Racial disparities in COVID-19 test positivity among people living with HIV in the United States.

OBJECTIVE: This study aimed to compare COVID-19 positivity by HIV status and race/ethnicity using data from the U.S. National COVID Cohort Collaborative (N3C).
METHODS: The N3C cohort (≥ 18 years) includes patients with any encounter after 1/1/2020 with SARS-CoV-2 laboratory tests. Detailed electronic medical records are centralized and harmonized across health-care organizations (34 sites). COVID-19 diagnosis was defined by RT-PCR or antibody testing. HIV infection was defined by standard diagnostic codes within 2 years prior to COVID-19 testing. Descriptive statistics and multivariable logistic regression were utilized to compare COVID-19 positivity and HIV status by patient's race/ethnicity.
RESULTS: Over 2.1 million patients were captured in the N3C as of 01/29/2021, of whom 372,716 (15%) were positive for COVID-19; 17,820 (0.7%) were PLWH of whom 2428 (13.6%) tested positive for COVID-19. COVID-19 positive PLWH were more likely to be 30+ years of age (90% vs. 70%; χ2p < 0.001), male (67% vs. 46%, χ2p < 0.001), and Black (44% vs. 15%, χ2p < 0.001) compared to HIV-negative patients. Compared to non-Hispanic/Latinx (NH)-White PLWH, NH-Black (aOR: 1.59, 95% CI: 1.37-1.86), Latinx (aOR: 2.17, 95% CI: 1.68-2.83), and NH-Asian (aOR: 2.18, 95% CI: 1.30-3.63) PLWH were more likely to have COVID-19 after adjustment for age, sex, and CharlsonDeyo comorbidity score.
CONCLUSION: PLWH and minoritized communities, including NH-Black and Latinx or Hispanic adults, appear to be disproportionately impacted by the COVID-19 pandemic.

Introduction

The advent of the global COVID-19 pandemic has brought to the forefront the significance of systemic barriers to equity in the United States (U.S) and the role of structural racism[1] in the spread of novel outbreaks leading to public health emergencies. In the US, the COVID-19 pandemic has disproportionately impacted racial and ethnic minority populations including Black, Indigenous, and Hispanic and Latinx peoples.[2] Minoritized communities in the US have experienced higher rates of incident infections, hospitalization, and mortality associated with COVID-19 compared to their White counterparts.[3-5] Racial/ethnic disparities observed during the COVID-19 pandemic are characterized by unequal exposure to social risk factors, including crowded living conditions, living in multigenerational households, frequent use of public transportation, and being an essential worker.[6,7] Unequal exposure to these risk factors among minoritized communities is due to US systemic barriers in equal access to financial institutions, housing markets, and employment opportunities.[3]

In addition to social conditions leading to unequal exposure, minoritized communities also have a disproportionate burden of preexisting chronic conditions or underlying comorbidities, such as HIV, that may increase their risk of adverse COVID-19 related outcomes. Recent hospital-based reports in the United States suggest that despite an observed lower overall SARS-CoV-2 seroprevalence among people living with HIV (PLWH) compared to their HIV-negative counterparts, PLWH are more likely to have severe COVID-19 outcomes.[8] An elevated risk of severe COVID-19 outcomes due to chronic immunosuppression is further exacerbated by existing racial disparities among PLWH in the US. For decades, non-Latinx/Hispanic (NH)-Black adults have had the highest burden of HIV-related morbidity and mortality; in 2018, although NH-Black adults constituted only 13% of the U.S. population, NH-Black persons accounted for 42% of new HIV diagnoses and 44% of HIV-related deaths.[9] Given the unequal burden of COVID-19 and HIV conditions among minoritized communities, it is important to quantify racial disparities in COVID-19 infection among PLWH. Data examining racial disparities in health-related outcomes provides rationale for future intervention-level work to alleviate the disproportionate burden of negative health outcomes. Therefore, our objective was to characterize racial and ethnic disparities in COVID-19 test positivity among people living with HIV (PLWH) in the US using a large multi-institutional database known as the National COVID Cohort Collaborative (N3C).

Methods

The US N3C is a centralized, harmonized, high-granularity electronic health record (EHR) repository stewarded by the US National Institutes of Health National Center for Advancing Translational Science (NIH/NCATS).[10] Designs, data collection, and data harmonization methods of the N3C phenotype 2 used in the present study have been described in prior publications.[10] For this analysis, we used data collected from adults ≥ 18 years captured in the N3C between 1 January 2020 and 29 January 2021, which included 34 institutional sites contributing data for over 2.1 million patients using the Observational Medical Outcomes Partnership Common Data Model. Each participating institutional site contributes demographics, medications, laboratory tests, diagnoses, and vital status to the central data repository for data harmonization. Deidentified data were shared through the NCATS Data Enclave for statistical analyses through institutional data use agreements. The N3C data Enclave is approved through the NIH Institutional Review Board (IRB) and the institutional IRB of each study location with an approved study protocol.

We leveraged the N3C COVID-19 positive definition, which is publicly available on GitHub (https://github.com/National-COVID-Cohort-Collaborative/Phenotype_Data_Acquisition). Briefly, positive cases were defined as patients with any encounter on or after 1/1/2020 with either (1) a positive result for one of a set of a priori-defined SARS-CoV-2 laboratory tests, or (2) a “strong positive” diagnostic code, or (3) two “weak positive” diagnostic codes on the same date or during the same encounter. Over 99% of the COVID-19 cases in N3C were confirmed by real-time PCR, and < 1% were confirmed by antibody testing.[10] HIV was defined by documentation of any of the following[1]: HIV condition (International Classification of Diseases or ICD-9 or 10 codes)[2], HIV-related laboratory results (e.g., HIV viral load), or[3] HIV-related medications excluding pre-exposure prophylaxis. Race/ethnicity (NH-White, NH-Black, Latinx, and N-Asian), and other covariates including, age, sex, and modified CharlsonDeyo comorbidity scores[11] were extracted from EHR during encounters prior to the first COVID-19 test for each patient. We compared COVID-19 test positivity by race/ethnicity across HIV and COVID-19 status using descriptive statistics and χ2-tests. We estimated adjusted odds ratio (aOR) with 95% confidence intervals (95% CI) to evaluate racial disparities in COVID-19 positivity among PLWH specifically, adjusting for sex, age, and CharlsonDeyo comorbidity scores.

Results

Overall, of the > 2.1 million patients included in the harmonized dataset, 372,716 (15%) patients were COVID-19 positive; 17,820 were PLWH (0.7%) of whom 2428 (13.6%) had COVID-19. Compared to HIV-negative patients with COVID-19, PLWH with COVID-19 were more likely to be older (p < 0.001), male (p < 0.001), and with greater comorbidities (p < 0.001) (Table1). We observed no significant differences in age, sex, and comorbidities when comparing PLWH with and without COVID-19 (Table 1).

Table 1.

Differences in COVID-19 positivity* by sociodemographic measures among people living with HIV (PLWH) and HIV-negative adults (n = 372,716), National COVID Cohort Collaborative (N3C) from 1 January 2020 to 29 January 2021.

	HIV Negative and COVID-19 positive (n = 370,288)		All People Living with HIV (n = 17,820)>		People Living with HIV and nonCOVID-19 positive (n = 15,392)		People Living with HIV and COVID-19 Positive (n = 2428)
	n	%	n	%	n	%	n	%
Age (years)	—	—	—	—	—	—	—	—
< 30	116,190	31.3	1957	10.9	1706	11.1	251	10.3
30–49	85,990	23.2	6724	37.7	5752	37.4	972	40
50–64	81,918	22.1	6606	37.1	5803	37.7	803	33.1
65+	81,179	21.9	2504	14.1	2117	13.8	387	15.9
Sex	—	—	—	—	—	—	—	—
Male	170,153	45.9	12,086	67.8	10,460	67.9	1626	66.9
Female	199,771	53.9	5730	32.2	4929	32.0	801	32.9
Modified CharlsonDeyo Comorbidity Indexa	—	—	—	—	—	—	—	—
0	271,226	73.2	10,073	56.5	8680	56.4	1393	57.4
1–2	62,115	16.8	3966	22.3	3414	22.2	552	22.7
3+	36,947	9.9	3781	21.2	3298	21.4	483	19.8

aCOVID-19 positive is based on PCR positivity, and negative is defined as any negative or no result

bExcluding HIV/AIDS from score

Compared to PLWH without COVID-19, PLWH with COVID-19 were more likely to be NH-Black (p < 0.001) and Latinx (p < 0.001), and conversely less likely to be NH-White (< 0.001; Figure 1). In multivariable analyses, compared to NH-White PLWH, NH-Black (aOR: 1.59, 95% CI: 1.37–1.86), Latinx or Hispanic (aOR: 2.17, 95% CI: 1.68–2.83), and NH-Asian (aOR: 2.18, 95% CI: 1.30–3.63) PLWH were more likely to have COVID-19 after adjustment for age, sex, and CharlsonDeyo comorbidity score.

Figure 1.

Racial/ethnic differences by COVID-19 positivity and HIV-status in the United States National COVID Cohort Collaborative (N3C) from 1 January 2020 to 29 January 2021 (n = 372,716).

Discussion

In this large EHR-based data resource, we observed significant racial/ethnic disparities among PLWH with COVID-19. NH-Black, Latinx or Hispanic, and Asian adults living with HIV were more likely to be COVID-19 positive compared to their NH-White counterparts even after adjustment for comorbidity score, age, and sex. It is important to underscore that these racial disparities observed are not due to biological race but systemic race.[12] As defined in Millet’s article, systemic or structural racism is defined as “A system in which public policies, institutional practices, cultural representations, and other norms work in various, often reinforcing ways to perpetuate racial group inequity.”[13] Through our analysis, we quantifiably demonstrate that the existing racial/ethnic disparities in the spread of HIV are further compounded by the COVID-19 epidemic, demonstrating the disproportionate burden of COVID-19 among minoritized populations in the US.

Inequities in both COVID-19 and HIV morbidity and mortality have been attributed to a synergy of social conditions that put socially disadvantaged groups at heightened risk of infection.[12,14] For example, data show that the most effective strategy to curbing the spread of the COVID-19 pandemic and reducing risk of exposure is through social distancing. However, as described by Yancy early in the pandemic,[15] the ability to maintain social distance from others while working from home or leveraging internet-based options for everyday activities are privileges not afforded by certain communities. Additionally, Millett has described several structural racism pathways that operate through social determinants of health leading to disparities in the burden of SARS-CoV-2 and HIV among minoritized communities. These pathways are multi-level factors including individual-level determinants such as income, homelessness, and unemployment status; health-care level determinants such as insurance/health-care access and provider bias; and environmental or community level-factors such as incarceration patterns, federal laws, and local policies.[12]

Future directions for this work include contextualizing the racial disparities we observed by evaluating social determinants of health as identified by Millett and their impact on racial/ethnic differences in HIV and COVID-19, both in the US and elsewhere. The current study is limited to data captured in the N3C dataset, which although are large, are not fully representative of the general or HIV-population in the US. Additionally, important individual-level factors such as sexual orientation or risk behaviors (e.g., injection drug use patterns) are not routinely captured in EMR and, therefore, are unavailable for consideration in these analyses. In the present analysis, we did not consider adherence to antiretroviral therapy to contextualize differences in uptake across racial/ethnic categories and potential associations with SARS-CoV-2 positivity. A recently published report by our group using N3C data demonstrated that among 1544 (11%) PLWH with available data, we observed that lower CD4 cell counts was indeed associated with adverse COVID-19 outcomes.[16] Although antioretroviral medication adherence was outside the scope of the current study, further analyses are ongoing to evaluate racial/ethnic differences in antioretroviral regimens and the potential role of antiretroviral therapy adherence on COVID-19 outcomes. Future research should prioritize profiling multi-level factors associated with HIV and COVID-19 outcomes, including the underlying structural or social determinants of health that may impact adherence to antiretroviral therapy or access to treatment. Nonetheless, our descriptive analyses provide the first insights into racial/ethnic disparities of the concurrent burden of HIV and COVID-19 in the US, highlighting the need to center work for HIV and/or COVID-19 on racial equity and justice.

National COVID Cohort Collaborative (N3C)

Roslyn MannonAlfred AnzaloneHarold P LehmannChristine SuverChristopher G. ChuteMelissa A. HaendelStephanie S. Hong