Data from 20,400 participants who completed either a laptop-based or interviewer-administered version of the Collaborative Psychiatric Epidemiology Surveys (CPES) were used. The CPES is a nationally-representative study of adults 18 years and older living in households in the contiguous United States between 2001 and 2003. The CPES encompasses three cross-sectional surveys to ensure participation of people representing a wide range of racial and ethnic experiences: (1) The National Survey on American Life (NSAL), (2) The National Comorbidity Survey Replication (NCS-R), and (3) The National Latino and Asian American Study (NLAAS). Data collection of the CPES were based on the selection of a probability sample of respondents using a multi-stage area probability sampling process of the following population sampling units: (1) US Metropolitan Statistical Areas and counties, (2) area segments, and (3) housing units within the selected area segments. Upon identification of the respondent sample, interviewers contacted housing units in-person conducted a short screening interview with a knowledgeable adult (19). Each participant underwent a validated core interview questionnaire assessing the nature, severity, and impairment of various mental health disorders, including measures of primary mental health diagnostic symptoms, symptom severity, and utilization of mental health services, conducted at different time points (20–23). NSAL, NLAAS, and NCS-R datasets were combined for analysis using a study specific weighting scheme to create combined-analysis weights. Combined analysis weights were created by using the relative size of the study subgroup, defined by race or ethnicity and geographic region.

The original study consisted of 20,400 participants. Participants with incomplete data across any variable used for the statistical analysis were excluded (40.82%). As a result, the final sample size for the analysis was 11,844 participants.

Participants in the subsample exhibited a higher unweighted prevalence of psychosis (4.4%, subsample vs. 0%, non-participants), a higher prevalence of comorbid psychosis (7.1%, subsample vs. 3.2%, non-participants), and higher prevalence of mental illness (34.4%, subsample vs. 9.6%, non-participants) compared to those who did not participate. Additionally, a higher proportion of subsample participants were female (53.3%, subsample vs. 51.6%, non-participants), and attained a college level of education or higher (54.8%, subsample with college education or higher vs. 47.5%, non-participants with college education or higher). A lower proportion were between the ages of 18–45 (51.8%, subsample vs. 57.7%, non-participants). These differences were statistically significant (p < 0.0001). Consequently, the demographic characteristics of the subsample differs compared to that of the full CPES sample. Further, the subsample has a higher prevalence of psychosis, comorbid psychosis and mental illness compared to the full CPES.

Most participants (52.7%) identified as White. Asian participants accounted for 8.2% of the sample, 19.2% of participants identified as Black, and 20.0% of participants identified as Hispanic ( Table 1 ). Most participants (78%) completed 12 or more years of education. Similarly, most participants (58.8%) were defined as non-poor. The majority of participants (40.7%) were between the ages of 26–45 ( Table 1 ).

About 34.4% of all participants reported experiencing mental illness only, 4.4% reported psychosis only, and 7.1% experienced comorbid psychosis ( Table 1 ).

Approximately 1.5% of White participants were affected by psychosis only, while 0.5% of Asian and 1.2% of Black as well as Hispanic participants were affected. Approximately 21.3% of White participants, 1.7% of Asian participants, and 5.7% of Black as well as Hispanic participants were affected by mental illness only. Approximately 3.5% of White participants, 0.3% of Asian participants, 1.8% of Black participants, and 1.5% of Hispanic participants were affected by comorbid psychosis.

Compared to White participants, Hispanic participants were 87% more likely to report psychosis only (OR = 1.87, CI = 1.18 – 2.95, p = 0.008) ( Table 2 ). Black participants were 95% more likely to report psychosis only (OR = 1.95, CI = 1.30 – 2.96, p = 0.002) compared to White participants. No significant association was detected for Asian participants.

Non-White participants were on average less likely to report having mental illness only compared to White participants ( Table 2 ). For example, Asian participants were 64% less likely to report experiencing a mental illness (OR = 0.36, 95% CI = 0.30 – 0.44, p < 0.0001), compared to White participants. Hispanic and Black participants were 39% and 33% less likely to report experiencing a mental illness (OR_Hispanic = 0.61, 0.95% CI = 0.52 – 0.73, p < 0.0001; OR_Black = 0.67, 95% CI = 0.57 – 0.79, p < 0.0001, Hispanic and Black, respectively), compared to White participants. Similarly, compared to White participants, Asian participants were 63% less likely to report experiencing comorbid psychosis (OR = 0.37, 95% CI = 0.27–0.53, p < 0.0001).

Compared to White participants, Hispanic participants were 94% more likely to report only experiencing psychosis (AOR_Hispanic = 1.94, 0.95% CI = 1.2 – 2.98, p = 0.002), and Black participants were 86% more likely to report only experiencing psychosis after accounting for covariates (AOR_Black = 1.86, 95% CI = 1.24 – 2.82, p = 0.003) ( Table 3 ).

Compared to White participants, Asian participants were 70% less likely to report experiencing mental illness only (AOR = 0.30, 95% CI = 0.25–0.36, p < 0.001). Hispanics and Black participants were 47% and 43% less likely to report experiencing mental illness only, (AOR_Hispanic = 0.53, 95% CI = 0.44–0.65, p < 0.001; AOR_Black = 0.57, 95% CI = 0.48–0.68, p < 0.001).

Compared to White participants, Asian participants were 68% less likely to experience comorbid psychosis (AOR_Asian = 0.32, 95% CI = 0.22–0.47, p < 0.001). Similarly, Hispanic participants were 34% less likely to experience comorbid psychosis (AOR = 0.66, 95% CI =0.48–0.92, p = 0.001) compared to White participants ( Table 3 ).

This study expands on the prior literature by identifying trends of comorbid psychosis by self-identified race or ethnicity. The findings align with our hypothesis that there are discernible differences in prevalence of comorbid psychosis by self-identified race or ethnicity in the general population. Approximately 7.1% of all study participants experienced comorbid psychosis, which is lower than previously reported studies (10, 13, 32). However, these studies focused on clinical or patient-focused studies and often the majority of participants self-identify as White. Therefore, this estimate is likely to be lower because it reflects a prevalence rate for the general population context, rather than patient-based samples.

Overall, non-White racial or ethnic participants had a lower prevalence of comorbid psychosis compared to White participants. Black participants (1.8%) had a lower prevalence of comorbid psychosis compared to White participants (3.5%). A similar trend was identified in Hispanic (1.5%) and Asian participants (0.3%). Our results for comorbid psychosis are consistent with many prior studies on comorbid psychosis. However, several studies have yielded varied conclusions concerning comorbid psychosis based on age of onset and accessing and engaging with mental health infrastructure (32–34). Further, multinomial regression results adjusted for the effects of covariates indicated that the odds of being diagnosed with comorbid psychosis was lower for Asian and Hispanic participants compared to White participants. However, this pattern did not extend to Black participants.

It is possible that the detected racial and ethnic differences may reflect measurement non-invariance between racial or ethnic groups. For instance, studies have suggested that symptoms such as paranoia and delusions may be more likely to be labeled as cultural expressions (i.e., creative expression of cultural identity) rather than symptoms of psychosis in certain racial or ethnic groups (35–39). Further, individuals who identify as non-White may be more likely to receive a misdiagnosis or delayed diagnosis of a psychotic disorder due to these measurement differences (40–42). These measurement differences could practically lead to a delay in access to appropriate treatment. Consequently, future studies are encouraged to assess the consistency of measurement non-invariance by race and ethnicity as well as the implications in the epidemiology of these conditions as well as their treatment. These results highlight the need for further research to understand the specific contributions to these disparities.

We were able to reproduce prior results related to a psychosis-only diagnosis in our sample. The odds of being affected with psychosis only in Black and Hispanic participants was almost two times higher than White participants. This result is consistent with previous research, which has also reported higher rates of psychosis among Black and Hispanic participants (4). Moreover, these results align with prior literature that indicates individuals of recent African or Afro-Caribbean descent are more susceptible to experiencing psychosis (8).

Interestingly, despite the higher rates of psychosis diagnosis, Black and Hispanic participants in this study had lower overall rates of mental illness. This paradoxical finding suggests that factors beyond the mere presence of mental illness contributes to the disparities in psychosis rates among different racial and ethnic groups. In particular, the stigma surrounding mental health in certain communities could result in underreporting or underdiagnosing of other mental illnesses among non-White racial or ethnic groups. Consequently, members of some non-White racial or ethnic groups tend to experience higher rates of specific mental health conditions, such as psychosis or substance use, while simultaneously reporting lower overall rates of mental illness when compared to White individuals (43). Avoidance or delay in seeking help could lead to a higher prevalence of severe mental health conditions (e.g., psychosis) when individuals from these communities do eventually access mental health care. The mental health paradox highlights the complexity of mental health disparities and challenges prevailing assumptions about mental health prevalence solely based on racial or ethnic categorizations (43). To address this paradox, it is imperative to conduct more nuanced research that addresses the multifaceted factors influencing mental health outcomes in different communities including income, gender, and age (i.e., birth cohorts). Consequently, these results replicate prior results and highlight the importance of understanding the mental health paradox as it applies to psychosis and comorbid psychosis.

These results should be interpreted while considering the following limitations. First, these data were collected in 2001–2003 and do not capture more recent definitions as detailed in DSM-5, such as changes in duration of symptoms or increased specification across many psychotic disorders. Consequently, these findings may not be generalizable to the current generation of DSM primary psychotic disorders. However, it is a broad definition that may be more applicable to Non-White racial or ethnic communities. Second, the analytic sample size was reduced after removing participants with missing data. However, there were significant differences in the distribution of variables between the full sample and study subsample. The analysis of an imputed sample could result in biased parameter estimates (44), and as such imputation was not conducted. Third, the use of self-reported lifetime prevalence was estimated which may be subject to rater and/or recency bias. However, amongst those who reported experiencing psychosis within their lifetime, 63% had a psychotic experience within the last 12 months. Fourth, participants self-identified their race or ethnicity. Such categorizations are based on socially-derived constructs. Consequently, these results cannot disaggregate the degree to which these results reflect a purely social etiology from social processes that may be subject to genetic confounding (45). Fifth, a significant concern revolves around the potential influence of diverse diagnostic approaches employed across countries and individuals, casting doubt on the generalizability of the results. Psychosis, known for its stability in treatment and diagnosis, is notably consistent, irrespective of cultural constructs (46). The absence of a standardized diagnostic procedure, particularly in distinguishing comorbid psychiatric disorders, raises questions about the transferability of the conclusions to broader contexts. These findings underscore the necessity for additional contextualization of these phenomena by creating more clear diagnosis instructions. Sixth, a critical issue within the study pertains to the diagnostic reliance on the CIDI psychosis screener. While this tool serves as a useful preliminary assessment, it provides only a probable diagnosis rather than a formal one. This limitation introduces an element of uncertainty into the accuracy of the findings. Caution should be exercised when extrapolating implications for clinical practice, although it remains useful when discussing the general population. Furthermore, due to small sample sizes in some individual race and ethnic categories, the study combined certain groups into broader race or ethnicity categories. While this contributed to statistical power of the analysis, it also limited the ability to differentiate how specific groups within the same broader category (e.g., aggregating African Americans and Black Caribbeans in a single Black category) might uniquely experience mental illness.

Despite these limitations, this analysis possesses several significant strengths that contribute to its relevance and value. The study utilizes a large sample that reflects the diversity and complexities of the broader United States population. By incorporating participants from various backgrounds, including immigrant populations and non-White racial or ethnic groups, the research captures a comprehensive picture of mental health experiences across different communities.