Pathways Through Early Psychosis Care for U.S. Youths From Ethnically and Racially Minoritized Groups: A Systematic Review

Of the 1,699 potentially relevant studies, we reviewed 342 full-text articles in detail, 18 of which satisfied our eligibility criteria (7, 29–45) (Table 1) (see flowchart in the online supplement). Each article was then reviewed for domain-specific results. Most studies addressed more than one domain. Most of the 18 articles (N=16, 89%) were published within the past 5 years. More than half (N=11, 61%) had an explicit focus on race and ethnicity (as defined by the study authors), and the remainder (N=7, 39%) analyzed race or ethnicity among other factors. The proportions of ethnically and racially minoritized groups in the samples of these studies ranged from 21% in one study to ≥90% in four studies. Among the studies having more than one racial group in the sample population, eight included information about participant ethnicity. Four used “Hispanic” or “Latino” as a race category, and four used it in conjunction with race, creating categories of “non-Hispanic White,” “Hispanic White,” and “non-Hispanic Black.” In most studies, results fell into more than one of the four domains, and we discuss them in each relevant domain below.

Twelve of the 18 articles discussed early psychosis and utilization (7, 29, 31, 34, 37, 38, 40–45). None were pilot studies or randomized controlled trials (RCTs) that directly tested interventions for subpopulations of individuals receiving services for first-episode psychosis (FEP).

In one study, African American participants who had begun using coordinated specialty care (CSC) services were less likely at baseline to have private insurance, more likely to be homeless, and had significantly poorer quality of life, compared with Caucasian participants (41). Black and Latino patients were reluctant to use the mental health care options of pharmacologic treatments (7) and had lower rates of outpatient behavioral health care utilization (including outpatient visits with mental health professionals and claims for psychotherapy or pharmacologic interventions), compared with White participants (34). Hispanic participants were less likely to utilize inpatient services than were non-Hispanic White participants (43). Instead, individuals sought accessible treatments focused on building social relationships and helping them thrive in school- or work-related activities (7).

When Black participants did use services, they disengaged earlier from treatment programs, despite higher rates of outreach efforts to retain them in FEP treatment (42, 45). Even so, findings of other studies suggested that race did not predict the therapeutic alliance in FEP care (29) or treatment satisfaction (31). Additionally, when analyses factored in and matched intervention services by race, early intervention services continued to reduce overall health care costs (37), which can be a factor in utilization (7).

Six studies (30, 32, 35, 36, 39, 44) met our inclusion criteria for the outcomes domain. All were focused on outcomes, and none were about recovery. Most studies (N=4, 67%) included data from a naturalistic sample of youths who were followed up after an inpatient hospitalization or initiation of outpatient treatment (32, 35, 36, 39). Only two (33%) presented data from an RCT—both from the same national RCT (30, 44).

The studies presenting results in this domain addressed disparate research questions, and few common outcomes were reported across them, making it difficult to compare their findings. Studies focused on individuals undergoing treatment as usual (36, 44) or participating in CSC (39) and other early intervention programs (32, 35). In general, it appeared that ethnoracially minoritized youths experienced worse outcomes. Symptoms data were the most commonly reported outcome (in N=3 articles, 50%), with two articles each reporting on positive symptoms (36, 44) and on negative symptoms (32, 36) and one each on affective symptoms (36) and thought disorder (44).

Three articles reported differences between African American and Caucasian participants (32, 36, 44). In a study of individuals hospitalized for FEP (32), social functioning of Black participants was negatively associated with negative symptoms, depressive symptoms, and general psychopathology. In a 1-year follow-up study of inpatient and outpatient participants who received care as usual in the community (medication management and community-based supportive individual or group therapy), no significant racial differences in positive symptoms were noted (36). However, African American participants had significantly less improvement, compared with Caucasian participants, in thought disorder, negative symptoms (avolition and anhedonia in particular), affective symptoms, and global functioning. In another study (44), which compared 2-year outcomes among participants in a CSC program and in treatment as usual (nonspecialized services in the community), Black participants in the treatment-as-usual group had significantly higher scores (indicating worse outcomes) on measures of positive symptoms, disorganized thoughts, uncontrolled hostility, and anxiety-depression but not on negative symptoms, compared with non-Hispanic White participants in the treatment-as-usual group. For participants who received CSC, symptom outcomes were similar in all groups.

Two studies (35, 39) reported differences between non-Hispanic White and Black participants, and one also included Hispanic and Asian participants. In one study (35), White participants were more likely than other participants to indicate that they had “met all goals” (e.g., symptom remission or an ability to manage any existing or residual symptoms, and full-time employment). One study of CSC clients (39) found that after 1 year, non-Hispanic White and non-Hispanic Black participants were more likely to report persistent cannabis use, compared with Hispanic and Asian participants, and that persistent cannabis use was associated with worse symptoms for all groups, compared with reduced use.

Two articles discussed the relevance of clinic-level characteristics (29, 33). Neither were pilot studies or RCTs. In one RAISE trial (29), race did not significantly predict the quality of the therapeutic alliance, as measured by video ratings of individual resilience therapy. More severe positive symptoms, less severe negative symptoms, and female gender were significantly associated with a better therapeutic alliance (29). Provider characteristics, including race, were not investigated.

In the other study (33), efforts to recruit Latino research participants with early psychosis and their family members were described and thematically summarized. The researchers found that using bilingual and bicultural clinicians and administrative staff, explicit efforts to focus on “relationship building,” and procedures designed to engage family members were important elements in the recruitment of participants and families who decided to participate in the research study after the outside clinical care that led to their referral had ended. Relationship building included help explaining mental health systems, connecting with families while they were still in treatment, and supporting general resource needs.

We identified six relevant studies focused on understanding family-level factors in the pathway through care (7, 33, 38, 42, 44, 45). None of these interventions were for families receiving services for FEP; rather, the focus was on understanding family perspectives and ways to enhance family engagement and level of participation in services.

Three studies (7, 33, 38) focused on the initial part of the pathway through care, identifying issues that were important to families and their experiences. One study examined young people experiencing early psychosis and their key supporters after an initial hospitalization but before they engaged in treatment (7). That study identified and ranked factors that youths and key supporters indicated as important when making treatment decisions. The top three identified factors included a desire that the young person “get back to normal,” with a focus on work and school goals; concerns that the mental health care on offer was insufficient; and distress related to police involvement in the pathway to care.

In a cohort of Latino persons with a first episode of psychosis and their caregivers, poor psychosis literacy was observed among young persons, whereas caregivers showed greater psychosis literacy (38). Moreover, young persons with early psychosis and their caregivers who reported having a greater knowledge of psychosis and who attributed those symptoms to serious mental illness reported being more likely to seek out professional help (38). Engaging caregivers was important, because they were more likely than the young person with psychosis to recognize that the individual had a significant mental health problem and were also more likely to seek help. In addition, sustained engagement occurred more often when providers ensured that the information they gave to patients and their families about the nature of follow-up care directly addressed immigrant Latino families’ fears about rehospitalization, financial burden, or involvement of law enforcement (33). The authors noted that delivering information about psychosis in the primary language of the main support person, and that directly addressed concerns and misconceptions and helped with concrete needs, seemed essential for initial and continued engagement.

Three additional studies focused on family treatment and consistently found that Black families engaged less in care (42, 44, 45). In OnTrackNY and NAVIGATE programs, providers had less success engaging Black families and providing ongoing psychoeducation, compared with White families (44, 45). In another study, even though teams made significantly more outreach attempts to Black families than to White families, Black families were scheduled for significantly fewer psychoeducation sessions (42).

Even though we conducted our searches with many culture-level terms, we did not find any studies focused on early psychosis at the cultural level. In many studies, family and culture seemed to be conflated such that “family,” and its identity as being from an ethnoracially minoritized group, was a stand-in for “culture.”

Several of the reviewed studies highlighted problems with engaging ethnoracially minoritized youths in early psychosis care (34, 43). Structural barriers to care (e.g., cost of care and police involvement in the pathway to care) may cause many Black youths and family members to avoid treatment (7, 40), and even when Black youths initially engage with care, they disengage earlier than their White counterparts (42, 45). Studies estimate that 30% of Black people, 20% of Latino people, and 23% of Indigenous people avoid any medical care because of experiences of ethnoracially based discrimination in health care settings (16, 46).

Decisions to engage in or disengage from care may hinge on access and availability of services, stigma, and financial strain. Findings from one study suggested that race and ethnicity did not predict the strength of the therapeutic alliance (29), and results from another study suggested a lack of identifiable differences across diverse groups in terms of treatment satisfaction (31). Early intervention also reduced health care costs across all racial groups (37), an outcome that was thought to be important for treatment engagement (7, 33). Designing, implementing, and ensuring access to affordable services that focus on improving engagement are important for mental health equity and may enhance treatment utilization (7, 31).

Research with Latino and Black youths and their families, “supporters,” or “key supporters” suggested that more culturally sensitive care may better engage people in treatment. Two studies highlighted that families from minoritized groups have fear of the police (7, 33), which are often reported to be part of emergency pathways to care (5, 47). This observation is consistent with those of other literature (3, 48, 49). Police involvement has a negative impact on the mental health of Black Americans (50) and contributes to health inequities for Latino populations (51, 52), including those who are undocumented (27, 38, 53). Black youths and their supporters feared being perceived as a threat (3), and undocumented Latino individuals feared deportation (27, 38). Police violence, because it is state sanctioned, may be especially distressing (8, 53). The role of the police in the pathway to early psychosis care needs further attention.

Another engagement issue involved the inability of early psychosis programs to connect with Black families to provide ongoing psychoeducation (44, 45), even when extra outreach attempts were made (42). Delivering information about psychosis while addressing misconceptions and concrete needs in a language that the primary supporter could understand was important for engaging Latino youths and families (33)—and likely other cultural groups. Black families may also need information tailored to their questions and concerns, as well as more flexibility on when and where family education is provided to accommodate their schedules and facilitate access.

Our review found limited discussion focused on ethnoracially minoritized status, even when samples consisted primarily of persons from ethnoracially minoritized groups. We also found no studies about pathways through care for early psychosis that focused on Asians (although one study included an Asian sample), Native Americans, refugees, asylum seekers, and other groups that experience discrimination in the United States, resulting in negative health and mental health impacts. For example, Muslim individuals would be classified as White but may not identify that way (54, 55), and they also face numerous barriers to representation in health care studies (56). Similarly, the studies we reviewed had difficulty identifying whether “Hispanic” was a racial category or an ethnicity; some reported the percentage of Hispanic participants separately from race but did not incorporate it into analyses, whereas others created a Hispanic racial category if participants self-identified as Hispanic. Survey data suggest that when asked about their race in an open-ended format, Latino individuals often use pan-ethnic terms (e.g., Hispanic or Latino) or responses that linked their racial origin to the country or region of their ancestors (57). More conversation about these issues is important.

Moreover, studies focusing on non–English-speaking populations were almost completely absent from the eligible literature. Services and written materials were rarely available in a language other than English, and there was little discussion of efforts to recruit clinical staff who look and speak like their clients or to place services and treatment milieus in the communities they seek to reach. The most frequent form of language accommodation was the use of a translator, which is not ideal for helping families feel affirmed and welcomed in a new treatment context or to help them understand how to navigate and manage complicated cultural formulations around stigma and persons deemed “mentally ill” (58–60).

To the best of our knowledge, almost no research has been done on how clinic-level factors, such as the race or ethnicity of providers, affect care utilization or outcomes. Even the modest literature on the therapeutic alliance in psychosis omitted individual-level characteristics of the provider beyond age and sex or gender (61).

Culture-level moderators were limited to research on families, rather than the contours of the broader social context that affect everyday life. Moreover, although stigma among Black and Latino families is well documented, we found no research discussing how to address stigma within the family in culturally relevant ways. In fact, we found few evidence- and family-based interventions. One family-based intervention that has not yet been widely implemented for Latino families includes a focus on family collectivism and spiritual beliefs and practices (62), which may be more culturally appealing to families and thus more effective.

We note three primary limitations of this review. First, using “first-episode psychosis” and “early psychosis” to identify young people with psychotic symptoms typically within the first 2 years of symptom onset is a relatively new construct in the United States, and therefore the opportunity to do this kind of research has been limited. Second, many other minoritized statuses should be considered, including those of women, individuals identifying as LGBTQIA+, and persons with disabilities (e.g., deafness) (63). Serious mental illness, with the social stigma and socioeconomic marginalization that often accompanies it, is also a minoritized status (64, 65). These statuses likely intersect to create further health inequities for ethnoracially minoritized youths but were beyond the scope of the present review (66). Finally, identifying ethnoracially minoritized groups required us to use terms that are social constructs and that were often used in inconsistent ways across articles, which limits the generalizability and consistency of findings.

This review supports researchers arguing for a need to clarify and standardize use of the terms and categories of “race” and “ethnicity” in research, which warrants a larger conversation (1). Race is a social construct, and we run the risk of playing into biological essentialism. Ideally, we would move toward a disaggregation of factors that compose “race,” such as socioeconomic status, higher rates of stigma, and so on. Current racial categories required by the NIH, for example, are based on self-report data, and we note issues with how the categories in use have been constructed. For example, one study (34) used race-ethnicity data from the Optum database, which reported race via algorithmic calculations based on items such as zip code.

Research that promotes mental health equity by engaging with more diverse groups and using more inclusive samples, such as pilot studies of interventions that map onto cultural preferences and studies with the statistical power to test racial-ethnic differences, is now within reach. The EPINET (Early Psychosis Intervention Network; https://nationalepinet.org) platform, which includes 101 clinics across 17 states and eight regional hubs, presents new opportunities for recruiting diverse samples to engage in sufficiently powered research that is explicitly focused on persons from ethnoracially minoritized groups (and persons from other minoritized groups) and pilot interventions for specific populations across hubs.

Future research on “culture” might more explicitly study the ways in which history, economics, politics, stigma, stereotypes, cognitive bias, and other such factors are relevant for youths with early psychosis and their families. Studies could be more explicitly designed to examine the impacts of structural racism, intersectionality, or social determinants of mental health. For example, issues related to discrimination against persons with a minority status, general mistrust of the mental health system among individuals from minoritized groups, and structural barriers that prevent such individuals from engaging in research and clinical care (e.g., transportation) may affect study results, but they are often not considered in investigations. Intersectionality of minority status and disadvantage (e.g., sexual or gender minority status, disability status, and socioeconomic status) could also be attended to more often. Rather than being mostly descriptive (as these studies were), future studies might focus on naming the cultural systems and their mechanisms of harm that perpetuate inequity and mental health disparities for youths with early psychosis and their families (67).

It is also difficult to disentangle variables related to socioeconomic status and ethnoracially minoritized status on the basis of the studies available, and more evidence is needed to better understand these differences. Around 70% of the studies in a systematic review of pathways to care for Black youths focused on no- to low-income Black youths (3), suggesting that a socioeconomic gradient may be involved in differential outcomes in these studies, including in family engagement. Findings from a recent analysis of the RAISE study suggest that Black individuals are more likely to have lower socioeconomic status and that higher socioeconomic status positively correlates with better outcomes (68), prompting long-standing questions about the intersections of race and socioeconomic status (69). Even so, results from most outcomes studies indicate that people from ethnoracially minoritized groups generally experience worse outcomes. We need tools and research to better understand the implications of such findings.

Young Black and Latino people diagnosed as having early psychosis and their families also wanted to better understand how to manage the social use of substances in everyday life without exacerbating symptoms or losing social opportunities (7). Substance use interventions can focus on helping youths with early psychosis overcome social and structural barriers (such as the overconcentration of alcohol outlets in Black neighborhoods) (70) to cutting back or quitting.

Given that providers of early psychosis services (including some U.S. providers) have reported many issues related to a lack of cultural understanding and culturally meaningful and sensitive support in their clinics (71), it is also important to have more U.S.-based research on these issues, with diverse populations. More research is needed on the therapeutic alliance and provider characteristics, such as gender and race-ethnicity, as well as how work roles may be subject to minoritization (e.g., White male psychiatrists vs. Black female social workers), staff turnover of providers from minoritized groups, and how to recruit and retain diverse providers in the mental health workforce.