Contact Between Police and People With Mental Disorders: A Review of Rates

A systematic search was conducted in PubMed, PsycINFO, Web of Science, JSTOR, Criminal Justice Abstracts, Sociological Abstracts, and the National Criminal Justice Reference Service. The search was first performed in September 2009 to scope the literature as part of a larger study focused on gathering self-reported data from people with mental illnesses about their encounters with the police. After the primary data were analyzed and published, the search was updated in June 2015 and more carefully analyzed for patterns. Combinations of the following terms were searched: (police* OR law enforce*) AND (bipolar* OR mania* OR mental ill* OR mental disorder OR schizophreni* OR psycho*). Searches with large numbers of records (>250) were narrowed by adding the interact* term. Additional studies were located by reviewing the reference lists of all included records and by using the related articles and cited articles tools in the Web of Science database.

Studies were included if they were published in peer-reviewed journals, reported primary research findings, and were written in English. No restrictions were placed on publication year or methodological design. Studies were excluded if they reported data in such a way that an exact numerator and denominator, used for calculating rates, could not be ascertained. In addition to studies that met general eligibility criteria, studies on arrest rates were included if they provided data on histories of arrest over the lifetime or over extended periods among persons with mental disorders and were excluded if they used only police-involved samples or examined arrest rates over a brief period (fewer than five years). Studies of rates of police involvement in care pathways were included if they reported data on the proportion of people who interacted with police on their way to inpatient or outpatient mental health services, including emergency or compulsory psychiatric care, and were excluded if they used only police-involved samples. Studies of police calls for service were included if they provided data on the proportion of police dispatches or encounters that involved people with mental disorders and were excluded if they focused only on use-of-force incidents, used only samples detained in police custody, or focused only on calls to police–mental health crisis teams. [The excluded studies are listed in a table in an online supplement to this article.]

For every included study, data about selected variables were extracted and entered into Excel and SPSS. Data were extracted by a research assistant and then verified by the author. Student’s t tests, Spearman rank correlation, and analyses of variance were performed to assess differences in average rates across study characteristics. Because of the possibility that studies with exceptionally large samples—exceeding the combined samples of the remaining studies—would distort the findings, those studies were removed from calculations of the overall rates described in the text.

Data from smaller studies—those with fewer than 500 individual cases—were combined for summary in Tables 1–3. Rates that include all studies are reported in the tables. In total, 85 unique studies with 329,461 cases involving contact between the police and people with mental disorders were included and synthesized. Cases covered individuals and events.

This review synthesized rates reported in 85 unique studies of contact between police and people with mental disorders. The estimates varied considerably depending on the type of rate. Overall, the findings suggest that typically one in four people with mental disorders have histories of police arrest, which is substantially higher than estimated rates among general adult populations in Canada (95) and the United Kingdom (96) but in line with rates in the United States (97,98). This review also found that about one in ten individuals encountered police in their pathway to mental health care, and one in 100 police dispatches and encounters involve people with mental disorders. Such results shed empirical light on claims made by journalists, politicians, police personnel, and scholars about the frequency with which the police and people with mental disorders interact.

As is the case with most reviews of this nature, the included studies reported a wide range of rates, varying across the populations studied and the methods employed. This body of research predominantly consists of samples drawn from the United States, with some indication that U.S. police officers, compared with those in other countries, may play a greater role in connecting people to mental health services. When viewed optimistically, this might suggest that U.S. police officers are more actively involved, compared with their counterparts in other countries, in facilitating precharge diversion processes aimed at preventing people with mental disorders from further involvement with the criminal justice system. Conversely, this finding may also signal serious problems associated with the criminalization of people with mental disorders in the United States, whereby criminal justice processes, agents, and institutions are used to manage mental health issues. Another source of variation in rates was the diverse methods used for establishing the presence of an arrest history and a mental disorder, such as self-report versus official records, although these findings were mixed across the three types of rates examined in this review. Methodological quality of the studies was not formally appraised and controlled in this review, but the rates varied across different methodological characteristics, such as the size of the samples and the various ways in which mental disorder was operationalized.

One problem that has plagued research in this area is how to establish valid and reliable indicators of mental disorder in the context of police contacts. Scholars have developed rigorous field observation methods to reduce biases inherent in police-generated data (90,94). A limitation of this approach is that it is resource-intensive and difficult to perform on a large scale. Sophisticated data algorithms also have been produced to identify persons with mental disorders within administrative police databases (80,81,83,84). Although this method can be carried out on large data sets and is relatively easy to replicate, it is dependent on the reporting practices of dispatchers and police officers and, as such, likely underestimates the number of encounters that actually involve people with mental disorders. Triangulating data and methods, as was done in this review, enhances the validity of such rates, but the underreporting of mental disorders remains an issue. A number of factors influence whether mental disorders will be discovered by, or reported to, the police, including relevance to the encounter, concealability, the skills of police personnel, and the fact that people who more often interact with the police, such as men and persons from racial-ethnic minority groups, may be the least motivated to disclose mental health issues (99). It would be prudent for researchers and decision makers to be mindful of these methodological issues when comparing the rates of different jurisdictions and when drawing conclusions about the nature and magnitude of the phenomenon.

Rates of contact between the police and people with mental illness tell us more about the patterned social arrangements in society than the criminality of individuals with mental disorders. Numerous studies expose the powerful social forces and diverse situations that bring these two groups together. Police interactions involving people with mental illnesses are often casual in nature, unrelated to offending behavior, and resolved through informal means (80,93,100). The criminalization of illicit substances, the coercive underpinnings of mental health policies and services, and the intersecting functions of our mental health and criminal justice institutions are all implicated in these trends (101,102). Consistent with other reviews (103), this study revealed that rates of contact between police and people with mental disorders were correlated with a number of individual, organizational, and contextual factors, including homelessness, substance use problems, gender and race, and social disorganization. In addition to methodological differences, these characteristics are likely major sources of variation in rates between studies and jurisdictions. Interestingly, one study indicated that building the capacity for police services to respond appropriately to people with mental disorders, such as through training or specialized response teams, may increase the rate of mental health–related calls for services; however, the drivers of such changes (reporting practices versus greater reliance on police services) are not well understood and warrant greater research attention. What is clear is that reducing the rates revealed in this study will require that improvements be made to the undesirable social context in which people with mental disorders find themselves: terrible situations requiring routine and recurrent intervention by the police.

A number of methodological limitations must be noted. First, it is likely that a sizeable number of studies reporting on arrest histories among people with mental disorders were not captured by the literature search because they focused on other substantive issues and their titles, abstracts, or keywords did not contain necessary terms (that is, police* OR law enforce*). Second, the literature search was first performed in 2009 and then replicated in 2015; consequently, it was not possible to determine and report on the total number of studies screened and assessed for eligibility. Third, methodological quality of the included studies was not formally appraised, and variability in quality was not controlled. Fourth, an initial search of the gray literature revealed few unpublished documents reporting primary data and satisfying other inclusion criteria. As such, the review was restricted to published research, which may pose a threat to the validity of the findings on account of publication bias.

The gross estimates of rates presented in this study will be of interest to researchers and police personnel who focus their efforts on understanding and improving such interactions; however, the study did not engage with a number of central questions. For instance, are these interactions increasing in frequency? Is the nature of these encounters changing? Are repeated contacts becoming more common? In what ways do interactions with police differ for people with or without mental disorders? How are rates affected by the implementation of specific strategies, including police training and specialized response models, aimed at improving encounters between police officers and people with mental disorders? What role does the presence of a mental disorder, as opposed to other factors, actually play in these interactions? Such questions have important implications for policy makers, police leadership, and health care planners and are worthy of additional research.

Numerous coroner inquests and public inquiries after tragic events have created a heightened awareness of issues surrounding police encounters with people who have mental disorders. A prevailing sentiment is that many of these interactions are undesirable, unnecessary, and avoidable. By providing estimates regarding the rates of contact between police and people with mental disorders, this study illuminates the magnitude of the issue and supplies an empirically supported reference point to scholars and practitioners in this area.

The author thanks Zachary MacMillan, Karen Chu, and Michelle Pritchard for their contributions to the project.