The Psychiatrist's Blind Spot: How Diagnostic Bias Against Black Patients Persists Despite Decades of Evidence
A young Black woman waits in a psychiatric emergency room. She is grieving, traumatized, and depressed. Yet the clinician describes her as "guarded," "hostile," and "paranoid" in the chart. She leaves with a schizophrenia diagnosis, a prescription for a first-generation antipsychotic, and a future shaped by a label she should never have received.
This is not a hypothetical. It is a real case documented in the Harvard Review of Psychiatry. The woman was eventually rediagnosed with major depressive disorder and PTSD, but only after years of harm caused by a misdiagnosis rooted in racial bias.[1] Her story is not unique. It is the result of a diagnostic system that has never fully addressed its own biases.
The Numbers We Already Know
This evidence is not new. A 2024 review in JAMA Psychiatry looked at 64 studies and over 54,000 people with schizophrenia. It found that Black individuals were more than twice as likely as White individuals to be diagnosed with schizophrenia (odds ratio 2.07, 95% CI, 1.64–2.61).[2] Of 16 studies in general mental health settings, 12 found higher rates of schizophrenia diagnoses for Black individuals, and all seven studies using research-based diagnostic tools confirmed this gap.[2] An earlier meta-analysis by Garb also found that race bias affects the diagnosis of schizophrenia and related disorders, and that how symptoms are expressed, including Black cultural expressions of depression, strongly influences diagnoses.[3] These patterns have stayed the same for thirty years.
A 2025 national household study using the Structured Clinical Interview for DSM-5 found that non-Hispanic Black individuals had a 4.1% prevalence of schizophrenia spectrum disorders, compared to 1.2% among non-Hispanic White individuals (adjusted OR = 3.49; 95% CI, 1.37–8.91).[4] However, when researchers adjusted for neighborhood social vulnerability, the difference was no longer significant. This suggests that structural factors, not biology, explain the gap.[4] Anglin and colleagues have shown that a history of structural racism in the United States has shaped social conditions, with neighborhood factors, ongoing trauma and stress, and complications before and after birth all reflecting social and environmental issues that may increase psychosis risk through a pathway shaped by structural racism.[5]
Providers are more likely to diagnose schizophrenia based on race than any other demographic factor, and this gap remains even when symptom ratings are the same between groups.[1] When terms like hostility, suspiciousness, and dangerousness are influenced by race, they contribute to these diagnostic gaps and affect treatment decisions.[6]
A History Written in Pseudoscience
This is not a new problem. It is part of a long history. From the beginning, psychiatry has been influenced by white supremacist ideas. Pseudoscientific theories like "drapetomania"—the supposed illness that made enslaved people want to escape—and "dysaesthesia aethiopis" have left a legacy of anti-Black racism that still affects mental health today.[6][7] Psychiatric diagnoses have been used to hold back Black people by suggesting they are naturally immoral or criminal, and as a way to justify control and oppression. For example, the idea of "protest psychosis" labeled civil rights activism as a symptom of schizophrenia.[6] The DSM-5-TR itself says that race is a social, not biological, concept, and that discrimination based on race deeply affects health.[8] Still, the diagnostic system has never been fully cleansed of this history.
The Downstream Consequences
The effects of misdiagnosis go far beyond a label. Black patients who are misdiagnosed are less likely to get antidepressants or electroconvulsive therapy for their real mood disorders.[6] They are more likely to be given first-generation antipsychotics, higher doses of medication, and long-acting injections.[6] They also face more forced treatment, more use of physical restraints, and higher rates of civil commitment.[8] A meta-analysis in JAMA Internal Medicine found that Black patients are at higher risk of being physically restrained in emergency departments. This may be partly because overdiagnosis of psychotic disorders in Black patients limits their access to outpatient care and leads to more severe illness when they seek help.[9]
The social impact is just as severe. A national study found that Black people diagnosed with schizophrenia faced more disadvantages than White people with the same diagnosis. They had lower rates of employment, education, income, marriage, and social support.[10] They also reported much higher levels of discrimination throughout their lives, including at work (Cohen's d = 0.85), in the health system (d = 0.70), and from public race-based abuse (d = 0.55).[10] Even with these challenges, they were less likely to get psychiatric treatment.[10]
As Faber and colleagues wrote, misdiagnosis leads to "a progression of lifelong punitive societal implications, including reduced opportunities, substandard care, increased contact with the legal system, and criminalization." Misra and colleagues have shown that racialized policing, incarceration, economic exploitation, and the violence of racism all increase both the risk of psychotic disorders and the inequalities in how they are treated.[11][12] The diagnostic label itself becomes a form of structural harm, following patients through housing applications, job screenings, custody cases, and interactions with law enforcement.
Our own professional organizations recognize this problem. The DSM-5-TR clearly states that misdiagnosing schizophrenia in African Americans with mood disorders is a result of clinician bias and structural racism.[8] The APA's 2021 Resource Document on Anti-Black Racism calls the overdiagnosis of psychotic disorders in Black populations "one of the most impactful and well-described inequities in psychiatry.[6] We know this. We have known it for decades.
So why hasn't it changed?
The Implicit Bias Training Trap
Most institutions have responded with implicit bias training. While these programs are well-intentioned and are now often required, they are mostly ineffective.
A 2024 systematic review of 77 implicit bias training programs in healthcare, published in Science Advances, found that the primary training target was race/ethnicity (49.3%), but trainings commonly lack specificity on addressing implicit prejudice or stereotyping (67.5%), diverge from current scientific literature in up to 67.5% of cases, and lack internal validity in 99.9% of cases.[13] A separate review in the Annual Review of Public Health concluded that while bias training can raise awareness and engage providers in establishing egalitarian goals, "these changes are not sustained, and the interventions have not demonstrated change in behavior in the clinical or learning environment."[14]
Even worse, some interventions can backfire. Cooper and colleagues argued in JAMA Health Forum that poorly designed training can create a "false sense of confidence in training that has had no benefit," and that telling clinicians to avoid stereotyping can actually make stereotypes more active.[15] When institutions simply complete "bias training" as a requirement, they risk normalizing the very biases they aim to fix, replacing real change with mere procedural compliance.
Awareness matters, but without real changes to the system, it is just for show. Training alone cannot solve a structural problem.
From Awareness to Architecture
If the diagnostic system produces biased outputs, it must be redesigned. The evidence points to several concrete structural interventions.
First, measurement-based care. The APA recommends using quantitative symptom measures in all psychiatric evaluations.[16] Regular use of validated rating scales, similar to checking blood pressure for hypertension, can help reduce bias in clinical impressions. At least nine review articles show that measurement-based care leads to better outcomes and larger improvements than usual care.[17] New evidence suggests it can also improve the relationship between clinicians and patients of color, helping them take a more active role in their care and reducing early dropout.[18] Problems with the therapeutic relationship and treatment relevance often cause early dropout for communities of color, and measurement-based care offers a more personalized and transparent approach.[18] Still, less than 20% of practitioners use measurement-based care, and only about 5% use it as often as recommended.[17] Expanding this practice is overdue.
Second, structured diagnostic interviews. While these assessments have not completely removed racial disparities in schizophrenia diagnosis, they may help reduce them somewhat.[3] More importantly, they require clinicians to systematically check for mood, trauma, and substance use disorders instead of relying on first impressions, which too often label Black patients as "psychotic." Garb's review found that using more structured interviews, self-report tools, and statistical prediction rules can help decrease diagnostic bias.[1][3]
Third, the Cultural Formulation Interview. The DSM-5-TR includes this interview because cultural context affects how symptoms appear.[8] Cultural ways of expressing distress—such as certain behaviors, words, or ways of describing suffering—can be mistaken for paranoia or thought disorder by clinicians who are not familiar with them.[8] The CFI can be used by any clinician in any setting and is especially useful when there are significant differences in culture, religion, or socioeconomic background between the clinician and the patient, or when it is unclear whether symptoms meet the usual criteria.[8] It is not just a formality; when used regularly, it can help prevent the misunderstandings that lead to misdiagnosis.
Fourth, trauma-informed assessment. The 2025 APA Clinical Practice Guideline for PTSD clearly states that clinicians have often missed PTSD in Black patients and instead diagnosed depression or a psychotic disorder.[19] Severe PTSD can look like hypervigilance, mistrust, dissociation, intrusive thoughts, and emotional withdrawal—symptoms that may seem like psychosis if the full context is not considered.[19] For many marginalized communities, the line between past trauma and a "safe" present is not clear, as many still face high levels of crime, racism, and both large and small acts of aggression.[19] Black Americans experience more trauma from violence, social disadvantage, and discrimination, so careful trauma assessment is crucial before labeling symptoms as a primary psychotic disorder.[5][12]
Fifth, use dimensional diagnostic approaches. Categorical diagnoses—like deciding if someone has schizophrenia or not—are especially prone to bias at the edges. Dimensional models, which look at the full range of psychotic, mood, and trauma symptoms, can help reduce mistakes that often harm Black patients.[3] When the choice is "schizophrenia or major depression," Black patients are too often diagnosed with schizophrenia. But if we ask, "Where does this patient fall on a spectrum of psychotic, mood, and trauma symptoms?" we are more likely to get it right.
The Coming Challenge: Artificial Intelligence
These structural reforms are urgent not only because of the current system's failures, but because of what is coming next. Artificial intelligence is increasingly integrated into mental health care, and early evidence suggests it may replicate — or even amplify — existThese structural reforms are urgent not just because the current system is failing, but also because of what is coming next. Artificial intelligence is becoming more common in mental health care, and early evidence shows it may repeat or even worsen existing biases. A 2025 study in NPJ Digital Medicine looked at racial bias in four major language models used for psychiatric cases and found that these models often suggested worse treatments when a patient's race was clear.[20] Straw and Callison-Burch also found strong biases in language processing models used in psychiatry, including biases related to race, religion, gender, nationality, sexuality, and age. A [21]A review in the Annals of Internal Medicine found that algorithms can mitigate, perpetuate, or exacerbate racial and ethnic disparities, regardless of whether they explicitly use race and ethnicity, and that intentionality and implementation determine their effect on disparities.[23] If we build AI diagnostic support tools on a foundation of biased clinical data — data generated by the very system that overdiagnoses schizophrenia in Black patients — we will automate the bias at scale. The structural reforms described above are therefore not only corrections for the present but prerequisites for an equitable future.
Expanding the Workforce and the Care Model
Diversifying the workforce is still essential. Doctors from underrepresented minority groups are more likely to work in underserved areas and care for Medicaid patients.[6][24] A 2023 study in JAMA Network Open found that having more Black primary care doctors was linked to longer life expectancy and lower death rates in the population.[24] Having a psychiatric workforce that matches the communities it serves is not just about diversity—it is about patient safety.
Beyond physician diversity, community mental health workers and peer support specialists are valuable but underused. Peer support has been linked to more use of outpatient mental health services and smaller gaps in care for Black and Latinx youth.[25] Peer providers can achieve results as good as or better than those of non-peer professionals, and they are especially effective at reducing hospital admissions and reaching patients with severe illness.[26] The APA's Resource Document on Anti-Black Racism recommends supporting culturally responsive, community-based care and using the Collaborative Care Model to improve access and reduce diagnostic and treatment gaps.[6] How Psychiatrists Can Talk with Patients and Their Families About Race and Racism notes that frank discussions about race can enhance therapeutic alliance, decrease mistrust, and create an atmosphere of openness — while also addressing existing racial power imbalances in the therapeutic relationship.[27] A study in The Lancet Psychiatry found that racial and ethnic minority patients with depression symptoms had an increased preference for providers who share or understand their culture, but had low rates of accessing providers who could provide such care.[28] Cultural competence is not an abstract ideal; it is a measurable clinical skill that patients value and that affects outcomes.
A Professional Obligation
Psychiatry is the only medical field where the main way we diagnose—a clinical interview—has no objective test, lab value, or scan to confirm it. This makes our diagnoses especially open to the biases we all have. That vulnerability is not a moral failing. But refusing to address it, even after decades of evidence, is.
We do not need more workshops. We need structured interviews built into everyday clinical practice. Measurement-based care should be the standard, not just a goal. The Cultural Formulation Interview should be used regularly, not just sometimes. Trauma-informed assessments must recognize the ongoing stress many Black Americans face. We must prepare now for AI tools that could either reinforce or change these patterns. And we need a workforce and care model that truly reflects the communities we serve.
The young woman in the emergency room deserved better. Every patient who follows her deserves better, too.
References
1. Psychotic Misdiagnosis of Racially Minoritized Patients: A Case-Based Ethics, Equity, and Educational Exploration. Jegarl AM, Jegede O, Isom J, Ciarleglio N, Black C. Harvard Review of Psychiatry. 2023 Jan-Feb 01;31(1):28-36. doi:10.1097/HRP.0000000000000353.
2. Ethnoracial Risk Variation Across the Psychosis Continuum in the US: A Systematic Review and Meta-Analysis. van der Ven E, Olino TM, Diehl K, et al. JAMA Psychiatry. 2024;81(5):447-455. doi:10.1001/jamapsychiatry.2023.5497.
3. Race Bias and Gender Bias in the Diagnosis of Psychological Disorders. Garb HN. Clinical Psychology Review. 2021;90:102087. doi:10.1016/j.cpr.2021.102087.
4. Neighborhood Social Vulnerability and Racial Disparities in Schizophrenia Spectrum Disorder Prevalence. Anglin DM, Olfson M, van der Ven E, et al. The American Journal of Psychiatry. 2025;182(9):850-860. doi:10.1176/appi.ajp.20240906.
5. From Womb to Neighborhood: A Racial Analysis of Social Determinants of Psychosis in the United States. Anglin DM, Ereshefsky S, Klaunig MJ, et al. The American Journal of Psychiatry. 2021;178(7):599-610. doi:10.1176/appi.ajp.2020.20071091.
6. Resource Document on Advocating for Anti-Racist Mental Health Policies with a Focus on Dismantling Anti-Black Racism. Joshua R. Anthony MD MBA, Amanda Calhoun MD MPH, Enrico G. Castillo MD MSHPM, et al. American Psychiatric Association (2021).
7. Modern-Day Relics of Psychiatry. Tripathi S, Messias E, Spollen J, Salomon RM. The Journal of Nervous and Mental Disease. 2019;207(9):701-704. doi:10.1097/NMD.0000000000001059.
8. Diagnostic and Statistical Manual of Mental Disorders. Dilip V. Jeste, Jeffrey A. Lieberman, David Fassler, et al. American Psychiatric Association (2022).
9. Racial Disparities in Emergency Department Physical Restraint Use: A Systematic Review and Meta-Analysis. Eswaran V, Molina MF, Hwong AR, et al. JAMA Internal Medicine. 2023;183(11):1229-1237. doi:10.1001/jamainternmed.2023.4832.
10. Comparison of Black and White Individuals Who Report Diagnoses of Schizophrenia in a National Sample of US Adults: Discrimination and Service Use. Bommersbach TJ, Rhee TG, Stefanovics EA, Rosenheck RA. Schizophrenia Research. 2023;253:22-29. doi:10.1016/j.schres.2021.05.017.
11. The Weaponization of Medicine: Early Psychosis in the Black Community and the Need for Racially Informed Mental Healthcare. Faber SC, Khanna Roy A, Michaels TI, Williams MT. Frontiers in Psychiatry. 2023;14:1098292. doi:10.3389/fpsyt.2023.1098292.
12. Structural Racism and Inequities in Incidence, Course of Illness, and Treatment of Psychotic Disorders Among Black Americans. Misra S, Etkins OS, Yang LH, Williams DR. American Journal of Public Health. 2022;112(4):624-632. doi:10.2105/AJPH.2021.306631.
13. The Nature and Validity of Implicit Bias Training for Health Care Providers and Trainees: A Systematic Review. Hagiwara N, Duffy C, Cyrus J, et al. Science Advances. 2024;10(33):eado5957. doi:10.1126/sciadv.ado5957.
14. Eliminating Explicit and Implicit Biases in Health Care: Evidence and Research Needs. Vela MB, Erondu AI, Smith NA, et al. Annual Review of Public Health. 2022;43:477-501. doi:10.1146/annurev-publhealth-052620-103528.
15. Mandated Implicit Bias Training for Health Professionals—A Step Toward Equity in Health Care. Cooper LA, Saha S, van Ryn M. JAMA Health Forum. 2022;3(8):e223250. doi:10.1001/jamahealthforum.2022.3250.
16. Resource Document on Implementation of Measurement-Based Care. Kathryn Ridout, Erik R. Vanderlip, Carol L. Alter, et al. American Psychiatric Association (2023).
17. Implementing Measurement-Based Care in Behavioral Health: A Review. Lewis CC, Boyd M, Puspitasari A, et al. JAMA Psychiatry. 2019;76(3):324-335. doi:10.1001/jamapsychiatry.2018.3329.
18. Leveraging Measurement-Based Care to Reduce Mental Health Treatment Disparities for Populations of Color. Barber J, Childs AW, Resnick S, Connors EH. Administration and Policy in Mental Health. 2025;52(1):172-178. doi:10.1007/s10488-024-01364-4.
19. Clinical Practice Guideline for the Treatment of Posttraumatic Stress Disorder (PTSD) in Adults (2025). Lori A. Zoellner, Priscilla M. Schulz, Lucindra Campbell-Law, et al. American Psychological Association.
20. Racial Bias in AI-mediated Psychiatric Diagnosis and Treatment: A Qualitative Comparison of Four Large Language Models. Bouguettaya A, Stuart EM, Aboujaoude E. NPJ Digital Medicine. 2025;8(1):332. doi:10.1038/s41746-025-01746-4.
21. Artificial Intelligence in Mental Health and the Biases of Language Based Models. Straw I, Callison-Burch C. PloS One. 2020;15(12):e0240376. doi:10.1371/journal.pone.0240376.
22. Considering Biased Data as Informative Artifacts in AI-Assisted Health Care. Ferryman K, Mackintosh M, Ghassemi M. The New England Journal of Medicine. 2023;389(9):833-838. doi:10.1056/NEJMra2214964.
23. The Impact of Health Care Algorithms on Racial and Ethnic Disparities : A Systematic Review. Siddique SM, Tipton K, Leas B, et al. Annals of Internal Medicine. 2024;177(4):484-496. doi:10.7326/M23-2960.
24. Black Representation in the Primary Care Physician Workforce and Its Association With Population Life Expectancy and Mortality Rates in the US. Snyder JE, Upton RD, Hassett TC, et al. JAMA Network Open. 2023;6(4):e236687. doi:10.1001/jamanetworkopen.2023.6687.
25. The Availability of Peer Support and Disparities in Outpatient Mental Health Service Use Among Minority Youth With Serious Mental Illness. Ojeda VD, Munson MR, Jones N, Berliant E, Gilmer TP. Administration and Policy in Mental Health. 2021;48(2):290-298. doi:10.1007/s10488-020-01073-8.
26. Three Nontraditional Approaches to Improving the Capacity, Accessibility, and Quality of Mental Health Services: An Overview. Grant KL, Simmons MB, Davey CG. Psychiatric Services (Washington, D.C.). 2018;69(5):508-516. doi:10.1176/appi.ps.201700292.
27. Resource Document on How Psychiatrists Can Talk with Patients and Their Families About Race and Racism. American Psychiatric Association (2020).
28. Racial and Ethnic Differences in Perception of Provider Cultural Competence Among Patients With Depression and Anxiety Symptoms: A Retrospective, Population-Based, Cross-Sectional Analysis. Eken HN, Dee EC, Powers AR, Jordan A. The Lancet. Psychiatry. 2021;8(11):957-968. doi:10.1016/S2215-0366(21)00285-6.