Abstract
Objective:
To describe associations between patient race and ethnicity with emergency department disposition for mental health visits in the United States.
Methods:
We identified 674,821 visits for mental health in the 2019 National Emergency Department Sample and classified them by ICD-10 diagnostic group: schizophrenia-spectrum, bipolar, major depressive, anxiety, or other disorders. Racial and ethnic categories were White, Black, Hispanic, or other. Logistic regression models, adjusted for age, sex, insurance status, and medical comorbidities, were used to describe differences in odds of inpatient admission by race/ethnicity and diagnosis.
Results:
After covariate adjustment, we did not find overall differences in the likelihood of admission between racial/ethnic groups. However, compared to White patients, admission rates were lower for visits by Black patients for bipolar disorder, OR=0.71 (95%CI=0.57–0.81), and major depressive disorder, OR=0.70 (95%CI=0.59–0.83), and lower for Black (OR=0.61; 95%CI=0.50–0.74) and Hispanic patients (OR=0.67; 95%CI=0.49–0.91) for anxiety disorders. There were no significant racial/ethnic differences in admission rates for schizophrenia-spectrum disorders.
Conclusions:
Overall admission rates were comparable for Black and White patients. After covariate adjustment, there were no differences across racial/ethnic groups, though some racial/ethnic differences persisted within diagnostic subsets of mood and anxiety disorders.
Keywords: health policy, mental health services, psychiatric emergency services, health disparities
Introduction
Racial and ethnic disparities have been described throughout multiple facets of mental healthcare in the United States, though the nature of disparities differs by group and care setting.1,2 Black patients are overrepresented in acute psychiatric settings and on inpatient units compared to White patients.1,3–7 At the same time, rates of outpatient mental health utilization tend to be lower among Black and Hispanic populations.3,8,9 Several hypotheses have been developed to explain these patterns in mental healthcare. Some have suggested that while Black and Hispanic populations report lower overall rates of mental illness than White populations, conditional on having a psychiatric diagnosis, Black and Hispanic patients are more likely to have severe illness.2 Others have suggested that these differences are due largely to differential access to care.1 Still others have suggested that, compared to White patients, non-White patients have a greater reluctance to seek mental health care.10–12
Perhaps most concerningly, racial/ethnic differences in mental health use may be generated by provider biases during evaluation.13–18 Differences in rates of diagnosis by race/ethnicity in the United States have been described since at least the 1970s, particularly, higher rates of psychotic disorder diagnosis among Black compared to White Americans.13,19,20 Because psychotic disorders may be associated with more coercive or restrictive forms of psychiatric care, this has led to concern that unjustified overdiagnosis of psychotic disorders among Black patients has contributed to elevated rates of inpatient treatment or involuntary psychiatric treatment among Black Americans.5–7,21 At the same time, there is concern that culturally insensitive diagnostic practices may contribute to underdiagnosis of psychiatric illness among Hispanic or other minority patients.22,23
Emergency department (ED) records may provide insights into usual patterns of care. EDs are an important point of contact and the main route to inpatient admission for patients with severe psychiatric symptoms.24–26 For racial/ethnic groups with relatively low use of outpatient services, evaluation and triage in the ED may be a particularly relevant pathway to mental healthcare access.14 However, because diagnostic precision improves with structured approaches,15 culturally-tailored assessment,27 and length of observation,28 and ED assessments may be completed under time pressure, ED disposition decisions may also carry a higher risk of racial/ethnic diagnostic bias.
To explore these questions, we examined a large nationally representative sample of ED mental health discharges by diagnosis and patient race/ethnicity to explore whether there were racial/ethnic differences in rates of inpatient admission. If Black and Hispanic patients have less access to outpatient care, they may use the ED as a substitute for outpatient care, meaning that the average Black or Hispanic psychiatric ED patient may have lower average severity than the average White patient. Conversely, personal or historical experiences of racism may be associated with a greater reluctance to seek treatment among Black or Hispanic patients.10–12 If the threshold to seek care is higher for Black or Hispanic patients, those who present to the ED may have more severe illness on average. We examined distributions of severity by race/ethnicity as measured by both diagnostically labelled severity and by rates of inpatient admission as a proxy for high risk. Because Black patients, and particularly Black male patients, have been historically overrepresented in inpatient psychiatry,18,21 we examined the data for evidence of biases in admission rates among visits by Black patients in general, as well as more specifically Black male patients and Black male patients diagnosed with schizophrenia-spectrum disorders.18
Methods
Data Sources
The 2019 National Emergency Department Sample (NEDS) is a stratified sample of discharges from U.S. community hospitals, encompassing approximately 30 million ED visits, including patient race/ethnicity. Strata and sample weights are provided by NEDS to produce nationally representative estimates. Adult population estimates for 2019 produced by the U.S. Census Bureau and obtained from the Annie E. Casey Foundation.29
Sample Assembly
We restricted the sample to ED visits for those aged 18–90 excluding older age groups for which psychiatric inpatient admission was less likely. We selected study visits for a mental disorder in the first diagnostic field, excluding mental disorders due to physiologic conditions or substances (Supplementary Table 1).
Dependent and Independent Variables
The primary outcome of interest was ED disposition coded as hospital admission or discharge. Because not all hospitals have psychiatric units, psychiatric admission may require hospital transfer;30 we considered transfers as admissions.
Race and ethnicity were the primary independent variables of interest and contained the non-overlapping classifications of non-Hispanic White (White), non-Hispanic Black (Black), Hispanic, or other (non-White, non-Black, non-Hispanic). Because NEDS racial/ethnic categories are collected by partner organizations collection practices may vary substantially by organization; details regarding specific collection practices are limited.31 Race/ethnicity data were missing for 11,997 visits (1.8% of the sample).
Any insurance was defined as including any of the payers (Medicare, Medicaid, or private insurance) as well as any other insurance, such as Worker’s Compensation, CHAMPUS, CHAMPVA, or Title V, as either a primary or secondary payer. The any private insurance indicator was defined by having either a primary or secondary payer of private insurance.32 Other independent variables included patient sex, 10-year age group, and medical comorbidities as defined by the Elixhauser Comorbidity Index (ECI).33–35
Analyses were subsetted by diagnosis in accordance with the four largest diagnostic categories of illness by visit number plus a residual category: schizophrenia-spectrum disorders, bipolar disorder, major depressive disorder (MDD), anxiety disorder, or other mental health disorders (Supplementary Table 1). The composition of visits for other mental health disorders was explored in descriptive analyses.
Analyses
All analyses were done in STATA 17.0. Survey weights handled using the svy function in STATA.
Descriptive studies
Descriptive variables including rates of inpatient admission, demographic, and comorbidity data were summarized using NEDS sample weights and stratified by diagnostic category. The significance of racial/ethnic differences in admission rates within diagnostic subsets were compared using an F-test.
Logistic regression
Because race/ethnicity is correlated with many other variables associated with admission risk such as socioeconomic status, insurance status, and medical comorbidity, logistic regression was used to model odds of admission as a function of race/ethnicity, having any insurance, private insurance status, patient sex, age group, and ECI comorbidities. Models included overall and analyses by diagnosis. Both unadjusted and covariate-adjusted models are shown. The reference categories for these regressions were male, White race/ethnicity, with no insurance, and no medical comorbidity. White race/ethnicity was chosen as a reference demographic category due to it being the largest population size; the authors do not intend to imply this group as normative. We also explored racial/ethnic differences in mania versus depression visits among bipolar disorder visits.
Subset analyses
Affective and Psychotic Symptom Variation
Prior literature suggests that patient expressions and provider interpretations of affective and psychotic symptoms vary by race/ethnicity.13,15,17,36,37 Further, for a variety of reasons including diagnostic,15 historical and structural factors,36 and service availability,1 the distributions of symptom severity among patients may differ by race/ethnicity.2 To explore these hypotheses, we leveraged diagnostic specifiers. Specifically, we explored affective symptom severity through depression specifiers of severe psychosis, and schizoaffective disorder diagnosis among schizophrenia-spectrum disorders visits.
Admission Rates for Men with Schizophrenia-Spectrum Illness
Informed by hypotheses of social control,18,21 we also explored racial/ethnic differences among visits by male patients with schizophrenia-spectrum illness and specifically paranoid schizophrenia, a discontinued38 but historically racialized diagnosis.21
Human Subjects Research
The New York State Psychiatric Institute Institutional Review Board deemed this analysis of de-identified data to be exempt from human subjects research review.
Results
Demographic characteristics
The study sample contained 674,821 mental health visits including those for schizophrenia-spectrum disorders (22.5%), bipolar disorder (9.7%), MDD (23.2%), anxiety disorders (40.7%), and other mental health disorders (3.9%) (Table 1). Of other diagnoses, the largest subcategories were unspecified mood disorders (33.7%), personality disorders (17.3%), conduct disorders (17.0%), and mental disorders not otherwise specified (13.9%). The sample was 59.6% White, 20.1% Black, 12.4% Hispanic, and 6.3% other. By race/ethnicity, this implies a rate of ED visits per 1,000 adult population of 18.5 (95%CI=16.9–20.2) among Blacks, 10.6 (95%CI=10.2–11.0) among Whites, 8.5 (95%CI=7.6–9.4) among Hispanics, and 8.2 (95%CI=7.1–9.3) among other race/ethnicities.29
Table 1.
Demographic and clinical characteristics for ED mental health visits by primary diagnosis, United States, 2019.
| Overall (N=674,821) | SCZ (N=154,814) | BPD (N=65,060) | MDD (N=154,699) | ANX (N=273,467) | Other dx (N=26,781) | |||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Variable | Mean | 95% CI | Mean | 95% CI | Mean | 95% CI | Mean | 95% CI | Mean | 95% CI | Mean | 95% CI |
|
| ||||||||||||
| Admit | 26.4% | (24.0-28.8) | 41.5% | (37.1-45.9) | 50.7% | (47.2-54.2) | 38.6% | (35.7-41.6) | 5.5% | (4.7-6.3) | 24.0% | (20.7-27.4) |
| Age | 40.80 | (40.55-41.0) | 41.74 | (41.29-42.2) | 40.69 | (40.36-41.0) | 40.14 | (39.80-40.5) | 40.94 | (40.72-41.2) | 38.03 | (37.48-38.6) |
| Female | 51.0% | (50.4-51.6) | 37.8% | (37.2-38.5) | 52.0% | (51.1-52.9) | 50.1% | (49.3-50.8) | 59.4% | (58.8-59.9) | 43.9% | (42.7-45.1) |
| White | 59.6% | (57.2-61.9) | 46.5% | (43.3-49.7) | 66.3% | (63.3-69.3) | 67.1% | (64.9-69.3) | 60.7% | (58.7-62.8) | 61.1% | (57.6-64.5) |
| Black | 20.1% | (18.3-21.9) | 33.2% | (30.1-36.2) | 17.7% | (15.4-20.0) | 16.4% | (14.7-18.1) | 15.5% | (14.2-16.8) | 20.9% | (18.8-23.1) |
| Hispanic | 12.4% | (11.1-13.7) | 11.1% | (9.5-12.8) | 8.3% | (7.1-9.5) | 9.4% | (8.3-10.5) | 16.0% | (14.4-17.6) | 9.3% | (7.8-10.9) |
| Other race/ethnicity | 6.2% | (5.3-7.0) | 7.4% | (6.0-8.7) | 5.7% | (4.7-6.7) | 5.2% | (4.5-5.9) | 6.2% | (5.4-6.9) | 6.7% | (5.1-8.3) |
|
| ||||||||||||
| Insurance Status | ||||||||||||
|
| ||||||||||||
| Any insurance | 84.5% | (83.6-85.5) | 85.8% | (84.4-87.3) | 88.2% | (87.1-89.4) | 85.2% | (83.9-86.5) | 82.5% | (81.7-83.4) | 85.3% | (83.3-87.2) |
| Any Medicare | 23.6% | (22.9-24.3) | 32.9% | (31.5-34.2) | 27.7% | (26.6-28.9) | 19.7% | (18.9-20.4) | 19.5% | (19.0-20.1) | 25.0% | (23.4-26.5) |
| Any Medicaid | 42.6% | (41.0-44.3) | 53.2% | (50.6-55.8) | 46.1% | (43.8-48.4) | 40.4% | (38.6-42.3) | 36.3% | (35.0-37.7) | 51.5% | (49.1-53.8) |
| Any private Insurance | 26.4% | (25.1-27.7) | 13.0% | (11.8-14.2) | 24.6% | (22.9-26.3) | 30.9% | (29.2-32.5) | 32.4% | (31.2-33.5) | 20.0% | (18.2-21.7) |
|
| ||||||||||||
| Elixhauser Sum | 0.93 | (0.90-0.97) | 1.12 | (1.05-1.19) | 1.24 | (1.19-1.30) | 1.07 | (1.03-1.11) | 0.68 | (0.66-0.70) | 0.90 | (0.85-0.95) |
|
| ||||||||||||
National Emergency Department Sample, 2019, analysis limited to those age 18-90 years. Results based on weighted sampling. Elixhauser weighted sum does not include diagnoses of depression or psychosis.
Abbreviations. ANX=anxiety disorders, BPD=bipolar disorder, MDD=major depressive disorder, Oth dx=other diagnoses, SCZ=schizophrenia-spectrum diagnoses.
Visits aggregated by diagnoses differed in their racial/ethnic composition. Compared to their overall sample representation of 20.1%, Black patients comprised a higher percentage of schizophrenia-spectrum disorder visits (33.2%; 95%CI=30.1–36.2) and a lower percentage of MDD visits (16.4%; 95%CI=14.7–18.1) and anxiety disorder visits (15.5%; 95%CI=14.2–16.8) (Figure 1). Compared to their overall representation of 12.4%, Hispanic patients comprised 11.1% (95%CI=9.5–12.8) of schizophrenia-spectrum disorder visits, 9.4% (95%CI=8.3–10.5) of MDD visits and 16.0% (95%CI=14.4–17.8) of anxiety disorder visits.
Figure 1. ED mental health visits by primary diagnosis group and race/ethnicity, United States, 2019.
National Emergency Department Sample, 2019, analysis limited to those age 18–90 years. Results based on weighted sampling. Percentages in diagnostic group graphs do not sum to 100 due to missing data values for race/ethnicity (1.8% of total sample).
Abbreviations. ANX=anxiety disorders, BPD=bipolar disorder, MDD=major depressive disorder, Oth dx=other diagnoses, SCZ=schizophrenia-spectrum diagnoses.
Unadjusted Admission Rates
Inpatient admission rates varied significantly by diagnosis (Table 1). Bipolar disorder was associated with the highest overall admission rate (50.7%; 95%CI=47.2–54.2), followed by schizophrenia-spectrum disorders (41.5%; 95%CI=37.1–45.9). Anxiety disorders were associated with the lowest overall admission rate (5.5%; 95%CI=4.7–6.3).
The racial/ethnic composition of admitted patients was 61.9% White, 21.6% Black, 9.5% Hispanic, and 5.2% other (data not shown). For bipolar disorder visits, admission rates were 44.7% (95%CI=39.5–49.9) for Black, 49.3% (95%CI=43.5–55.2) for Hispanic, 53.1% (95%CI=49.6–56.5) for White and 45.6% (95%CI=38.2–53.1) for other race/ethnicity patients (Figure 2). For MDD visits, admission rates were 33.3% (95%CI=29.0–37.6) for Black patient visits, 34.9% (95%CI=29.9–39.9) for Hispanic patient visits, 32.9% (95%CI=28.2–37.6) for other race/ethnicity patient visits, and 40.9% (95%CI=37.9–43.9) for White patient visits. Race/ethnicity was also a significant predictor of inpatient admission for visits for anxiety disorders and other mental health disorders. There were no significant differences in admission rates by race/ethnicity for visits for schizophrenia-spectrum disorders.
Figure 2. Admission rates for mental health ED visits, by diagnosis group and race/ethnicity.
National Emergency Department Sample, 2019, analysis limited to those age 18–90 years. N=674,821. 95% confidence interval given by error bars. Percentages based on weighted sampling. Independent F-tests for differences between rates of admission by racial/ethnic group within diagnostic category were inclusive of White, Black, Hispanic, and other race/ethnicity (not shown); omnibus p-values displayed above diagnostic category groups.
Abbreviations. ANX=anxiety disorders, BPD=bipolar disorder, MDD=major depressive disorder, Oth dx=other diagnoses, SCZ=schizophrenia-spectrum diagnoses.
Adjusted Admission Rates
In an adjusted model for all visits, race/ethnicity was not significantly associated with inpatient admission (Table 2).
Table 2.
Associations of patient insurance, race/ethnicity, sex, and discharge diagnosis with emergency department disposition among emergency department mental health visits.
| Unadjusted | Covariate Adjusted | |||
|---|---|---|---|---|
|
| ||||
| OR | 95%CI | OR | 95% CI | |
|
| ||||
| White (reference) | 1 | 1 | ||
| Black | 1.05 | (0.87-1.27) | 0.85 | (0.70-1.02) |
| Hispanic | 0.68** | (0.55-0.83) | 0.85 | (0.71-1.01) |
| Other race/ethnicity | 0.76** | (0.64-0.90) | 0.82 | (0.66-1.01) |
| Any Insurance | 1.64** | (1.30-2.09) | ||
| Any Private Insurance | 1.12** | (1.01-1.25) | ||
| Male (reference) | 1 | |||
| Female | 1.07** | (1.04-1.11) | ||
| Psychiatric Diagnoses | ||||
| SCZ (reference) | 1 | |||
| BPD | 1.27** | (1.14-1.41) | ||
| MDD | 0.81** | (0.72-0.91) | ||
| ANX | 0.09** | (0.07-0.11) | ||
| Other | 0.42** | (0.35-0.49) | ||
N=674,811. Binary outcome variable is an indicator of admission/transfer. Reference category is male, white race/ethnicity, no insurance, schizophrenia-spectrum diagnosis. Any private insurance reflects both primary and secondary expected payers and is a nested category within Any insurance. Reference category is male, white race/ethnicity, no insurance, schizophrenia-spectrum diagnosis. Unadjusted models contains race/ethnicity categories and constant term alone. Adjusted model includes covariates shown as well as 10-year age group and indicator variables for Elixhauser comorbidities excluding depression and psychosis.
indicates that the Wald statistic is significant at the 0.01 level.
Abbreviations. ANX=anxiety disorders, BPD=bipolar affective disorder, MDD=major depressive disorder.
Among bipolar disorder visits (Table 3), Black patient visits had lower odds of admission than White patient visits (OR=0.71; 95%CI=0.59–0.84). These results primarily reflected differences in admission rates among those presenting with bipolar depression (OR=0.75; 95%CI=0.57–0.99), or unspecified episode type (OR=0.60; 95%CI=0.4–0.76) compared to mania. Inpatient admission for MDD visits also differed between White and Black patients even after adjustment for covariates with Black patients having a lower odds of admission (OR=0.70; 95%CI=0.59–0.83). Compared to White patient visits, inpatient admission for anxiety disorder visits were also significantly lower for Black (OR=0.61; 95%CI=0.50–0.74) and Hispanic (OR=0.67; 95%CI=0.49–0.91) patient visits. Racial/ethnic differences in admission for other diagnosis were observed for all non-White groups. No racial/ethnic differences among admission rates for visits for schizophrenia-spectrum disorders were observed.
Table 3.
Associations of patient insurance, race/ethnicity, and discharge diagnosis with emergency department disposition among emergency department mental health visits, subset analyses by discharge diagnosis group.
| SCZ (N=154,794) | BPD (N=65,056) | |||
|---|---|---|---|---|
|
| ||||
| OR | 95% CI | OR | 95% CI | |
|
| ||||
| White | 1.00 | 1.00 | ||
| Black | 1.04 | (0.82 –1.33) | 0.71** | (0.59 –0.84) |
| Hispanic | 1.06 | (0.83 –1.36) | 0.93 | (0.76 –1.13) |
| Other race | 0.84 | (0.64 –1.11) | 0.89 | (0.67 –1.18) |
| Any Insurance | 1.81** | (1.31 –2.51) | 1.74** | (1.31 –2.31) |
| Any Private Insurance | 0.96 | (0.82 –1.12) | 1.25** | (1.08 –1.45) |
| Male | ||||
| Female | 1.04 | (0.99 –1.08) | 1.15** | (1.08 –1.22) |
|
| ||||
| MDD (N=154,681) | BPD, mania (N=18,919) | |||
| OR | 95% CI | OR | 95% CI | |
|
| ||||
| White | 1.00 | 1.00 | ||
| Black | 0.70** | (0.59 –0.83) | 1.09 | (0.91 –1.31) |
| Hispanic | 0.85 | (0.70 –1.02) | 1.02 | (0.78 –1.34) |
| Other race | 0.83 | (0.68 –1.03) | 1.17 | (0.87 –1.56) |
| Any Insurance | 1.61** | (1.27 –2.03) | 1.90** | (1.38 –2.63) |
| Any Private Insurance | 1.25** | (1.11 –1.42) | 0.97 | (0.83 –1.14) |
| Male | 1.00 | 1.00 | ||
| Female | 1.13** | (1.07 –1.18) | 1.09** | (1.00 –1.19) |
|
| ||||
| Anxiety (N=273,440) | BPD, depression (N=11,233) | |||
| OR | 95% CI | OR | 95% CI | |
|
| ||||
| White | 1.00 | 1.00 | ||
| Black | 0.61** | (0.50 –0.74) | 0.75** | (0.57 –0.99) |
| Hispanic | 0.67** | (0.49 –0.91) | 0.89 | (0.65 –1.22) |
| Other race | 0.52** | (0.35 –0.76) | 0.82 | (0.54 –1.24) |
| Any Insurance | 1.37** | (1.03 –1.84) | 1.84** | (1.27 –2.68) |
| Any Private Insurance | 1.07 | (0.90 –1.28) | 1.26** | (1.01 –1.58) |
| Male | 1.00 | 1.00 | ||
| Female | 1.22** | (1.10 –1.36) | 1.05 | (0.94 –1.17) |
|
| ||||
| Other dx (N=26,771) | BPD, unspecified state (N=34,893) | |||
| OR | 95% CI | OR | 95% CI | |
|
| ||||
| White | 1.00 | 1.00 | ||
| Black | 0.61** | (0.50 –0.74) | 0.60** | (0.47 –0.76) |
| Hispanic | 0.67** | (0.49 –0.91) | 0.85 | (0.67 –1.08) |
| Other race | 0.52** | (0.35 –0.76) | 0.76 | (0.57 –1.02) |
| Any Insurance | 1.37** | (1.03 –1.84) | 1.59** | (1.16 –2.18) |
| Any Private Insurance | 1.07 | (0.90 –1.28) | 1.31** | (1.10 –1.56) |
| Male | 1.00 | 1.00 | ||
| Female | 1.22** | (1.10 –1.36) | 1.19** | (1.10 –1.30) |
Eight logistic regressions with outcome variable indicator of admission/transfer. The three models BPD, mania, BPD, depression and BPD, unspecified are subset analyses of BPD. Reference category is male, white race/ethnicity, no insurance, schizophrenia-spectrum diagnosis. Any private insurance reflects both primary and secondary expected payers and is a nested category within Any insurance. All models include covariates shown as well as 10-year age group and indicator variables for Elixhauser comorbidities excluding depression and psychosis.
indicates that the Wald statistic is significant at the 0.01 level.
Abbreviations. ANX=anxiety disorders, BPD=bipolar disorder, MDD=major depressive disorder, Oth dx=other diagnoses, SCZ=schizophrenia-spectrum diagnoses.
Subset analyses (data not shown)
Affective and Psychotic Symptom Variation
A higher proportion of White patient visits (36.9%; 95%CI=35.7–38.2) than Black (27.5%; 95%CI=26.0–29.0) or Hispanic (24.1%; 95%CI=22.6–25.7) patient visits were for mood disorders. Among MDD visits, a higher proportion of White (28.3%; 95%CI=26.0–30.5) than Hispanic (25.6%; 95%CI=22.1–29.1) or Black (22.9%; 95%CI=20.5–25.3) patient visits were for severe disorders. The proportion of schizoaffective disorder diagnosis among visits for schizophrenia-spectrum diagnosis did not differ significantly by race/ethnicity.
Schizophrenia-spectrum disorder visits accounted for a higher proportion of Black (37.2%; 95%CI=34.7–39.7) than Hispanic (20.3%; 95%CI=17.9–22.7) or White (17.6%; 95%CI=16.8–18.4) patient visits. Among MDD visits, Hispanic patients had the highest proportion with psychotic features (9.0%, 95%CI=6.9–11.0), followed by Black (7.9%; 95%CI=6.9–8.9), and White (5.7%; 95%CI=5.1–6.3) patients.
Admission Rates for Men with Schizophrenia-Spectrum Illness
Among schizophrenia-spectrum disorder visits, Black patients had a higher proportion for paranoid schizophrenia visits (12.4%; 95%CI=10.0–14.8) than Hispanic (9.3%; 95%CI=7.7–10.9) or White (8.4%; 95%CI=7.6–9.2) patients. Among male schizophrenia-spectrum disorder visits, there were no significant racial/ethnic differences in rates of inpatient admission.
Discussion
On a population basis, ED mental health visit rates differed substantially by race/ethnicity with Black individuals having the highest rate followed by White individuals then Hispanic and other individuals. We found some unadjusted differences in admission rates by race/ethnicity, most notably for Hispanic patients, however, after controlling for insurance status, mental health diagnosis, sex, age, and medical comorbidities, we did not find differences in the likelihood of inpatient admission. Inpatient admission rates differed by race/ethnicity within diagnostic subsets: among anxiety disorder visits, Hispanic patients were admitted at lower rates than White patients and among anxiety and mood disorders, Black patients were admitted at lower rates than White patients. We did not find higher rates of admission for Black compared to White patients, including for men with schizophrenia-spectrum diagnoses.
The discrepancy between overall, diagnostic category, and subset findings illustrates several features of racial/ethnic disparities. The literature has described robust correlations between race/ethnicity and other predictors of inpatient admission such as insurance status32,39 and medical comorbidity40, which influence the likelihood of admission. Insurance status is significantly correlated with inpatient admission, and we found the tendency for Black or Hispanic patients to be uninsured relative to White patients contributed to lower overall likelihood of admission for Black and Hispanic patients. Greater medical comorbidity was associated with increased likelihood of admission. On average, Black patients had higher rates of medical comorbidity compared to White patients, who had higher rates than Hispanic patients. The tendency for Hispanic patients to have lower average medical comorbidity contributed to their lower likelihood of admission compared to White or Black patients.
In contrast to prior epidemiologic literature suggesting that, conditional on diagnosis, severity is higher among Black and Hispanic compared to White patients,2 we found that 1) clinically diagnosed severity within MDD was lower among Black and Hispanic patients compared to White patients and 2) admission rates were lower for both mood and anxiety disorders among Black than White patients and lower for anxiety disorders among Hispanic than White patients, suggesting lower severity. These findings would be consistent with a hypothesis that even controlling for insurance status, Black and Hispanic patients have poorer access to outpatient services and may substitute with higher rates of ED use6,7 over a hypothesis that lower outpatient service use rates are driven by a higher symptom severity threshold among Black and Hispanic patients to seek mental health emergency care.
Psychiatric diagnosis was strongly correlated with disposition with anxiety disorders associated with the lowest odds of admission. We reproduce a previously reported finding that Black patients were overrepresented in schizophrenia-spectrum disorders.13,19,20 By contrast, Hispanic patients had higher representation than other racial/ethnic groups among visits for anxiety disorders, a finding that contrasts with population prevalence studies that have described lower rates of anxiety disorders among Hispanic individuals in the U.S. compared to non-Hispanic Whites.23
There has been significant interest in explanations for racial/ethnic diagnostic differences. Prior work from the ED has suggested that providers may collect less or lower quality information from Black than White patients, particularly about mood symptoms.14 Others have suggested that some providers may inappropriately interpret reserve among Black patients due to normative wariness of healthcare settings as signs of psychosis,36 or use different processes to link symptoms to diagnosis in Black versus White patients.13,15,17,36,37 Due to data limitations, we can only comment on these hypotheses to a limited extent. We found that differences between Black and White patients in admission rates for mood disorders were driven by visits for depressed mood over mania. We also found that rates of diagnosed psychosis were higher among Black and Hispanic patients including those presenting with MDD. Previous literature has suggested provider management choices for depression may differ by patient race/ethnicity.41 Alternatively, diagnostic error and symptom severity may be correlated. For example, if the subset of Black patients with the most severe depression were more likely to be misdiagnosed with a schizophrenia-spectrum diagnosis,17 this would leave a healthier group of patients among the population diagnosed with MDD without psychotic features for Black compared to White patients. However, without independent diagnostic assessments, we cannot test these hypotheses. Similarly some evidence suggests that expression of some anxiety symptoms differ between Hispanic and non-Hispanic Whites,22 which may affect provider-assigned diagnoses. However, diagnostic studies of Hispanic patients have focused more on generating instruments to match clinical diagnoses than on measuring clinician bias.42
Despite the diagnostic composition of visits by Black versus White patients, rates of inpatient admission did not differ between these groups. Higher rates of Black inpatient admissions in our sample reflected higher ED visit per population rates among Black patients over measurable biases towards inpatient admission for Black patients conditional on ED presentation. These findings highlight the relative importance of racial/ethnic differences in pathways to clinical presentation,43–45 outpatient care access, as well as broader socioeconomic, legal, historical, and structural factors18 in generating existing disparities in psychiatric care and suggest these as avenues for future study.
This study has several limitations. First, we lacked information on history-taking practices, relevant symptoms, social supports, structured instruments for interview assessment, or other relevant clinical data available in prior studies that have explicitly focused on the nature of psychiatric diagnosis. Second, we are limited by the granularity of race/ethnicity data available in this record. We cannot report on collection practices for race/ethnicity variables. Further, race and ethnicity are imprecise social constructs.46 Within broad ethnic categories, such as Hispanic, rates of psychological distress, experience, geography and other factors vary considerably based on country of origin, foreign- versus U.S.-born, and many other variables we are unable to examine here.47,48 We are limited further by power in the examination of patients of non-White, non-Black, non-Hispanic race/ethnicity. Third, we cannot distinguish medical admissions from psychiatric inpatient admissions, nor can we determine whether admissions were voluntary or involuntary. Consequently, we consider our findings to be exploratory of differences between admitting practices by race/ethnicity stratified by diagnosis. Finally, our analyses were performed at the visit-level, which overweights higher acuity patients who may have presented repeatedly within a year relative to a person-level analysis.
Conclusions
The ED can play an important role in directing patients with mental health symptoms. Prior research has suggested that ED triage practices contribute to greater representation of Black than White patients on inpatient units.14 In this nationally representative sample of emergency department visits for mental health conditions, we found no differences in the overall admission rates for Black or Hispanic versus White patients. After adjusting for covariates, we further found no differences in overall admission rates across race/ethnicity, though findings from diagnostic subsets suggest systemic differences in the ways that patients of different race/ethnicity interact with mental health care systems. These findings from a nationally representative survey may help us to understand how to provide more equitable access to mental health services.
Supplementary Material
Clinical Points.
Racial and ethnic disparities in acute psychiatric settings and emergency departments have been described for many years in the United States, however, nationally representative data permitting the population-level study of racial/ethnic disparities in acute psychiatric practice settings have not been widely available until recently.
Nationally, racial/ethnic differences in admission practices were greater for patients presenting to the emergency department with depression or anxiety than for schizophrenia or bipolar mania.
Racial/ethnic differences by diagnosis are persistent, however, care access, socioeconomic, historical, legal, and structural factors contribute significantly to disparities in acute psychiatric admission practices.
Acknowledgements
YNG was supported in part by a Moynihan Clinical Research Fellowship from the Leon Levy Foundation and Award Number R25MH086466 from the National Institute of Mental Health. Neither funding organization had any role in the design, development, analysis, or interpretation of this study.
Footnotes
COI
The authors have no conflicts of interest to disclose.
References
- 1.Chow JCC, Jaffee K, Snowden L. Racial/Ethnic Disparities in the Use of Mental Health Services in Poverty Areas. Am J Public Health. 2003;93(5):792–797. doi: 10.2105/AJPH.93.5.792 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.McGuire TG, Miranda J. New Evidence Regarding Racial And Ethnic Disparities In Mental Health: Policy Implications. Health Affairs. 2008;27(2):393–403. doi: 10.1377/hlthaff.27.2.393 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Hu TW, Snowden LR, Jerrell JM, Nguyen TD. Ethnic populations in public mental health: services choice and level of use. Am J Public Health. 1991;81(11):1429–1434. doi: 10.2105/AJPH.81.11.1429 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Theriault KM, Rosenheck RA, Rhee TG. Increasing Emergency Department Visits for Mental Health Conditions in the United States. J Clin Psychiatry. 2020;81(5). doi: 10.4088/JCP.20m13241 [DOI] [PubMed] [Google Scholar]
- 5.Lawson WB, Hepler N, Holladay J, Cuffel B. Race as a factor in inpatient and outpatient admissions and diagnosis. Hosp Community Psychiatry. 1994;45(1):72–74. doi: 10.1176/ps.45.1.72 [DOI] [PubMed] [Google Scholar]
- 6.Snowden LR, Cheung FK. Use of inpatient mental health services by members of ethnic minority groups. Am Psychol. 1990;45(3):347–355. doi: 10.1037//0003-066x.45.3.347 [DOI] [PubMed] [Google Scholar]
- 7.Snowden LR, Hastings JF, Alvidrez J. Overrepresentation of Black Americans in Psychiatric Inpatient Care. PS. 2009;60(6):779–785. doi: 10.1176/ps.2009.60.6.779 [DOI] [PubMed] [Google Scholar]
- 8.Cook BL, Trinh NH, Li Z, Hou SSY, Progovac AM. Trends in Racial-Ethnic Disparities in Access to Mental Health Care, 2004–2012. PS. 2017;68(1):9–16. doi: 10.1176/appi.ps.201500453 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Ojeda VD, McGuire TG. Gender and Racial/Ethnic Differences in Use of Outpatient Mental Health and Substance Use Services by Depressed Adults. Psychiatr Q. 2006;77(3):211–222. doi: 10.1007/s11126-006-9008-9 [DOI] [PubMed] [Google Scholar]
- 10.Giacco D, Matanov A, Priebe S. Providing mental healthcare to immigrants: current challenges and new strategies. Current Opinion in Psychiatry. 2014;27(4):282–288. doi: 10.1097/YCO.0000000000000065 [DOI] [PubMed] [Google Scholar]
- 11.Campbell RD, Long LA. Culture as a Social Determinant of Mental and Behavioral Health: A Look at Culturally Shaped Beliefs and Their Impact on Help-Seeking Behaviors and Service Use Patterns of Black Americans with Depression. Best Practices in Mental Health. 2014;10(2):48–62. [Google Scholar]
- 12.Nadeem E, Lange JM, Edge D, Fongwa M, Belin T, Miranda J. Does Stigma Keep Poor Young Immigrant and U.S.-Born Black and Latina Women From Seeking Mental Health Care? PS. 2007;58(12):1547–1554. doi: 10.1176/ps.2007.58.12.1547 [DOI] [PubMed] [Google Scholar]
- 13.Strakowski SM, Flaum M, Amador X, et al. Racial differences in the diagnosis of psychosis. Schizophrenia Research. 1996;21(2):117–124. doi: 10.1016/0920-9964(96)00041-2 [DOI] [PubMed] [Google Scholar]
- 14.Strakowski SM, Hawkins JM, Keck PE, et al. The effects of race and information variance on disagreement between psychiatric emergency service and research diagnoses in first-episode psychosis. J Clin Psychiatry. 1997;58(10):457–463; quiz 464–465. doi: 10.4088/jcp.v58n1010a [DOI] [PubMed] [Google Scholar]
- 15.Neighbors HW, Trierweiler SJ, Ford BC, Muroff JR. Racial Differences in DSM Diagnosis Using a Semi-Structured Instrument: The Importance of Clinical Judgment in the Diagnosis of African Americans. Journal of Health and Social Behavior. 2003;44(3):237–256. doi: 10.2307/1519777 [DOI] [PubMed] [Google Scholar]
- 16.Strakowski SM, Shelton RC, Kolbrener ML. The effects of race and comorbidity on clinical diagnosis in patients with psychosis. The Journal of Clinical Psychiatry. 1993;54(3):96–102. [PubMed] [Google Scholar]
- 17.Strakowski SM, Keck PE, Arnold LM, et al. Ethnicity and Diagnosis in Patients With Affective Disorders. J Clin Psychiatry. 2003;64(7):747–754. doi: 10.4088/JCP.v64n0702 [DOI] [PubMed] [Google Scholar]
- 18.Rosenfield S Race Differences in Involuntary Hospitalization: Psychiatric vs. Labeling Perspectives. Journal of Health and Social Behavior. 1984;25(1):14–23. doi: 10.2307/2136701 [DOI] [PubMed] [Google Scholar]
- 19.Simon RJ, Fleiss JL, Gurland BJ, Stiller PR, Sharpe L. Depression and schizophrenia in hospitalized black and white mental patients. Arch Gen Psychiatry. 1973;28(4):509–512. doi: 10.1001/archpsyc.1973.01750340047007 [DOI] [PubMed] [Google Scholar]
- 20.Raskin A, Crood TH, Herman KD. Psychiatric history and symptom difference in black and white depressed inpatients. J Consult Clin Psychol. 1975;43(1):73–80. doi: 10.1037/h0076322 [DOI] [PubMed] [Google Scholar]
- 21.Metzl J The Protest Psychosis: How Schizophrenia Became a Black Disease. Beacon; 2011. [Google Scholar]
- 22.Cintrón JA, Carter MM, Suchday S, Sbrocco T, Gray J. Factor structure and construct validity of the Anxiety Sensitivity Index among island Puerto Ricans. Journal of Anxiety Disorders. 2005;19(1):51–68. doi: 10.1016/j.janxdis.2003.10.007 [DOI] [PubMed] [Google Scholar]
- 23.Asnaani A, Richey JA, Dimaite R, Hinton DE, Hofmann SG. A cross-ethnic comparison of lifetime prevalence rates of anxiety disorders. J Nerv Ment Dis. 2010;198(8):551–555. doi: 10.1097/NMD.0b013e3181ea169f [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Betz ME, Wintersteen M, Boudreaux ED, et al. Reducing Suicide Risk: Challenges and Opportunities in the Emergency Department. Annals of Emergency Medicine. 2016;68(6):758–765. doi: 10.1016/j.annemergmed.2016.05.030 [DOI] [PubMed] [Google Scholar]
- 25.Olfson M, Wall M, Wang S, Crystal S, Gerhard T, Blanco C. Suicide Following Deliberate Self-Harm. Am J Psychiatry. 2017;174(8):765–774. doi: 10.1176/appi.ajp.2017.16111288 [DOI] [PubMed] [Google Scholar]
- 26.Doupnik SK, Rudd B, Schmutte T, et al. Association of Suicide Prevention Interventions With Subsequent Suicide Attempts, Linkage to Follow-up Care, and Depression Symptoms for Acute Care Settings: A Systematic Review and Meta-analysis. JAMA Psychiatry. 2020;77(10):1021–1030. doi: 10.1001/jamapsychiatry.2020.1586 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Lewis-Fernández R, Aggarwal NK, Bäärnhielm S, et al. Culture and Psychiatric Evaluation: Operationalizing Cultural Formulation for DSM-5. Psychiatry. 2014;77(2):130–154. doi: 10.1521/psyc.2014.77.2.130 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Sohler NL, Bromet EJ. Does racial bias influence psychiatric diagnoses assigned at first hospitalization? Soc Psychiatry Psychiatr Epidemiol. 2003;38(8):463–472. doi: 10.1007/s00127-003-0653-0 [DOI] [PubMed] [Google Scholar]
- 29.Adult population by race | KIDS COUNT Data Center. Accessed August 29, 2022. https://datacenter.kidscount.org/data/tables/6539-adult-population-by-race [Google Scholar]
- 30.Zhu JM, Singhal A, Hsia RY. Emergency Department Length-Of-Stay For Psychiatric Visits Was Significantly Longer Than For Nonpsychiatric Visits, 2002–11. Health Affairs. 2016;35(9):1698–1706. doi: 10.1377/hlthaff.2016.0344 [DOI] [PubMed] [Google Scholar]
- 31.Healthcare Cost and Utilization Project (HCUP) NEDS Notes. Accessed January 10, 2023. https://www.hcup-us.ahrq.gov/db/vars/race/nedsnote.jsp
- 32.Gao YN, Olfson M. Insurance and inpatient admission of emergency department patients with depression in the United States. General Hospital Psychiatry. 2022;78:28–34. doi: 10.1016/j.genhosppsych.2022.07.003 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33.Elixhauser A, Steiner C, Harris DR, Coffey RM. Comorbidity measures for use with administrative data. Med Care. 1998;36(1):8–27. doi: 10.1097/00005650-199801000-00004 [DOI] [PubMed] [Google Scholar]
- 34.Manitoba Centre for Health Policy, Garland A. The Epidemiology and Outcomes of Critical Illness in Manitoba. University of Manitoba, Faculty of Medicine, Dept. of Community Health Sciences; 2012. [Google Scholar]
- 35.Stagg V ELIXHAUSER: Stata module to calculate Elixhauser index of comorbidity. Statistical Software Components. Published online September 16, 2015. Accessed April 21, 2022. https://ideas.repec.org/c/boc/bocode/s458077.html [Google Scholar]
- 36.Trierweiler SJ, Muroff JR, Jackson JS, Neighbors HW, Munday C. Clinician race, situational attributions, and diagnoses of mood versus schizophrenia disorders. Cultural Diversity and Ethnic Minority Psychology. 2005;11(4):351–364. doi: 10.1037/1099-9809.11.4.351 [DOI] [PubMed] [Google Scholar]
- 37.Adebimpe VR. Overview: white norms and psychiatric diagnosis of black patients. Am J Psychiatry. 1981;138(3):279–285. doi: 10.1176/ajp.138.3.279 [DOI] [PubMed] [Google Scholar]
- 38.American Psychiatric Association, ed. Diagnostic and Statistical Manual of Mental Disorders: DSM-5-TR. Fifth edition, text revision. American Psychiatric Association Publishing; 2022. [Google Scholar]
- 39.Ruger JP, Richter CJ, Lewis LM. Association between Insurance Status and Admission Rate for Patients Evaluated in the Emergency Department. Academic Emergency Medicine. 2003;10(11):1285–1288. doi: 10.1197/S1069-6563(03)00500-1 [DOI] [PubMed] [Google Scholar]
- 40.Wang HY, Chew G, Kung CT, Chung KJ, Lee WH. The use of Charlson comorbidity index for patients revisiting the emergency department within 72 hours. Chang Gung Med J. 2007;30(5):437–444. [PubMed] [Google Scholar]
- 41.McGuire TG, Ayanian JZ, Ford DE, Henke REM, Rost KM, Zaslavsky AM. Testing for Statistical Discrimination by Race/Ethnicity in Panel Data for Depression Treatment in Primary Care. Health Services Research. 2008;43(2):531–551. doi: 10.1111/j.1475-6773.2007.00770.x [DOI] [PMC free article] [PubMed] [Google Scholar]
- 42.Snipes C Assessment of Anxiety with Hispanics. In: Benuto LT, ed. Guide to Psychological Assessment with Hispanics. Springer US; 2013:153–162. doi: 10.1007/978-1-4614-4412-1_11 [DOI] [Google Scholar]
- 43.Rosen DL, Travers D. Emergency department visits among patients transported by law enforcement officers. PLOS ONE. 2021;16(1):e0244679. doi: 10.1371/journal.pone.0244679 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 44.Rotenberg M, Tuck A, Ptashny R, McKenzie K. The role of ethnicity in pathways to emergency psychiatric services for clients with psychosis. BMC Psychiatry. 2017;17(1):137. doi: 10.1186/s12888-017-1285-3 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 45.Hanchate AD, Paasche-Orlow MK, Baker WE, Lin MY, Banerjee S, Feldman J. Association of Race/Ethnicity With Emergency Department Destination of Emergency Medical Services Transport. JAMA Network Open. 2019;2(9):e1910816. doi: 10.1001/jamanetworkopen.2019.10816 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 46.Flanagin A, Frey T, Christiansen SL, AMA Manual of Style Committee. Updated Guidance on the Reporting of Race and Ethnicity in Medical and Science Journals. JAMA. 2021;326(7):621–627. doi: 10.1001/jama.2021.13304 [DOI] [PubMed] [Google Scholar]
- 47.Alegría M, Canino G, Shrout PE, et al. Prevalence of Mental Illness in Immigrant and Non-Immigrant U.S. Latino Groups. AJP. 2008;165(3):359–369. doi: 10.1176/appi.ajp.2007.07040704 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 48.González HM, Tarraf W, Whitfield KE, Vega WA. The epidemiology of major depression and ethnicity in the United States. Journal of Psychiatric Research. 2010;44(15):1043–1051. doi: 10.1016/j.jpsychires.2010.03.017 [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.