Police Involvement in Involuntary Psychiatry Admission: A Report From the Dublin Involuntary Admission Study

We studied all admissions to two psychiatry units at Tallaght University Hospital and Mater Misericordiae University Hospital, Dublin, Ireland, between July 1, 2015, and December 31, 2018. This project formed part of a larger study of psychiatry admission trends in Dublin inpatient psychiatry units, the Dublin Involuntary Admission Study. The psychiatry unit at Tallaght Hospital is a 52-bed admission unit serving a suburban catchment area of 273,419 people. The psychiatry unit at Mater Hospital is a 15-bed admission unit, with an inner-city catchment of 32,869 people.

We recorded gender, date of birth, employment and marital status, date of admission, and date of discharge, as well as clinical discharge diagnosis using the ICD-10. For patients whose status was involuntary for part or all of their admission, we recorded the duration of involuntary admission, who had instigated the admission, and whether the status was involuntary from the outset or underwent a change.

We obtained national data about psychiatry admissions in Ireland for 2017 (midway through this study) from the National Psychiatric Inpatient Reporting System (14). The study was approved by the local research ethics committees and performed in accordance with the Declaration of Helsinki. Data protection legislation was adhered to at all times.

We analyzed data with IBM SPSS Statistics, version 25. For bivariable analysis, we used the Pearson chi-square test and independent-samples t test. For multivariable analysis, we generated a binary logistic multivariable regression model with whether or not the involuntary admission was initiated by police as the dependent variable. Independent variables were those significantly associated with the dependent variable in bivariable testing.

During the 3.5-year period studied, 2,715 fully recorded admissions to the two units occurred, of which 2,272 (84%) were voluntary and 443 (16%) were involuntary (246 males and 197 females). Compared with national rates, admission rates to the two units were lower for overall admissions (250.4 vs. 351.6 per 100,000 people per year), voluntary admissions (209.6 vs. 306.2), and involuntary admissions (40.8 vs. 45.4). Notably, however, the balance of voluntary and involuntary admissions differed between the units in this study and the national data; the proportion of admissions that were involuntary in the study units was higher than the national proportion (16% vs. 13%).

Overall, 1,384 male patients (51% of the total sample) and 1,333 female patients (49%) were admitted; 18% (N=246) of the admissions of males and 15% (N=197) of females were involuntary (see table in online supplement). The M±SD age was 41±15 years (range 16–91). Some data were missing for 53 of the involuntary admissions, reducing the total included in the following analyses to 390. Twenty percent (N=78) of applications for involuntary admission were made by police officers, comprising 22% (N=48) of male and 17% (N=30) of female patients. About one-quarter of applications (27%, N=104) were made by members of the public, 18% (N=70) by family members, and 11% (N=43) by authorized officers of the health service. About one in five involuntary admissions (22%, N=86) was a change of status to involuntary, and 2% (N=9) were involuntary transfers from other hospitals.

Schizophrenia, schizotypal, and delusional made up the most common diagnostic category among involuntarily admitted patients, with 24.7 admissions per 100,000 population per year (see table in online supplement). Affective disorder was the most common diagnosis among voluntary patients, with 57.9 admissions per 100,000 population per year.

The mean age of involuntary patients whose admission was initiated by the police was 38.1±11.6 years (range 22–76), which did not statistically significantly differ from the age of those whose involuntary admission was not initiated by the police (40.5±14.5 years, range 18–86). Involuntary patients admitted through the police and those not admitted through the police did not differ significantly in gender, marital status, employment status, and diagnosis (see table in online supplement). Involuntary patients whose admission was initiated by the police were significantly more likely to be admitted under the “risk criterion” than those whose involuntary admission was without police involvement (11% vs. 5%, N=327; χ²= 8.94, df=2, p=0.011) (see table in online supplement).

Mean duration of admission was longer for involuntary patients whose admission was through police (66.4±154.3 days) than for those whose involuntary admission was without police (39.8±50.7 days) (t=2.49, df=369, p=0.013), but the mean number of days spent as an involuntary patient did not significantly differ between the two groups (24.5±26.9 days and 28.6±39.0 days, respectively), indicating that many patients were moved from involuntary to voluntary status during their admission.

We generated a multivariable model in which involuntary admission with or without police involvement was the dependent variable. Independent variables were those significantly associated with the dependent variable in bivariable testing, including the Mental Health Act 2001 criteria under which the patient was involuntarily admitted and duration of admission. Multivariable testing indicated that neither of these variables significantly and independently predicted whether the involuntary admission was initiated by police (see table in online supplement).

In this study of two admission units in Dublin, voluntary and involuntary admission rates for psychiatric care were lower than national rates. The proportion of admissions that were involuntary in the study units was higher than the national proportion, but the rate of involuntary admission per 100,000 population per year was lower, which may be attributable to several catchment area–and service-related factors. One of the admission units studied here, Tallaght University Hospital, which accounted for almost two-thirds of our data, is in a part of Ireland with the lowest number of psychiatric beds per 100,000 population (10). The impact of bed availability on decision making during admission is important to consider in examining our findings. The relatively high proportion of involuntary admissions in this area suggests that in areas with relatively fewer beds available, involuntary admissions are prioritized over voluntary admissions. Alternatively, community services and alternatives to inpatient care may be more readily available in the areas studied than in other areas. Future research could examine this relationship between bed availability and decision making in urgent psychiatric care in greater detail.

In this study, applications for involuntary admission were most commonly made by members of the public, including emergency department staff. Nationally, in 2018, family members most commonly made applications for involuntary admission (10). These differences suggest fewer social supports in our study areas and thus a greater tendency to access urgent psychiatric care through emergency services. In both our study data and national data, police officers were the second most common group that initiated involuntary admission.

Both in our study and at the national level, the most common diagnostic category among involuntary patients was schizophrenia, schizotypal, and delusional disorders, and the most common category among voluntary patients was affective disorders. The rate of voluntary admission of persons with affective disorders in our study (57.9 per 100,000 population) was almost half the national rate (113.6 per 100,000 population), suggesting a practice of managing affective disorder in the community setting in our study areas.

We found that patients whose involuntary admission involved the police did not differ significantly from involuntary patients whose admission did not involve the police in terms of age, gender, marital status, employment, or diagnosis. Results of bivariable analyses indicated that patients with police involvement differed significantly from those without police involvement in that the mean duration of their admission was longer and that they were more likely to be admitted under the risk criterion of Ireland’s Mental Health Act. On multivariable analysis, however, neither of these variables independently predicted whether police were formally involved in involuntary admission.

Although we did not detect differences in diagnoses between the two groups, some clinical factors found to be relevant in previous studies, such as violence and presenting symptoms, were not recorded in our data set. Nationally, the rate of applications for involuntary admission by police officers has risen since 2013, while applications by family members have fallen (10). Nonclinical factors, such as homelessness, social exclusion, or criminogenic risks, might underlie police involvement, but this requires further study. Criminogenic risks include both clinical factors (e.g., antisocial thinking and antisocial attitudes) and social determinants (e.g., unemployment and antisocial associates). There is a growing literature on criminogenic risks among persons with serious mental illness.

Our study examined the relationship between police applications and involuntary admission, a relationship that is important in the context of national and international trends indicating increased police involvement in involuntary admission for mental health care. We examined several potentially significant diagnostic or demographic variables and recorded ICD-10 diagnoses and compared our findings with national data (14). Ours is the first study of this topic in a contemporary Irish context, and it builds on the previous work of the Dublin Involuntary Admission Study by examining a more recent period and additional variables of interest (15).

Limitations of this study included the following: both catchment areas studied are urban-suburban, and findings therefore may not be generalizable outside these settings; only one diagnosis was recorded per admission (but comorbid conditions, particularly substance misuse and antisocial personality traits, may also have been relevant); level of social support and forensic history were not examined; only patients who were eventually admitted were included in the study (instances in which police officers initiated an involuntary application that did not end in admission are not recorded nationally and difficult to estimate); and some patients involuntarily admitted via other routes might have had informal police involvement at an earlier stage in their pathway to care. These unexamined but potentially relevant factors could be explored in future studies.

Patients whose application for involuntary admission was made by police and those whose application was not made by police did not significantly differ in diagnostic or demographic characteristics. Other features may distinguish these two groups, such as homelessness, social exclusion, or criminogenic factors. Future research could examine socio-demographic factors that have been associated with differential involuntary commitment by police in other countries. The rising rate of applications by police raises questions about the availability of timely and appropriate mental health care and about the drivers of police involvement in some cases but not in others.