient psychiatry admissions more recently and therefore the likelihood that the estimated rates of victimisation would have been an underestimate of the true prevalence that occurred in the inpatient environment prior to deinstitutionalisation. While police can be, and are, called upon when crimes are committed in inpatient settings, it is highly likely that a number of additional decision-making processes among hospital staff deterred action from being taken, due to their high thresholds for violent and otherwise antisocial behaviour that may normally lead to police involvement in community settings.
A third explanation for the increase focuses on changes in policing practice and social policy. Since the 1980s, there has been a greater focus on community well-being and engagement, and ‘community policing’ is now a well-established model in many countries. Through the implementation of community policing models, a number of initiatives have been developed to improve relationships with vulnerable populations. For example, in Victoria, all new police recruits are now trained in recognising the signs of mental illness and are given strategies for interacting and communicating with this population . In some jurisdictions, mental health officers are employed to facilitate effective relationships between police and the mentally ill. Undoubtedly, this new approach has contributed to a more active ‘social welfare’ role for operational police , which may be reflected in an increasing number of schizophrenia-spectrum patients finding themselves in a position of reporting incidents of crime victimisation to the police.