Screening for Mental Illness Among Persons in Amsterdam Police Custody

The study was based on data collected during interviews of police detainees who were held in cell blocks by the Police Service Amsterdam-Amstelland between March and June 2009. Randomly selected detainees (N=402) at Amsterdam cell blocks were invited to participate in an on-site structured questionnaire. The survey was administered by trained interviewers employed by the Amsterdam Public Health Service. Oral consent for participation was obtained after informing the detainees about the objectives of the study. It was explained to participants that all data would be treated confidentially, that participation was voluntary, and that their decision whether to participate would not influence their legal case. Institutional approval for the study was provided by both the Amsterdam Public Health Service and the Police Service Amsterdam-Amstelland.

In total, 264 interviews were completed (response rate of 66%). Ninety-nine detainees refused to participate, 30 suffered from severe language problems, four were willing to participate but were transferred or interrogated at the time of the scheduled interview, and one could not be interviewed in the interest of the investigation. The reasons for nonresponse were not recorded for four detainees. Differences between responders and nonresponders in age and gender were not significant.

The survey covered demographic characteristics and questions on general medical and mental health, use of medication, and health care use in the past 12 months (1). The Brief Jail Mental Health Screen (BJMHS) was used to screen for serious mental illness, such as schizophrenia, bipolar disorder, and major depression (10). The BJMHS consists of eight yes-or-no questions. Section 1 includes six items about mental health symptoms. Section 2 comprises two items, one about current use of psychotropic medications and one about previous hospitalizations. Detainees are considered to be candidates for further mental health assessment if they endorse at least two items from the first section or one item from the second. Several validation studies comparing the results of the BJMHS with those of the Structured Clinical Interview for DSM-IV (SCID) have established its reliability and validity in custody settings (11–13).

Characteristics of detainees with positive and negative screens were compared by using chi-square tests for categorical data and t tests for continuous data. All analyses were performed by using SPSS 17.0 for Windows.

Table 1 displays characteristics of the interview sample. Most detainees belonged to an ethnic minority group (77%) and had a low level of education (69%). Nearly half were unemployed (46%). A total of 39% reported no permanent address, and almost one-quarter reported being indebted. Valid results for the BJMHS were available for 248 detainees, and almost 40% (N=97) screened positive. Among the 248 detainees, 71 endorsed at least two items from section 1 of the BJMHS (mental health symptoms), and another 55 endorsed at least one item from section 2 (current use of psychotropic medications and previous hospitalization). Those who screened positive were significantly more likely to be Native Dutch, to lack a permanent address, and to have financial problems compared with those who screened negative on the BJMHS.

An analysis comparing the same characteristics among detainees by section result found some significant differences (data not shown). Detainees without a permanent address (p=.04) and those with financial problems (p=.02) were more likely to report two or more items about symptoms from section 1. The Native Dutch were significantly more likely than non-Native Dutch to endorse at least one item from section 2 (p=.00). All other characteristics remained nonsignificant.

Almost 40% of survey respondents screened positive on the BJMHS, indicating a need for further mental health assessment. These findings were in line with results from other studies that demonstrated substantial rates of psychopathology among police detainees (3,7). In one study that also used the BJMHS, 58% of Australian police detainees were classified as needing referral (12). However, in a 2005 study by Steadman and others (11) that validated use of the BJMHS in U.S. county jails, the percentage who screened positive was much lower (11%). Research published in 2009 that used the eight-item version of the BJMHS at different U.S. jails found positive screens among 9% to 14% of inmates (14).

It is unclear how these different findings can be explained. In theory, several factors might influence the rate of positive BJMHS scores. First, it is known that women screen positive on the BJMHS more often than men (11,14). The percentages of females in our study and in the Australian and U.S. research were comparable, however. Second, differences between countries in the arrest or custody process might account for different estimates. The United States has the highest incarceration rate worldwide. The reasons for arrest and the offenses that are considered serious enough to warrant arrest, therefore, also might vary. As such, individuals who are arrested because of psychiatric issues may be underrepresented in the United States compared with the Dutch system.

Third, different practices governing the diversion of individuals with serious mental illness shortly after arrest might play a role. Detainees who are arrested for endangering themselves or others because of a psychiatric crisis (a potential reason for police custody) may be diverted to health services sooner in the Netherlands than in Australia. This might explain why a smaller percentage of detainees screened positive for serious mental illness in this study than in the study in Australia.

In addition, our study population included detainees who were transferred from police stations without proper facilities for overnight stays. Cell blocks generally are better equipped for this purpose. However, persons detained at police stations may be directly referred shortly after the arrest to psychiatric services throughout the city. Possibly, the rate of detainees with a positive BJMHS may have been even higher if the sample also included all individuals detained at police stations.

Fourth, the rate of respondents with a positive BJMHS screen might vary depending on who administers the interview. In the Australian study, nursing staff administered the interviews (12). In the 2005 study performed in U.S. jails, the BJMHS was administered twice—first by a correctional officer and later by a clinical research interviewer (11). The results demonstrated that respondents tended to underreport information to correctional officers, leading to false negatives on the BJMHS.

Fifth, differences in national health care systems might affect BJMHS results. Endorsement of either item concerning mental health care use (psychotropic medication use and prior hospitalization) constitutes a positive screening result. Dutch detainees may be more likely than U.S. detainees to have a positive screen because of better access to or better utilization of mental health care. Recent research using the BJMHS showed that in the United States, whites had higher odds than blacks and Latinos of having ever been hospitalized or of currently taking medications for mental illnesses (15). A similar effect was found in our study: Native Dutch detainees endorsed one of the two service items significantly more often than non-native Dutch detainees. The BJMHS attributes considerable weight to prior service use when determining the need for further assessment of detainees. Groups who are known for poor utilization, therefore, might be inadequately identified by the BJMHS (15). This might be the case not only for blacks and Latinos in the United States but also for the nonnative population in the Netherlands.

As does every study, this study had methodological limitations. A standardized screening tool, the BJMHS, was used to investigate the percentage of detainees who most probably suffered from a serious mental condition. The highest standard for the assessment of mental health conditions is a diagnostic interview such as the SCID. The SCID was not applied for two reasons, however. According to the experience of both police and health care staff, persons held in police cells are not inclined to comply with a lengthy interview because of anger about the detainment and feelings of stress and uncertainty concerning the subsequent legal procedure. Moreover, the longer the interview, the higher the chance that the detainee is removed for interrogation or transfer to court, resulting in incomplete data and attrition.

Another concern was the representativeness of the interviewed detainees. The response rate (66%) was considerable, but the possibility that BJMHS outcomes differed among participants and nonparticipants cannot be excluded, which may limit the representativeness of results. At least with regard to age and gender, no significant differences between participating and nonparticipating detainees were found. Future research should attempt to also collect information on detainees’ education level, prior incarceration, or primary language for the purpose of a more meaningful analysis of nonresponse.

Our findings have implications for those responsible for providing mental health services to police detainees and for policy makers in the area of public mental health. The high number of detainees with a positive screen for mental illness raises the question whether a standardized mental health assessment of every person coming into police custody is worthwhile. Systematic screening of police detainees could facilitate the identification of those who have lost contact with their mental health service providers in the community and could help to redirect these individuals to treatment.

Continuity of treatment can prevent persons with mental illness from cycling through the criminal justice and mental health systems and can benefit the public health and legal systems. Furthermore, standardized screening might help to improve patient safety in police custody, with those in need of referral obtaining attention for their mental health issues quickly after arrest. Systematic screening of all incoming detainees might also increase the safety of police and health care staff, who often face aggression and violence from detainees who have mental disorders or who are intoxicated. The benefits and costs of such a systematic approach should be carefully evaluated, given that it also would increase the workload of those involved in the care of police detainees.

This study was supported by institutional funds provided by the Amsterdam Public Health Service. The authors thank the Amsterdam-Amstelland police force for enabling this research.

The authors report no competing interests.