Abstract

Résumé

The need to improve mental health services and outcomes for Canada’s diverse population has been identified as a national priority.¹ Gaining a comprehensive understanding of the patterns of mental health service use by different parts of the population is crucial for improving access to services and providing equitable care.

The term pathways to care is used to describe the often complex series of help seeking contacts made by patients and their family members in response to the symptoms of a mental illness.² The patterns of health service use by patients with FEP have been of particular interest because the delay between the onset of psychotic symptoms and the start of treatment, known as the DUP, has been associated with poor clinical and functional outcomes.^3–5

Both pathways to care and DUP may be influenced by ethnic, cultural, or racial group membership. A person’s cultural background may influence their beliefs about the cause of psychiatric symptoms and the most appropriate course of action.^6,7 There may be differences between groups in the services and supports that are available and accessible when help is needed.^8,9 Language barriers or a lack of information on available services may impede a person’s ability to seek help.¹⁰ In addition, some groups may view mainstream services as not culturally competent or untrustworthy, and they may fear differential treatment based on their perceived race.¹¹

We recently conducted a systematic review and meta-analysis of ethnic differences in pathways to care and DUP, focusing on FEP.^12,13 Our findings suggest that people of African and Caribbean origin are less likely to have the involvement of a GP on the pathway to care, and more likely to have police or criminal justice involvement and to have an involuntary admission.¹² Three larger studies from the United Kingdom reported a shorter DUP for ethnic minority groups^14–16; however, we found little evidence overall that differences in the pathway to care translated into differences in the DUP.¹³ Many of the studies included in our review were not designed or adequately powered to examine the differences linked to specific cultural, ethnic, or racialized group membership, although there have been notable exceptions.^14,15

Our study aims to investigate pathways to care and DUP in 2 multicultural cities in Ontario. In our study, we disaggregated the African-Caribbean population into those of African origin and those of Caribbean origin, reflecting differences between these groups in culture, history of immigration, and socioeconomic conditions.¹⁷ Compared with people of Caribbean origin, people of African origin in Toronto are more likely to be refugees, to live in poverty, to have lower levels of educational attainment, and to speak English as a second language.¹⁷ Each of these factors is associated with negative pathways to care among patients with FEP.¹⁸ We compared these 2 groups to a White group of European ancestry, as this is the most prevalent group in Ontario.

Methods

Results

During the recruitment period, 299 people agreed to be screened for the study, and 208 met the inclusion criteria. Among eligible people, 18 declined participation, 12 could not be contacted, 2 withdrew after consenting, and 5 had incomplete interview data. The final sample included 171 participants (Black-African = 41; Black-Caribbean = 40; White-European = 90). More than 85% (n = 150) of the sample came from Toronto, and 82% (n = 141) were receiving care from a hospital-based EIS. Participants had been in EIS for a median of 11 months (IQR 3 to 30) at the time of the interview. The sociodemographic characteristics of the sample are presented in Table 1.

Differences in the Pathway to Care

The proportion of participants who made first contact with each type of service provider is shown in Figure 1. Black-African participants had nearly 4 times the odds of having first contact with the ED, relative to White-European participants (OR 3.78; 95% CI 1.31 to 10.92); however, the difference was not statistically significant for Black-Caribbean participants (OR 2.42; 95% CI 0.85 to 6.89) (Table 2). Additionally, participants who had GP involvement on their pathway to care had lower odds of having first contact with the ED (OR 0.13; 95% CI 0.05 to 0.33), as were those whose help seeking was initiated by somebody other than family or friends (OR 0.10; 95% CI 0.01 to 0.64) (Table 2). Nearly one-half of all participants had contact with police or crisis teams on the pathway to care (Table 3). There was no evidence of differences across the groups (Black-African OR 0.94; 95% CI 0.44 to 1.97; Black-Caribbean OR 0.88; 95% CI 0.42 to 1.87) (Table 2).

Open in a separate window

Figure 1

The first contact on the pathway to care for people with first-episode psychosis from 3 ethnic groups in Ontario, n = 171

^a Includes psychologists, social workers, nurses, school counsellors, priests, alternative and (or) native healers, and community organizations

Table 2

Results of the multivariate logistic and linear regression models, with variables included in each model chosen using the Bayesian Information Criteria method²⁴

Variable	First contact with ED OR (95% CI)	Police involvement OR (95% CI)	GP involvement OR (95% CI)	Inpatient referral OR (95% CI)	Total contacts exp(β) (95% CI)	DUP exp(β) (95% CI)	Referral delay exp(β) (95% CI)
Ethnic groupa
African	3.78 (1.31 to 10.92)	0.94 (0.44 to 1.97)	0.70 (0.29 to 1.70)	2.04 (0.94 to 4.44)	0.83 (0.65 to 1.06)	1.10 (0.60 to 2.03)	0.97 (0.52 to 1.81)
Caribbean	2.42 (0.85 to 6.89)	0.88 (0.42 to 1.87)	0.17 (0.07 to 0.46)	3.33 (1.46 to 7.60)	0.77 (0.60 to 0.99)	1.49 (0.82 to 2.69)	1.55 (0.84 to 2.83)
Age at entry	—	—	—	—	—	1.08 (1.03 to 1.14)	—
Diagnosisa
Affective				2.61 (1.25 to 5.42)	1.29 (1.03 to 1.62)
History of violencea
Yes						2.05 (1.24 to 3.35)	1.94 (1.16 to 3.22)
GP involvementa
Yes					1.17 (0.96 to 1.44)
Initiation of help seekinga
Family and (or) friend	1.52 (0.48 to 4.81)		0.16 (0.05 to 0.49)
Other	0.10 (0.01 to 0.64)		0.06 (0.02 to 0.24)

Open in a separate window

DUP = duration of untreated psychosis; ED = emergency department; GP= general practitioner

^aReference group is European, nonaffective diagnosis, no history of violence, no GP involvement, and self-initiation of help seeking

Table 3

Indicators of the pathway to care of people with first-episode psychosis across 3 ethnic groups in Ontario, n = 171

Variable	European n = 90	African n = 41	Caribbean n = 40
Police or criminal justice involvement, n (%)	41 (46)	18 (44)	17 (43)
General practitioner involvement, n (%)	56 (62)	21 (51)	14 (35)
First contact with emergency department, n (%)	13 (14)	14 (34)	16 (40)
Referral source inpatient admission, n (%)	38 (42)	23 (56)	26 (67)
Total number of contacts, median (IQR)	6 (4 to 9)	4 (3 to 6)	4 (3 to 7)
DUP, months, median (IQR)	7 (2 to 34)	9 (3 to 28)	16 (6 to 46)
Referral delay, months, median (IQR)	4 (2 to 17)	4 (1 to 8)	3 (1 to 21)

Open in a separate window

DUP = duration of untreated psychosis

Only 35% of Black-Caribbean participants had contact with a GP on the pathway to care, compared with 51% of Black-African and 62% of White-European participants (Table 3). The difference across ethnic groups was significant for the Black-Caribbean group, who had much lower odds of having GP involvement in the pathway to care, relative to European participants (OR 0.17; 95% CI 0.07 to 0.46; Black-African OR 0.70; 95% CI 0.29 to 1.70) (Table 2). Participants whose help seeking was initiated by family members or friends (OR 0.16; 95% CI 0.05 to 0.49), or by other people (OR 0.06; 95% CI 0.02 to 0.24), also had lower odds of GP involvement on the pathway to care (Table 2).

Participants in the White-European group had a median of 6 total contacts on the pathway to care (IQR = 4 to 9), whereas both the Black-African and the Black-Caribbean groups had a median of 4 total contacts (Black-African IQR = 3 to 6; Black-Caribbean IQR = 3 to 7) (Table 3). The difference across ethnic groups was significant for Black-Caribbean participants only (exp[β] = 0.77; 95% CI 0.60 to 0.99; Black-African exp[β] = 0.83, 0.65 to 1.06) (Table 2). Additionally, people who had a diagnosis of an affective psychotic disorder had a greater number of total contacts (exp[β] = 1.29; 95% CI 1.03 to 1.62) (Table 2).

The proportion of participants who were referred to EIS by each type of service provider is shown in Figure 2. Black-Caribbean participants had more than 3 times the odds of being referred to an EIS by the inpatient unit (OR 3.33; 95% CI 1.46 to 7.60), with no significant difference for Black-African participants (OR 2.04; 95% CI 0.94 to 4.44). Having a diagnosis of an affective disorder was an independent predictor of referral from the inpatient unit (OR 2.61; 95% CI 1.25 to 5.42).

Open in a separate window

Figure 2

The referral source to early intervention services people with first-episode psychosis from 3 ethnic groups in Ontario, n = 171

^a Includes psychologists, social workers, nurses, school counsellors, priests, alternative and (or) native healers, and community organizations

Differences in Duration of Untreated Psychosis and Referral Delay

The Black-Caribbean group had a median DUP of 16 months (IQR 6 to 46), compared with a median of 7 months for the White-European group (IQR 2 to 34) and 9 months for the Black-African group (IQR 3 to 28) (Table 3). These differences were not statistically significant (Black-African exp[β] = 1.10; 95% CI 0.60 to 2.03; Black-Caribbean exp[β] = 1.49; 95% CI 0.82 to 2.69) (Table 2). Older participants also tended to have a longer DUP (exp[β] = 1.08; 95% CI 1.03 to 1.14). People who had a history of violent or aggressive behavior had a DUP that was about twice as long as those who did not (exp[β] = 2.05; 95% CI 1.24 to 3.35) (Table 2).

The median referral delay was four months for the White-European and Black-African groups (IQR 2 to 17 and 1 to 8, respectively) and 3 months for the Black-Caribbean group (IQR 1 to 21) (Table 3), and this difference was not statistically significant (Black-African exp[β] = 0.97; 95% CI 0.52 to 1.81; Black-Caribbean exp[β] = 1.55; 95% CI 0.84 to 2.83) (Table 2). People with a history of violent or aggressive behaviour had a referral delay that was nearly twice as long as those with no history of such behaviour (exp[β] = 1.94; 95% CI 1.16 to 3.22) (Table 2).

Discussion

Our findings suggest that people in the Black-Caribbean group experience more aversive pathways to care, with an increased odds of referral to EIS from an inpatient admission, a decreased odds of GP involvement on the pathway to care, and fewer total contacts. People in the Black-African group had an increased odds of making first contact with an ED. We did not find significant differences between the ethnic groups in DUP or referral delay.

Few studies have examined ethnic differences in pathways to care in FEP, and ours is the most extensive Canadian study to date. Our findings are strengthened by the use of disaggregated analyses for African- and Caribbean-origin groups. We also measured the pathways to care using instruments that have been used in international studies across different cultures and service contexts.^25–28 Prior findings on this topic have been inconsistent,¹⁸ leading some to suggest that observed differences between groups arise only after the first episode of psychosis,^29,30 possibly owing to negative experiences with services or differences in the degree or type of stigma from within ethnic communities.³⁰

We found that Black-African participants had a greater odds of first contact with the ED. Most people in this group were first-generation immigrants, and the heavy reliance on the ED at first contact may reflect a lack of knowledge regarding the availability of services and how to navigate them.¹⁰ Indeed, immigration status has been previously found to be associated with the propensity to use emergency services by people with FEP.^15,31 We also found that the involvement of the GP on the pathway to care was an independent predictor of a reduced odds of making first contact with the ED, which is consistent with prior Canadian findings.^32,33

We found little evidence to suggest that there are ethnic differences in having police involvement on the pathway to care, in contrast to prior reports from the United Kingdom.^15,29,34,35 Canadian studies have found a trend toward an increased likelihood of police involvement for Black racial groups, but these findings did not reach statistical significance.^33,36 Jarvis et al³⁷ in Montreal found that Black patients with psychotic disorders in the ED were more likely to be brought by police or ambulance; however, their study looked at psychotic disorders generally rather than FEP specifically. These differences across different countries, and even across different jurisdictions within the same country, could reflect differences in mental health legislation and local laws. Also, we neither recruited participants from forensic services who would be expected to have the most extensive degree of police involvement, nor did we measure the frequency of experiencing other types of coercive pathways, such as involuntary hospitalizations or the use of seclusion or restraints.

Black-Caribbean participants in our sample had much lower odds of having GP involvement on their pathway to care. This is consistent with prior studies from the United Kingdom,^15,34 although Black-African people have also been found to have a reduced likelihood of GP involvement.³⁴ We speculate that this could be reflective of more negative attitudes toward primary care from ethnic minority groups, disparities in access to family physicians, or a greater tendency to delay help seeking to the point where emergency services are needed. Involvement of the GP on the pathway to care has been associated not only with a reduced frequency of contact with emergency services^32,33 and of compulsory admission and contact with police³⁰ but also with an increased number of total contacts and a longer treatment delay.^31–33

We found that Black-Caribbean participants were more frequently referred to an EIS by the inpatient unit. This differential referral for the Black-Caribbean group could be a consequence of the reduced odds of GP involvement for this ethnic group, although this variable was not selected for inclusion in our optimal regression model. Alternatively, it could represent a more severe clinical presentation necessitating an inpatient admission, although there were no differences across the groups in the median number of inpatient admissions (data not shown), and prior research has not found evidence of ethnic differences in clinical presentation at the first contact for FEP.³⁸

Our findings indicate that people in the Black-Caribbean group may have a longer DUP than those in the White-European group, although this did not reach statistical significance. The 2 previous Canadian studies looking at differences in DUP for African and Caribbean people reported no evidence of differences compared with the White-Canadian population^36,38; however, the studies lumped these 2 ethnic groups together as Black, which would obscure any differences between specific ethnic groups. Previous studies from the United Kingdom suggest that Black-African patients have a shorter DUP when compared with White-British patients, with no difference for the Black-Caribbean group.^14,15 These discrepant findings could be explained by differences in the social context of the African and Caribbean communities across the 2 countries. Interestingly, we did not find a difference between the ethnic groups in the time between the first contact after the onset of psychotic symptoms and entry into an EIS. These findings point toward ethnic differences in the sub-phases of DUP, specifically, the help seeking delay from the time of symptom onset to first contact with health services, and the referral delay from first contact to entry into an appropriate treatment program.^39–41 Research from the United Kingdom found that contact with mental health services came later for Black-Caribbean patients, but, once contact was made, this group received psychiatric care sooner.³⁵

Limitations

The patterns we observed may not be representative of all people who use EISs. We were unable to attain complete case ascertainment from the programs included in our study, as our recruitment strategy relied on referrals from the staff at participating EISs. Consequently, we do not know how many eligible people were missed from each program, how many were missed from programs that were not participating in our study, and how many people disengaged from EISs prior to being referred. We also do not have demographic information on people who refused screening, and therefore, we do not know if this was differential by ethnic group. Our sample is composed of people who have been successful at obtaining specialized care, and it is possible that the likelihood of receiving care from other health and social service providers differs across the 3 groups.

Our data were measured retrospectively, and thus are subject to recall errors and biases; however, we corroborated information with other sources, including medical records and family members, in an effort to improve accuracy. We did not assess the reliability of our data collection methods, and therefore, we cannot eliminate the possibility of systematic differences across the interviewers, or differential assessment methods across the ethnic groups. We did not have information on several factors that may impact the pathway to care, such as social support, substance use, and symptom severity, so residual confounding may remain. We also cannot rule out the possibility that observed ethnic differences are moderated by age, gender, or socioeconomic status, which has been reported previously,^29,35,42 as we did not have sufficient power to examine interaction terms in our analyses.

A further limitation is the categorization of the ethnic groups that we used. The categories used by the TDSB are pragmatic and attempt to reflect the identities of different groups in Toronto. It mixes country of origin and race, and does not have a clear conceptual basis. The use of these groups is informative for those involved in the development of services, although it is only one method to consider group categorization.

Conclusions

In summary, our findings suggest that there are significant differences in the pathways to EIS for psychosis for people of African and Caribbean origin in our Canadian context. There is some evidence to suggest that GPs may play a key role in mitigating these aversive pathways to care and that they could be a potential target for reducing ethnic disparities in patterns of service use. Other strategies for improving the pathways to care include hastening the care pathways, removing negative pathways to care, or adding beneficial alternative routes,⁴³ and examples of interventions aimed at improving the pathway to care for ethnic minority groups are available.⁴⁴ However, careful consideration will need to be given to the degree that various strategies are acceptable and effective for different ethnic groups to prevent further exacerbating the disparities in access to care.

Acknowledgments

Abbreviations

References