Medical coercion refers to the use of force, threats, or other means (
1,
2) to gain compliance from another person who is deemed an imminent danger to themselves or others (
3). Today, the overall rates of involuntary admissions and coercive interventions in psychiatric hospitals have increased internationally because of mandated community treatments (e.g., by the Mental Health Act of 1983), expansion of collaborative care models, and renewed institutionalization of psychiatric patients (
4). In Canada, the prevalence of involuntary admissions was as high as 77% in 2013, and these rates have increased since 2009 (
5,
6). Moreover, certain groups, particularly ethnic minority groups of African or Caribbean descent (
7,
8), are at greater risk for involuntary admission; this discriminatory treatment may be due to excessive application of the “danger” criterion for psychiatric commitment among members of these groups (
9–
11).
Examples of racial bias have already been documented in the research literature. Researchers in the United Kingdom have suggested that discrimination and racial bias may lead to higher rates of incorrect decisions to admit patients involuntarily (
12,
13). Meanwhile, researchers in the United States have argued that discrepant rates of psychosis may be due to misdiagnosis based on cultural differences and stereotypes about Black people (
14,
15). Stereotypes may include being perceived as agitated and aggressive when psychotic and being in need of forced treatment or care when ill (
14). Examples of cultural differences include nonmedical explanatory models of illness and healing as well as different health care beliefs (
13). Black patients may be at greater risk for coercive intervention because of such cultural misunderstandings, in combination with discrimination, racial bias, stereotypes, and assessment bias (i.e., whether a person is deemed a danger to self or others). Little is known about whether these findings also hold true in psychiatric settings in Canada (
16–
18) and what the implications may be for Black patients who are so affected.
Previous Canadian studies on ethnoracial group differences in negative pathways to care and coercive intervention practices have produced mixed findings. Some studies have shown that Black patients with psychosis are more likely to be coercively referred to emergency services via police or ambulance in Toronto (
16,
19) and Montreal (
20). In contrast, other studies have shown no correlation between ethnicity and negative pathways to care (
21) or for involuntary admission (
22). These discrepant findings may be due to differences in methods, patient populations, and local health service systems (
21). Therefore, in this study, we aimed to clarify this confusion by looking at the patient population with first-episode psychosis (FEP) in Montreal to determine which findings would be replicated.
Using a retrospective sample, we sought to determine whether Black patients with FEP were at a higher risk for coercive referral and coercive intervention by psychiatric services than non-Black patients with FEP. We hypothesized that Black patients with FEP would be at higher risk for coercive referrals and coercive interventions in psychiatric services than non-Black patients with FEP. Using quantitative methods to assess data retrospectively from patient charts, we examined the relationship between ethnoracial status and coercive interventions. Given the limited evidence on factors that place patients with FEP at risk for coercive referral and intervention, in this study we fill this knowledge gap by assessing the role of ethnoracial status in involuntary hospital admissions and by exploring the implications for the Black community.
Methods
Setting
The First Episode Psychosis Program (FEPP) is located in the Institute of Community and Family Psychiatry of the Jewish General Hospital in Montreal. The FEPP serves an urban inner city, highly diverse immigrant catchment of mostly low- to middle-income residents. The program has been receiving referrals since 2008. About half of the patients are from visible minority groups, meaning non-White, and one-third are from immigrant families. Two-thirds of the patients are male because of more frequent early onset of psychosis among men than among women (
7,
16,
20). Because referrals to the FEPP may be received outside the designated catchment area, patients may not be representative of the larger community. The research protocol and consent forms were authorized and approved by the research ethics board of the CIUSSS du Centre-Ouest-de-l’Île de Montréal.
Chart Review
To test the study hypothesis, we collected retrospective data from patients referred to the FEPP from January 14, 2008, to December 31, 2018, via chart review (N=279), permitting data extraction and analysis on a hospital server with electronic data abstraction forms.
Inclusion criteria for the study were applied as per routine clinical practice at the point of entry to the FEPP. These criteria included between ages 16 and 30 years; treatment with antipsychotic medication for <30 days; presence of psychotic symptoms for >3 consecutive days (i.e., hallucinations; delusions; paranoid ideation; disorganized speech, thought, or behavior; or prominent negative symptoms); no severe substance use that could impair participation in the program; no autism spectrum disorder; no intellectual disability; no serious medical conditions, such as epilepsy, that could account for the psychotic symptoms; and patient living within a 30-minute travel radius of the Jewish General Hospital.
Data Collection
Ethnoracial assignment.
Patients’ ethnoracial identity was embedded in physician notes and extracted from hospital files to determine participants’ ethnoracial status. Participants who were identified as Black in their patient chart were grouped together, and those who were not identified as Black were grouped together. For patients whose ethnoracial identity was not indicated, an ethnicity assignment procedure based on categories from Statistics Canada (
23,
24), including one’s country of birth, languages spoken, immigration status, religion, and family name, was used as a means to determine most probable ethnoracial identity. The term “visible minority” was used to refer to ethnoracial minority groups because it is a governmental designation by the Canadian government (
24). Someone who was marked as a multiple visible minority individual, meaning the patient was part of two or more minority groups, was categorized as “mixed.” Those who were mixed or biracial, in which one parent was of European origin and the other was from a visible minority group, were also considered as visible minority individuals.
For the purpose of this study, East Asian, South Asian, and Southeast Asian were pooled together under the racial category “Asian” to obtain larger numbers for statistical comparisons. Participants from North Africa (Morocco and Tunisia) were categorized as “Arab” because of the cultural, political, linguistic, and religious similarities to countries in the Middle East and West Asia (
25). This procedure resulted in five ethnoracial groups: White (European descent), Black (Caribbean and African descent), Latinx (Latin American descent), Arab (Middle Eastern descent), and Asian (East, South, and Southeast Asian descent). For purposes of comparison, three patient groups were created from the five ethnoracial groups: White, Black, and non-Black visible minority groups. Non-Black visible minority groups included Arab, Asian, and Latinx ethnoracial groups. These ethnoracial assignment procedures were based on Statistics Canada categories (
23,
24) and have been used by other Canadian studies (
17,
26,
27).
Coercive referral.
To determine whether patients were coercively referred, we reviewed hospital charts to reveal whether patients were brought to emergency services via police, ambulance, or court order.
Coercive intervention.
To determine whether patients experienced a coercive intervention, we extracted legal and medical variables from formal items in legal documents and descriptions in physician notes. Legal variables represented documented coercive practices rendered by the criminal justice system or other government bodies. Medical variables represented documented coercive practices rendered by medical doctors and staff.
Legal variables included
garde preventive (i.e., preventive confinement completed by a physician),
garde provisoire (i.e., temporary confinement for psychiatric assessment rendered by the Court of Quebec),
garde en établissement (i.e., confinement in an institution rendered by the Court of Quebec), forced outpatient treatment order (i.e., receiving ongoing treatment rendered by the Court of Quebec),
Tribunal administratif du Québec involvement (i.e., the patient is declared unfit to stand trial or is not criminally responsible and is placed in the follow-up of the tribunal), police contact (i.e., arrested by police, pressed charges), and other court involvement such as immigration detention (i.e., arrested at border) and youth protection (i.e., Department of Youth Protection). Other court involvement was included in data analyses because patients referred to the FEPP are as young as 16 years and may be legally obligated to receive treatment under these government regulations. Studies have shown the negative effect of detention on the mental health of migrants (
28) and of foster care on the mental health of youths (
29). Medical coercion variables included use of seclusion, physical restraints, “code white” (show of force when a patient poses as a threat to self or others), and urgent intramuscular antipsychotic.
Data Analysis
We calculated descriptive statistics for each demographic and study variable by using the statistical software JASP, version 0.9.2 (
30). Chi-square tests were conducted to determine whether a statistically significant effect was present and whether differences in cell frequencies warranted interpretation. Variables such as age, gender, ethnoracial status, substance abuse, and violent or threatening behavior predicted the dependent variable, which was coercive referral or intervention depending on the model. In total, six models were run.
Results
Of the 279 chart reviews, 71 were removed for the following reasons: missing data (N=22), lost medical record numbers (N=24), no psychiatry related charts at the time of referral (N=24), and duplication (N=1).
Table 1 summarizes the characteristics of the 208 patients with FEP in the final sample. The mean±SD age at referral was 22.9±4.5 years (range 16–40). Overall, 96 (46%) were White, 42 (20%) were Black, and 70 (34%) belonged to non-Black visible minority groups according to the definition provided above. More than two-thirds of the sample were male (N=141, 68%), and most participants (N=191, 92%) were single.
Chi-Square Analyses
Three-group (White, Black, and non-Black visible minority groups) chi-square analyses were conducted to identify disparities in coercive referral and coercive intervention (
Table 2). The results indicated that Black patients with FEP were more likely to be coercively referred (observed=21, expected=14) than would be expected by chance alone (χ
2=9.24, df=2, p=0.010), whereas White patients with FEP (observed=23, expected=32) were less likely to be coercively referred. No evidence was found that non-Black visible minority patients with FEP (observed=25, expected=23) were significantly more likely to be coercively referred.
Other important findings showed that Black patients with FEP were significantly more likely to experience any coercive intervention involving medical and legal measures (observed=34, expected=25; χ2=9.21, df=2, p=0.010), whereas White patients with FEP (observed=54, expected=58) and non-Black visible minority patients with FEP (observed=38, expected=42) were less likely to experience any coercive intervention involving these measures. Levels of medical coercive intervention were not differentially distributed among patients with FEP, but Black patients with FEP were significantly more likely to experience legal coercive interventions (observed=34, expected=25; χ2=10.92, df=2, p=0.004) than were White patients with FEP (observed=52, expected=56) and non-Black visible minority patients with FEP (observed=36, expected=41).
Logistic Regression Analyses
Two-group regression analyses were conducted in which the Black patient group was compared with a combined White and non-Black minority reference group as a factor variable.
Table 3 displays the results of the logistic regression analyses, with coercive referral to emergency services (yes or no) via police, ambulance, or court order as the dependent variable. Age, gender, ethnoracial group, presence of violent or threatening behavior, and presence of substance abuse were the control variables. Logistic regression models showed that those who were older (odds ratio [OR]=1.09, 95% confidence interval [CI]=1.01–1.17, p=0.025) or perceived to be violent or threatening (OR=6.28, 95% CI=3.17–12.46, p<0.001) were significantly more likely to be coercively referred to emergency services.
Other results showed that those who were older (OR=1.15, 95% CI=1.06–1.24, p<0.001), Black (OR=2.72, 95% CI=1.11–6.63, p=0.028), or perceived as violent or threatening (OR=5.01, 95% CI=2.34–10.74, p<0.001) were more likely to receive any coercive intervention when comparing the Black groups with a non-Black minority reference group (
Table 4). Logistic regression analyses also showed that being older (OR=1.14, 95% CI=1.05–1.23, p<0.001), Black (OR=3.01, 95% CI=1.23–7.35, p=0.016), or perceived as violent or threatening (OR=5.73, 95% CI=2.67–12.29, p<0.001) predicted higher likelihood of legal coercive intervention when comparing all patient groups (
Table 5). Medical coercive intervention was not significant when assessed separately as an independent variable.
Discussion
The purpose of this retrospective study was to determine whether Black persons with FEP in Canada were at a higher risk for coercive referral to psychiatric services and coercive psychiatric intervention than non-Black individuals with FEP. Findings from the chi-square analyses revealed that Black persons with FEP experienced more coercive referrals via police, ambulance, and court order than did other ethnoracial groups. These findings suggest that racial bias may be at play given pervasive negative stereotypes of Black men being violent (
31) and the frequent encounters that members of the Black community have with the police (
32). More information is needed to determine how authority figures assess danger, especially among Black patients with FEP, and how bias may manifest in institutional settings to produce disproportionate numbers of coercive referrals in this group.
Other important findings were that Black persons with FEP were at higher risk for coercive intervention, particularly legal coercion with court involvement, whereas persons with FEP from White and non-Black visible minority groups were less likely to experience coercive interventions overall. The significance of legal coercion, and not medical coercion, among Black people with FEP suggests that racial bias may exist within the judicial system, especially when it is brought to bear on young Black men and women with mental health issues. The significant combination of medical and legal coercion suggests that an interaction of psychiatry and the law may be present that is especially problematic for Black people with psychosis. Given the fact that mental health legislation is based on the criterion of “danger” and that Black people, particularly Black men, are often deemed to be dangerous (
9), courts may render more coercive measures to Black patients to manage risks to a public that may perceive Black persons as dangerous (
8,
10).
Implications
In this study, we have addressed the gap in the psychiatric and clinical literatures by examining the relationship between ethnoracial status and coercive intervention in retrospective patient charts. Canadian provinces and territories do not routinely collect ethnoracial data, perhaps to avoid labeling people by race-ethnicity (
33). The practice of collecting such data in countries such as Canada, the United States, and the United Kingdom is complex and controversial (
34,
35). Therefore, in this study, statistics on race-ethnicity were collected to examine potential differences between patient groups in Canada. This study also could help remind clinicians to carefully assess how they make these judgments and to exercise caution to prevent harm to Black patients. More medical education and training is needed to encourage culturally sensitive care, particularly among Black patients because of their long history of receiving inferior care in psychiatry (
36,
37). This article calls for the need to redress institutional cultures and norms, especially within the medicolegal system in which Black patients are disproportionately detained for coercive treatment compared with other racial groups. Policy makers must carefully assess the justification for coercion in balance with patient welfare and examine whether racial bias may be present in court proceedings.
Limitations
Several limitations were present in this study. For the quantitative sample, data from 71 individuals were removed from data analysis. Most of these were White persons with FEP, which was acceptable because sufficient numbers of individuals in this group remained to conduct the statistical analyses; however, bias could have been introduced as a result of these removals. This study focused on Black Canadians, and it is important to note that any differences between African and Caribbean groups were lost when these two groups were combined, because the scale of this study did not allow for separate data analyses of these two groups. The method of assigning ethnoracial identity from medical records was also a limitation because relying on physician notes may have resulted in misclassification or poor data quality. This method also involved an external attribution of identity, which may differ from participants’ perceptions and choices regarding their ethnoracial identity. However, this external attribution was appropriate to study potential clinician and system biases.
Participants were also limited to those who lived within a 30-minute travel distance to our hospital because of travel distance limitations enforced by the hospital. Findings therefore cannot necessarily be generalized to the larger community. Additionally, some “not noted” sociodemographic and clinical characteristics were transcribed in patients’ charts as a result of poor recording by health professionals. The logistic regression models had relatively small samples sizes, so we had to judiciously decide which control variables to use. For these reasons, migrant status, which is often associated with negative pathways to care (
38), along with primary language, previous hospitalizations, duration of untreated psychosis, and suicidal ideation, was not included in data analyses. These variables were of poor quality and were therefore excluded.
Finally, all comparative findings among the groups studied were correlational, so causality could not be inferred. Although Black persons with FEP were more likely to be coercively referred to mental health services than were other ethnoracial groups, it cannot be definitively concluded that the reason for this discrepancy was because these persons were Black. Intervening variables, such as poverty and educational status, may have accounted for the findings had the sample size and data collection permitted extended analyses. However, the findings of this study do suggest that ethnoracial status may play a role in mental health treatment. Moreover, despite these limitations, this study contributes to the literature on patients with FEP and coercion in Canada, which so far has been limited.
Conclusions
The findings of this study and those of the existing literature have shown that members of different ethnoracial groups are differentially treated in psychiatric care (
7,
14,
19,
39). This observation suggests that systemic racism may be an international problem embedded in medical and legal institutional frameworks of nations and health care systems. Racial discrimination and prejudice are important public health issues and must be addressed to ensure that medical, psychiatric, and legal interventions are implemented according to best practices for all.