Racial and Ethnic Disparities in Receipt of Medications for Treatment of COVID-19

United States, March 2020-August 2021

Jennifer L. Wiltz, MD; Amy K. Feehan, PhD; NoelleAngelique M. Molinari, PhD; Chandresh N. Ladva, PhD; Benedict I. Truman, MD; Jeffrey Hall, PhD; Jason P. Block, MD; Sonja A. Rasmussen, MD; Joshua L. Denson, MD; William E. Trick, MD; Mark G. Weiner, MD; Emily Koumans, MD; Adi Gundlapalli, MD, PhD; Thomas W. Carton, PhD; Tegan K. Boehmer, PhD

Disclosures

Morbidity and Mortality Weekly Report. 2022;71(3):96-102.

Abstract and Introduction

Introduction

The COVID-19 pandemic has magnified longstanding health care and social inequities, resulting in disproportionately high COVID-19–associated illness and death among members of racial and ethnic minority groups.^[1] Equitable use of effective medications^[2] could reduce disparities in these severe outcomes.^[3] Monoclonal antibody (mAb) therapies against SARS-CoV-2, the virus that causes COVID-19, initially received Emergency Use Authorization (EUA) from the Food and Drug Administration (FDA) in November 2020. mAbs are typically administered in an outpatient setting via intravenous infusion or subcutaneous injection and can prevent progression of COVID-19 if given after a positive SARS-CoV-2 test result or for postexposure prophylaxis in patients at high risk for severe illness.^† Dexamethasone, a commonly used steroid, and remdesivir, an antiviral drug that received EUA from FDA in May 2020, are used in inpatient settings and help prevent COVID-19 progression^§.^[2] No large-scale studies have yet examined the use of mAb by race and ethnicity. Using COVID-19 patient electronic health record data from 41 U.S. health care systems that participated in the PCORnet, the National Patient-Centered Clinical Research Network,^¶ this study assessed receipt of medications for COVID-19 treatment by race (White, Black, Asian, and Other races [including American Indian or Alaska Native, Native Hawaiian or Other Pacific Islander, and multiple or Other races]) and ethnicity (Hispanic or non-Hispanic). Relative disparities in mAb** treatment among all patients^†† (805,276) with a positive SARS-CoV-2 test result and in dexamethasone and remdesivir treatment among inpatients^§§ (120,204) with a positive SARS-CoV-2 test result were calculated. Among all patients with positive SARS-CoV-2 test results, the overall use of mAb was infrequent, with mean monthly use at 4% or less for all racial and ethnic groups. Hispanic patients received mAb 58% less often than did non-Hispanic patients, and Black, Asian, or Other race patients received mAb 22%, 48%, and 47% less often, respectively, than did White patients during November 2020–August 2021. Among inpatients, disparities were different and of lesser magnitude: Hispanic inpatients received dexamethasone 6% less often than did non-Hispanic inpatients, and Black inpatients received remdesivir 9% more often than did White inpatients. Vaccines and preventive measures are the best defense against infection; use of COVID-19 medications postexposure or postinfection can reduce morbidity and mortality and relieve strain on hospitals but are not a substitute for COVID-19 vaccination. Public health policies and programs centered around the specific needs of communities can promote health equity.^[4] Equitable receipt of outpatient treatments, such as mAb and antiviral medications, and implementation of prevention practices are essential to reducing existing racial and ethnic inequities in severe COVID-19–associated illness and death.

The PCORnet-distributed data infrastructure was queried,^¶¶ and 41 sites*** returned data on monthly receipt of medications for COVID-19 treatment during March 2020–August 2021. The monthly percentage of patients with a positive SARS-CoV-2 test result who received mAb (November 2020–August 2021) and of inpatients with a SARS-CoV-2 positive test result who received dexamethasone or remdesivir (March 2020–August 2021) was calculated separately by race and by ethnicity (as aggregated in PCORnet) for adults aged ≥20 years. Differences in treatment by race and ethnicity were assessed in two ways. First, pairwise Wilcoxon signed rank tests, with p-values indicated as p_w, were used to assess whether treatment receipt differed systematically over time (systematic temporal differences) by race or ethnicity. Second, relative monthly treatment disparities were calculated as the difference in percentage of patients treated between racial or ethnic minority (Black, Asian, Other for race; Hispanic ethnicity) and majority (White; non-Hispanic) groups divided by the percentage treated in the majority groups for each month.^††† The grand means (means of relative monthly treatment disparities) were calculated, and t-tests for statistical difference from zero, with p-values indicated as p_t, were used to assess presence of overall relative treatment disparities. Results were considered statistically significant for p-values <0.05. GraphPad Prism software (version 9.3.0; GraphPad Software, Inc) was used for analyses and visualization. This activity was reviewed by CDC and conducted consistent with applicable federal law and CDC policy.^§§§

During March 2020–August 2021, a total of 5,918,199 patients in PCORnet health care systems were tested^¶¶¶ for SARS-CoV-2, and 805,276 (13.6%) test results were positive (Table 1), representing approximately 3.0% of all positive results reported to CDC (Supplementary Table, https://stacks.cdc.gov/view/cdc/113252). These patients are similar demographically to those included in CDC case data by age, sex, race, and ethnicity. Geographically, patients in the Census Pacific division are underrepresented whereas those in the Mountain division are overrepresented. Among patients with a positive test result, 2.9% were Asian, 15.7% Black, 61.2% White, and 10.9% Other race; 18.6% were Hispanic and 71.7% were non-Hispanic ethnicity (Table 1). Compared with all persons with a positive SARS-CoV-2 test result, a higher proportion of patients with high-risk comorbidities**** were treated with mAb. Critical care^†††† was required by 3.4% of all persons with positive test results compared with 1.8% of those treated with mAb.

Mean monthly mAb use among all patients with positive SARS-CoV-2 test results who were White, Black, Asian, or Other race was 4.0%, 2.8%, 2.2%, and 2.2%, respectively; among patients of Hispanic or non-Hispanic ethnicity, mAb use was 1.8% and 4.0%, respectively. Patients who were Black, Asian, or Other race received mAb 22.4%, 48.3%, and 46.5%, respectively, less often than did White patients (Table 2); systematic temporal differences in mAb receipt were observed by race (all p_w<0.01) (Figure). SARS-CoV-2 positive patients of Hispanic ethnicity received mAb 57.7% less often (p_t<0.001) than did non-Hispanic patients; systematic temporal differences in mAb receipt were observed by ethnicity (p_w = 0.002).

(Enlarge Image)

Figure.

Monthly* percentage of COVID-19 patients (n = 805,276) receiving monoclonal antibody treatment,^† by race^§ and ethnicity^¶ — 41 health care systems in the National Patient-Centered Clinical Research Network — United States, November 2020–August 2021
*Systematic temporal differences in medication receipt by race and ethnicity were assessed by pairwise Wilcoxon signed rank test.
^†mAbs require administration by intravenous infusion or subcutaneous injection.
^§White race is the referent group; p-values for Black, Asian, and Other races are 0.004, 0.002, and 0.002, respectively.
^¶Non-Hispanic ethnicity is the referent group; p = 0.002 for Hispanic ethnicity.

Mean monthly dexamethasone use among inpatients who were White, Black, Asian, or Other race was 35.8%, 33.8%, 31.4%, and 34.2%, respectively; among patients of Hispanic or non-Hispanic ethnicity, dexamethasone use was 32.5% and 35.4%, respectively. Relative disparities in dexamethasone receipt by race were not statistically significant (Table 2); however, small but systematic temporal differences in dexamethasone receipt were observed among White inpatients and Black, and Asian inpatients (both p_w<0.05) (Supplementary Figure, https://stacks.cdc.gov/view/cdc/113252). Hispanic inpatients were treated with dexamethasone 6.2% less often than were non-Hispanic inpatients and systematic temporal treatment differences were also observed (p_w = 0.005).

Mean monthly remdesivir use among inpatients who were White, Black, Asian, or Other race was 29.0%, 31.2%, 26.2%, and 30.6%, respectively; among patients of Hispanic or non-Hispanic ethnicity, remdesivir use was 30.4% and 29.3%, respectively. Black inpatients received remdesivir 9.3% more often (p_t = 0.03) than did White inpatients; systematic temporal differences were also observed (p_w = 0.03). Asian, Other race, and Hispanic inpatients did not experience significant relative disparities or systematic temporal differences in remdesivir treatment compared with White and non-Hispanic inpatients.

1 2

Next Section

Abstract and Introduction
Discussion
References
Sidebar

*These authors contributed equally to this report.
^†Fact sheets for healthcare providers for FDA emergency use authorization are available from https://www.fda.gov/media/145611/download for REGEN-COV (casirivimab and imdevimab) and https://www.fda.gov/media/145802/download for bamlanivimab and etesevimab. The SARS-CoV-2 B.1.1.529 (Omicron) variant is not neutralized by bamlanivimab and etesevimab or casirivimab and imdevimab, the mAb-based COVID-19 treatments that were most frequently prescribed before the emergence of Omicron.
^§ https://www.covid19treatmentguidelines.nih.gov/management/clinical-management/
^¶PCORnet is a national network-of-networks developed to conduct patient-centered outcomes research. The PCORnet infrastructure supports large-scale studies using its distributed data network. https://doi.org/10.1016/j.jclinepi.2020.09.036
**mAbs included in this study include bamlanivimab, bamlanivimab and etesevimab, casirivimab, and imdevimab, and unspecified monoclonal antibodies. Medications are prescribed or administered in the 14 days before or after the index event.
^††All patients include 78.8% outpatient, 10.9% inpatient, and 10.3% with no associated care setting for mAbs. Care setting was designated with the test.
^§§Care setting was classified as the highest care setting within 16 days of a positive test result but does not necessarily reflect the care setting in which medications were provided. Patients initially tested in the outpatient setting would be assigned to the inpatient setting if they were admitted within 16 days of receipt of a positive test result.
^¶¶A query is a single statistical SAS package that runs at sites to generate the data required. This study used a modular program that generated aggregate data at the site level and combined all results returned to the coordinating center, resulting in a single aggregate report on data across all responding sites.
***Forty-one sites include Duke University, Medical University of South Carolina, University of North Carolina, Vanderbilt University Medical Center, Wake Forest Baptist Health, Allina Health, Intermountain Healthcare, Medical College of Wisconsin, University of Iowa Healthcare, University of Kansas, University of Nebraska, University of Texas SW Medical Center, University of Utah, University Medical Center New Orleans, Children's Hospital Colorado, Children's Hospital of Philadelphia, Cincinnati Children's Hospital, Nationwide Children's Hospital, Nemours Children's Hospital, Seattle Children's Hospital, St. Louis Children's Hospital, Columbia, Montefiore, Mount Sinai Health System, New York University Langone Medical Center, Weill Cornell Medicine, Lurie Children's Hospital, Northwestern University, Fenway Health, Health Choice Network, OCHIN, Inc, Johns Hopkins University, Ohio State University, Penn State College of Medicine and Penn State Health Milton S. Hershey Medical Center, Temple University, University of Michigan, University of Pittsburgh Medical Center, AdventHealth, Orlando Health System, University of Florida Health, and University of Miami. These sites represent academic and community hospitals; are located across all 50 states, Washington, D.C., Puerto Rico, U.S. Virgin Islands, U.S. Armed forces, and Guam; serve patients who are self-pay, public or privately insured; and total 3.0% of COVID-19 cases (as compared with CDC case surveillance.)
^††† https://pubmed.ncbi.nlm.nih.gov/16032956/
^§§§45 C.F.R. part 46.102(l)(2), 21 C.F.R. part 56; 42 U.S.C. Sect. 241(d); 5 U.S.C. Sect. 552a; 44 U.S.C. Sect. 3501 et seq.
^¶¶¶Testing was by polymerase chain reaction or antigen test; a positive, detected, or presumptive positive result was considered to be a positive test.
****High-risk criteria defined by the FDA include age ≥65 years, obesity, pregnancy, chronic kidney disease, diabetes, immunosuppression, cardiovascular disease, and lung disease, along with other underlying conditions that are not explicitly listed in the EUAs for these treatments.
^††††Critical care services are identified by International Classification of Diseases, Tenth Revision critical care codes (99291 and 99292) for the evaluation and management of the critically ill or critically injured patient.

Table 1. Demographic and medical risk characteristics of patients with positive SARS-CoV-2 test results, by clinical setting and medications received — 41 health care systems in the National Patient-Centered Clinical Research Network, United States, March 2020–August 2021
Characteristic	No. (%)*
Characteristic	All patients with positive SARS-CoV-2 test result	Patients receiving monoclonal antibodies	Inpatients with positive SARS-CoV-2 test result	Patients receiving dexamethasone	Patients receiving remdesivir
No. of unique patients	805,276	12,539	120,204	40,685	35,315
Demographics
Age group, yrs
20–39	312,680 (38.8)	1,639 (13.1)	20,966 (17.4)	4,966 (12.2)	3,354 (9.5)
40–54	209,202 (26.0)	2,933 (23.4)	23,296 (19.4)	8,285 (20.4)	6,885 (19.5)
55–64	128,550 (16.0)	3,045 (24.3)	24,025 (20.0)	8,874 (21.8)	7,779 (22.0)
65–74	86,848 (10.8)	3,075 (24.5)	24,267 (20.2)	9,124 (22.4)	8,257 (23.4)
75–84	47,047 (5.8)	1,425 (11.4)	18,016 (15.0)	6,420 (15.8)	6,056 (17.1)
≥85	20,949 (2.6)	422 (3.4)	9,634 (8.0)	3,016 (7.4)	2,967 (8.4)
Sex
Female	437,651 (54.3)	6,709 (53.5)	59,583 (49.6)	19,262 (47.3)	16,607 (47.0)
Male	367,359 (45.6)	5,828 (46.5)	60,603 (50.4)	21,416 (52.6)	18,704 (53.0)
Other^†/Missing^§	264 (0.0)	3 (0.0)	17 (0.0)	8 (0.0)	3 (0.0)
Race
Asian	22,968 (2.9)	206 (1.6)	4,396 (3.7)	1,219 (3.0)	1,003 (2.8)
Black or African American	126,166 (15.7)	1,904 (15.2)	28,403 (23.6)	8,879 (21.8)	8,172 (23.1)
White	493,181 (61.2)	9,366 (74.7)	59,212 (49.3)	22,910 (56.3)	19,318 (54.7)
Other^¶	88,026 (10.9)	773 (6.2)	20,729 (17.2)	6,151 (15.1)	5,366 (15.2)
Missing^§	74,935 (9.3)	280 (2.2)	7,449 (6.2)	1,511 (3.7)	1,443 (4.1)
Ethnicity
Hispanic	149,565 (18.6)	1,006 (8.0)	25,953 (21.6)	7,557 (18.6)	6,895 (19.5)
Non-Hispanic	577,394 (71.7)	11,189 (89.2)	88,007 (73.2)	31,627 (77.7)	27,147 (76.9)
Other**	5,553 (0.7)	20 (0.2)	273 (0.2)	84 (0.2)	104 (0.3)
Missing^§	72,764 (9.0)	318 (2.5)	5,955 (5.0)	1,410 (3.5)	1,161 (3.3)
Medical conditions associated with high risk^††
Anemia	72,830 (9.0)	2,187 (17.4)	33,072 (27.5)	9,762 (24.0)	8,553 (24.2)
Arrythmia	73,318 (9.1)	2,527 (20.2)	39,255 (32.7)	12,235 (30.1)	10,828 (30.7)
Asthma	60,080 (7.5)	1,890 (15.1)	16,045 (13.3)	5,301 (13.0)	4,944 (14.0)
COPD	26,636 (3.3)	879 (7.0)	15,330 (12.8)	5,551 (13.6)	5,513 (15.6)
Cancer	37,027 (4.6)	1,641 (13.1)	12,869 (10.7)	4,716 (11.6)	3,605 (10.2)
Chronic kidney disease	50,580 (6.3)	1,795 (14.3)	30,206 (25.1)	9,269 (22.8)	8,418 (23.8)
Chronic pulmonary disorders	100,625 (12.5)	3,219 (25.7)	32,617 (27.1)	11,282 (27.7)	10,582 (30.0)
Coagulopathy	33,374 (4.1)	985 (7.9)	23,070 (19.2)	7,442 (18.3)	6,469 (18.3)
Congestive heart failure	40,179 (5.0)	1,344 (10.7)	24,627 (20.5)	7,868 (19.3)	7,329 (20.8)
Coronary artery disease	54,051 (6.7)	2,074 (16.5)	28,799 (24.0)	9,305 (22.9)	8,607 (24.4)
Diabetes type 2	107,527 (13.4)	3,890 (31.0)	47,963 (39.9)	15,462 (38.0)	14,706 (41.6)
Hypertension	209,848 (26.1)	7,265 (57.9)	78,700 (65.5)	25,653 (63.1)	23,633 (66.9)
Mental health disorders	97,046 (12.1)	2,728 (21.8)	26,443 (22.0)	8,015 (19.7)	7,044 (19.9)
Peripheral vascular disorders	31,930 (4.0)	1,250 (10.0)	16,496 (13.7)	5,373 (13.2)	4,596 (13.0)
Severe obesity (BMI ≥40 kg/m²)	60,052 (7.5)	2,430 (19.4)	20,271 (16.9)	7,781 (19.1)	6,891 (19.5)
Outcome^§§
Critical care	27,585 (3.4)	225 (1.8)	21,412 (17.8)	10,675 (26.2)	8,244 (23.3)

Abbreviations: BMI = body mass index; CDM = common data model; COPD = chronic obstructive pulmonary disease; PCORnet = National Patient-Centered Clinical Research Network.
*Percentages are simple summary numbers (column percentages) out of the total in each category. Strata are not expected to sum to the total because the small cell masking by the data partners before submission of data.
^†For sex stratifications, Other includes all remaining PCORnet CDM values that are not male or female.
^§For sex, race, and ethnicity stratifications, Missing includes PCORnet CDM values of Refuse to answer, No Information, Unknown, and missing values.
^¶For race stratifications, Other includes PCORnet CDM values of Native Hawaiian or Other Pacific Islander, American Indian or Alaska Native, Multiple races, and Other.
**For ethnicity stratifications, Other includes PCORnet CDM values of Other.
^††Recorded history of the diagnoses in electronic health record (outpatient or inpatient) within 3 years before a positive test. Patients can have more than one condition.
^§§Fourteen days before to 30 days after a positive test result.

Table 2. Average monthly frequency and relative disparity in receipt of medications for treatment of COVID-19, by race and ethnicity — 41 health care systems in the National Patient-Centered Clinical Research Network, United States, March 2020–August 2021
Treatment/Race and ethnicity	Total no. eligible for treatment*	Total no. (%) treated	Mean of monthly percentage treated^†	p_w ^†	Mean of monthly relative disparity,^§ % (95% CI)	p_t ^§
Monoclonal antibodies (November 2020–August 2021)
Race
White	334,472	9,366 (2.8)	4.0	—	Ref.	—
Black	73,853	1,904 (2.6)	2.8	0.004	−22.4 (−38.7 to −6.1)	0.0125
Asian	14,744	206 (1.4)	2.2	0.002	−48.3 (−63.1 to −33.6)	<0.0001
Other	45,521	773 (1.7)	2.2	0.002	−46.5 (−51.1 to −41.9)	<0.0001
Ethnicity
Non-Hispanic	577,394	11,189 (1.9)	4.0	—	Ref.	—
Hispanic	149,565	1,006 (0.7)	1.8	0.002	−57.7 (−66.6 to −48.9)	<0.0001
Dexamethasone (March 2020–August 2021)
Race
White	59,212	22,910 (38.7)	35.8	—	Ref.	—
Black	28,403	8,879 (31.3)	33.8	0.024	−1.9 (−7.8 to 3.9)	0.498
Asian	4,396	1,219 (27.7)	31.4	0.020	−2.0 (−17.3 to 13.2)	0.782
Other	20,729	6,151 (29.7)	34.2	0.106	−1.3 (−9.1 to 6.6)	0.735
Ethnicity
Non-Hispanic	88,007	31,627 (35.9)	35.4	—	Ref.	—
Hispanic	25,953	7,557 (29.1)	32.5	0.005	−6.2 (−11.7 to −0.6)	0.032
Remdesivir (March 2020–August 2021)
Race
White	59,212	19,318 (32.6)	29.0	—	Ref.	—
Black	28,403	8,172 (28.8)	31.2	0.028	9.3 (0.9 to 17.7)	0.032
Asian	4,396	1,003 (22.8)	26.2	0.200	−15.1 (−30.3 to 0.1)	0.052
Other	20,729	5,366 (25.9)	30.6	0.323	1.7 (−9.4 to 12.8)	0.748
Ethnicity
Non-Hispanic	88,007	27,147 (30.8)	29.3	—	Ref.	—
Hispanic	25,953	6,895 (26.6)	30.4	0.423	8.8 (−0.4 to 18.0)	0.060

Abbreviation: Ref. = referent group.
*For monoclonal antibody therapy, all patients with a positive SARS-CoV-2 test result were considered eligible for treatment. For dexamethasone and remdesivir, inpatients with a positive SARS-CoV-2 test result were considered eligible for treatment.
^†Mean of monthly treated time series tested for differences using pairwise Wilcoxon signed rank tests with p value given as p_w. Mean of monthly percent treated = [(n treated/n eligible)_{March 2020} + (n treated/n eligible)_{April 2020} + … (n treated/n eligible)_{August 2021}]/n total no. months.
^§The difference in percentage of patients treated among racial (Black, Asian, or Other races) or ethnic minority (Hispanic) and majority (White or non-Hispanic) groups divided by the percentage treated in the majority groups for each month. Assessed as nonzero using t tests with p-value given as p_t. Total number of months for dexamethasone and remdesivir = 18 and for monoclonal antibodies = 10. Mean of monthly relative disparity, % = [(Minority − majority/Majority)_March2020 + (minority − majority/Majority)_{April 2020} … + (Minority − majority/Majority)_{August 2021}]/Total no. of months.