Disparities by Sex in COVID-19 Risk and Related Harms Among People with Opioid Use Disorder

Abstract

Objectives:

(1) Report sex-specific prevalence of coronavirus disease 2019 (COVID-19) test positivity among an opioid use disorder (OUD) cohort (2) Assess sex-specific rates of opioid overdose and mortality.

Methods:

A retrospective cohort study was performed on all adult patients with OUD who received a COVID-19 test in calendar year 2020 at a large academic medical center in Richmond, Virginia. Our study outcomes were positive COVID-19 test, opioid overdose, and all-cause in-hospital mortality. Sex-stratified multivariable logistic regression assessed sociodemographic factors associated with COVID-19 test positivity.

Results:

A total of 2,600 patients (males = 1,294, females = 1,306) with OUD received a COVID-19 test. Approximately 5% across both sexes tested positive for COVID-19 (p = 0.420), whereas 7% presented with an opioid overdose (males 10%; females 4%; p < 0.0001). However, mortality rates were similar across sex. Among males, individuals in the other racial group had increased odds of COVID-19 test positivity (adjusted odds ratio or AOR: 5.03, 95% confidence interval [CI]: 1.70–14.88), whereas black females had higher odds of COVID-19 test positivity (AOR: 1.92, 95% CI: 1.01–3.62) compared to their white counterparts.

Conclusions:

Opioid overdose, more often than COVID-19, impacted the health of patients with OUD presenting to a public safety net health system. Despite a female advantage documented in the general population for COVID-19 susceptibility, COVID-19 test positivity rates were similar across sex in an OUD cohort; yet, racial disparities emerged with notable sex-related variation. Sex and gender are important variables that modify health outcomes, including OUD and COVID-19, and should be further investigated using an intersectionality framework.

Introduction

The global pandemic of coronavirus disease 2019 (COVID-19) has caused numerous excess deaths in the United states¹ and has caused substantial medical,² social,³ and economic impacts.^4,5 In addition to the striking racial disparities unmasked in the COVID-19 global pandemic,^6,7 important differences by sex^8

–11 in COVID-19-related outcomes are also emerging in the literature. Data from the general population illustrate higher disease burden among males,¹² potentially related to sex-specific physiological¹³ and immunological mechanisms.¹⁴

However, both sex- (biological variable) and gender (continuum of sociocultural construct)-related factors, such as pregnancy, coexisting mood disorders, and the caregiver role, may intersect. This could place females at higher risk of COVID-19-related impacts,¹⁵ including those related to socioeconomic and emotional stress.^11,16,17

Given these important sex differences, evaluations of how COVID-19 impacts the health of individuals should prioritize utilizing sex-disaggregated data to investigate how intersecting identities impact outcomes; doing so ultimately can lead to proper elucidation of sex-related variations in epidemiologic trends and their underlying causes.¹⁸

Further, COVID-19 incidence and mortality are particularly high among vulnerable groups, including individuals with medical¹² and mental health comorbidities,¹⁹ particularly substance use disorders.²⁰ The COVID-19 pandemic has coincided with the overdose crisis, leading to devastating exacerbations in opioid use disorder (OUD)-related morbidity and mortality.²¹ In September 2020, more than 87,000 Americans died of drug overdoses in the United States, resulting in an increase of 29% compared to the previous 12-month period.^21,22

The state of Virginia also reported a substantial (40%) increase in opioid overdose deaths (2,013 in 2020 compared to 1,451 in 2019) during the same time period, largely driven by synthetic opioids such as fentanyl.²¹ Before the onset of the COVID-19 pandemic, black individuals had started outpacing white individuals in opioid overdose deaths,²³ and the historical gap between females and males for OUD-related mortality was narrowing.²⁴ However, how sex impacts COVID-19 risk among individuals with OUD is not well understood.

Therefore, to delineate disparities in need of urgent public health actions in the concurrent COVID-19 pandemic and overdose crises,²⁵ sex-informed investigations of COVID-19 outcomes among individuals with OUD are warranted. To address this gap, we conducted a retrospective study of OUD patients tested for COVID-19 from January through December 2020 at a public safety net health system in Richmond, Virginia.

Our primary objective was to assess sex-specific prevalence of COVID-19 and associated sociodemographic factors among individuals with OUD. Our secondary objective was to report sex-specific mortality and opioid overdose rates. Based on a prior study illustrating potential female disadvantage for COVID-19 susceptibility among individuals with OUD,²⁰ contrary to the general population,¹² we hypothesized that females and males would demonstrate similar COVID-19 prevalence, yet, differ in their identified risk factors for this outcome.

Methods

Study design, setting, and population

This retrospective, observational, cohort study included all consecutive patients tested for COVID-19 at Virginia Commonwealth University Health Systems (VCUHS) in Richmond, Virginia between January 1, 2020 and December 31, 2020. VCUHS is the largest public safety net health system in Virginia. The Institutional Review Board of Virginia Commonwealth University approved this study. All patients with OUD (inpatient and outpatient) who were at least 18 years old at the time of the COVID-19 test and whose sex and race were reported in their electronic medical records were included in the analyses.

Presence of OUD was determined by using OUD diagnosis codes from the International Classification of Diseases, 10th Revision, Clinical Modification [ICD-10-CM] (Supplementary Appendix A1) and/or presence of prescription of medication for OUD, including buprenorphine and naltrexone. Patients with an OUD medication prescription, but without an OUD diagnosis code, were excluded if they had a chronic pain disorder to optimize specificity of inclusion criteria. Data were abstracted from the electronic medical records by a VCU bioinformatics team, and a deidentified dataset was then provided to the study team.

Study variables

The primary outcome was a positive COVID-19 test documented in the electronic medical record. For patients with a positive COVID-19 test, we also reported on hospitalization, intensive care unit (ICU) admission, and mechanical ventilator use. Secondary outcomes (i.e., opioid overdose and mortality) were abstracted from the beginning of the study timeframe (January 1, 2020) through 60 days after the index COVID-19 test. The index test date was either the date of a positive test or the date of the most recent negative test (through December 31, 2020). Opioid overdose events were identified using ICD-10-CM codes (Supplementary Appendix A1) associated with health system encounters, and 60-day all-cause mortality²⁶ was identified by in-hospital deaths.

Additional variables abstracted included demographic characteristics (e.g., age, sex, race, smoking history), comorbidities, body mass index, smoking status, primary health insurance, symptomatic status at time of the COVID-19 test (classified as symptomatic or asymptomatic), and test location. Race was based on self-reported data recorded in the electronic medical record. Individuals were classified as black or white or “Other” race which included Native Hawaiian or Pacific Islander, Native American or Alaska Native, Asian, and individuals who were two or more races. Hispanic and Latino ethnicity was not assessed separately from race due to small sample size. The 5-year average (2015–2019) percentage of individuals living at or below 100% of the federal poverty level for each patient's zip code was obtained from the United States Census Bureau's American Community Survey.²⁷ Chronic conditions documented through ICD-10 codes were used to calculate the Charlson Comorbidity Index score.²⁸

Statistical analyses

First, sociodemographic factors were reported separately for males and females. Chi-squared analyses, Fisher's exact tests, or t-tests were utilized to examine bivariate associations. Then for the primary objective, sex-specific prevalence of having a COVID-19 positive test was reported and compared by sex using chi-squared tests. For the subset of patients testing positive, rates of hospitalization, ICU admission, and mechanical ventilation were also reported and compared by sex. Subsequently, sex-stratified multivariable logistic regression models were built for the primary outcome of COVID-19 test positivity. Variables included in the models were selected a priori based on existing literature.^6,7

For the secondary objective, sex-specific rates of 60-day all-cause mortality and opioid overdose were calculated and compared between males and females. For 60-day mortality, we also reported if the death occurred during the same encounter as a positive COVID-19 test or an opioid overdose. Opioid overdose event frequency was coded as “0” where there were no overdose encounters and >0 was the number of overdose encounters. Statistical significance was determined at an alpha level of 0.05. All statistical analyses were performed using SAS 9.4 (SAS Institute, Inc., Cary, NC).

Results

From January 1, 2020 to December 31, 2020, 2,600 patients (males = 1,294, females = 1,306) with OUD received a COVID-19 test and were included in our analysis. Overall, the majority were black (52%), less than 60 years old (69%), and insured by Medicaid (31%) or Medicare (31%). On average, 17% of individuals in patients' neighborhoods were living at or below 100% of the federal poverty level. More females (47%) than males (28%) were obese (p < 0.0001). More males (52%) than females (40%) were currently smoking cigarettes (Table 1). Almost all patients (90%) had been diagnosed with OUD within the last year or prior, and 26% had received buprenorphine or naltrexone treatment for OUD. The majority (64%) had an active OUD diagnosis within the past year with 18% receiving medication treatment during that time (data not shown).

Table 1.

Clinical and Sociodemographic Characteristics of Patients Receiving a COVID-19 Test from January 1, 2020 to December 31, 2020 (N = 2,600)

	Total sample (N = 2,600)	Males (n = 1,294)	Females (n = 1,306)	p
Mean age in years (SD)	50.94 (15.23)	51.90 (14.30)	49.99 (16.04)	0.0014
Racial group^a				0.4300
Black	1,349 (52.35)	689 (53.62)	660 (51.08)
White	1,143 (44.35)	554 (43.11)	589 (45.59)
Other	85 (3.30)	42 (3.27)	43 (3.33)
Insurance type				0.0616
Medicaid/uninsured/self-pay	812 (31.23)	432 (33.38)	380 (29.10)
Medicare	813 (31.27)	393 (30.37)	420 (32.16)
Private	975 (37.50)	469 (36.24)	506 (38.74)
BMI^b				<0.0001
Not obese	1,558 (62.34)	888 (71.96)	670 (52.96)
Obese	941 (37.66)	346 (28.04)	595 (47.04)
Neighborhood poverty rate (mean %, SD)^c	16.93 (8.50)	17.38 (8.63)	16.50 (8.36)	0.0089
Charlson comorbidity index score (mean, SD; range 1–29)^d	5.22 (4.91)	5.15 (4.93)	5.29 (4.89)	0.4888
0	502 (19.31)	254 (19.63)	248 (18.99)
1–2	582 (22.38)	294 (22.72)	288 (22.05)
3–4	287 (11.04)	144 (11.13)	143 (10.95)
5+	1,229 (47.27)	602 (46.52)	627 (48.01)
Tobacco smoking history^e				<0.0001
No use	675 (28.16)	251 (21.60)	424 (34.33)
Former use	631 (26.32)	308 (26.51)	323 (26.15)
Current use	1,091 (45.52)	603 (51.89)	488 (39.51)
Location of COVID-19 test				<0.0001
Inpatient	1,371 (52.73)	723 (55.87)	648 (49.62)
Outpatient	1,000 (38.46)	433 (33.46)	567 (43.42)
Emergency department	229 (8.81)	138 (10.66)	91 (6.97)
Symptomatic at time of COVID-19 test^f				0.0057
Symptomatic	507 (21.12)	277 (23.45)	230 (18.85)
Asymptomatic	1,894 (78.88)	904 (76.55)	990 (81.15)

Bolded p values indicate significance at an alpha level of 0.05.

Data available for n = 2,577.

Data available for n = 2,499. Obesity calculated from BMI in electronic medical record (BMI ≥30).

Data available for n = 2,539. Includes percentage of individuals living below the federal poverty line in the patient's neighborhood based on zip code.

Charlson comorbidity index calculated using Quan method.²⁸

Calculated for n = 2,397.

Calculated for n = 2,401.

BMI, body mass index; COVID-19, coronavirus disease 2019; SD, standard deviation.

Approximately 5% of males and females tested positive for COVID-19 (p = 0.420) with similar rates by sex of hospitalization, ICU admission, and mechanical ventilation (Table 2). More males (n = 124) than females (n = 50) had an opioid overdose event during the study timeframe (p < 0.0001) and died as a result (n = 4 males; n = 0 females). However, 60-day all-cause mortality rates were similar across sex (Table 2). Being symptomatic at the time of the COVID-19 test was associated with test positivity in the adjusted analyses across sex.

Table 2.

Sex-Specific COVID-19 Outcomes Among Patients Receiving a COVID-19 Test Between January 1, 2020 and December 31, 2020 (N = 2,600)

	Total sample (N = 2,600)	Males (n = 1,294)	Females (n = 1,306)	p
Positive COVID-19 test, n (%)	142 (5.46)	66 (5.10)	76 (5.82)	0.4199
Hospitalization, n (%)	70 (2.69)	38 (2.94)	32 (2.45)	0.4436
ICU admission, n (%)	27 (1.04)	16 (1.24)	11 (0.84)	0.3215
Mechanical ventilator use, n (%)	24 (0.92)	13 (1.00)	11 (0.84)	0.6651
Overdose event, n (%)^a	174 (6.69)	124 (9.58)	50 (3.83)	<0.0001
Overdose event frequency, n (%)^a				<0.0001
0	2,426 (93.31)	1,170 (90.42)	1,256 (96.17)
1	143 (5.50)	103 (7.96)	40 (3.06)
2	28 (1.08)	20 (1.55)	8 (0.61)
3	3 (0.12)	1 (0.08)	2 (0.15)
60-day all cause, in-hospital mortality, n (%)^a	122 (4.69)	65 (5.02)	57 (4.36)	0.4271
COVID-19-specific mortality, n (%)	7 (0.27)	3 (0.23)	4 (0.31)	1.0000
Opioid overdose related, n (%)	4 (0.15)	4 (0.31)	0 (0.00)	0.0612

Bolded p values indicate significance at an alpha level of 0.05.

Overdose events and mortality outcomes reported from January 1, 2020 through 60 days after date of index COVID-19 test.

ICU, intensive care unit.

Among males, individuals in the Other racial group had higher odds of COVID-19 test positivity (adjusted odds ratio or AOR: 5.03, 95% confidence interval [CI]: 1.70–14.88), whereas black females were more likely to have a positive COVID-19 test (AOR: 1.92, 95% CI: 1.01–3.62) compared to their white counterparts. Among females only, current smoking status was associated with reduced odds of COVID-19 test positivity, after adjustment (AOR: 0.26, 95% CI: 0.12–0.57) (Table 3).

Table 3.

Factors Associated with COVID-19 Positive Test Results Among Patients Receiving a COVID-19 Test from January 1, 2020 to December 31, 2020 (N = 2,600)

	Males (n = 1,294)		Females (n = 1,306)
	OR (95% CI)	AOR (95% CI)	OR (95% CI)	AOR (95% CI)
Age in years^a	1.05 (0.96, 1.14)	1.04 (0.92, 1.17)	1.02 (0.94, 1.09)	1.02 (0.92, 1.13)
Racial group
White	Ref.	Ref.	Ref.	Ref.
Black	1.56 (0.90, 2.71)	1.08 (0.54, 2.13)	2.02 (1.22, 3.35)	1.92 (1.01, 3.62)
Other	6.28 (2.58, 15.30)	5.03 (1.70, 14.88)	1.20 (0.27, 5.27)	1.38 (0.29, 6.58)
Insurance type
Private	Ref.	Ref.	Ref.	Ref.
Medicaid/uninsured/self-pay	1.09 (0.62, 1.93)	1.19 (0.58, 2.46)	0.61 (0.33, 1.14)	0.57 (0.26, 1.26)
Medicare	0.75 (0.40, 1.43)	0.85 (0.39, 1.87)	1.10 (0.65, 1.85)	0.96 (0.49, 1.89)
Obese	1.92 (1.14, 3.24)	1.75 (0.95, 3.24)	1.62 (1.00, 2.62)	1.12 (0.64, 1.96)
Neighborhood poverty rate^b	1.01 (0.98, 1.04)	1.00 (0.96, 1.04)	1.01 (0.98, 1.04)	1.00 (0.96, 1.04)
Charlson comorbidity index score^c	1.00 (0.95, 1.05)	0.99 (0.92, 1.05)	1.01 (0.96, 1.05)	0.99 (0.94, 1.05)
Tobacco smoking history
No use	Ref.	Ref.	Ref.	Ref.
Former use	0.90 (0.47, 1.74)	0.75 (0.34, 1.65)	0.65 (0.37, 1.14)	0.78 (0.42, 1.45)
Current use	0.51 (0.27, 0.97)	0.51 (0.24, 1.07)	0.23 (0.12, 0.45)	0.26 (0.12, 0.57)
Location of COVID-19 test
Outpatient	Ref..	Ref.	Ref.	Ref.
Inpatient	1.59 (0.87, 2.92)	1.68 (0.76, 3.71)	1.17 (0.70, 1.93)	1.32 (0.72, 2.43)
Emergency department	2.65 (1.21, 5.82)	1.88 (0.66, 5.34)	2.65 (1.27, 5.53)	1.79 (0.72, 4.47)
Symptomatic at time of COVID-19 test	4.76 (2.78, 8.15)	4.28 (2.28, 8.03)	5.06 (3.10, 8.27)	4.58 (2.63, 7.97)

Bolded odds ratios and 95% CIs indicate significance at an alpha level of 0.05.

n = 1,002 for adjusted model for males; n = 1,110 for adjusted model for females.

Odds ratio demonstrates change in odds associated with 5-year increment increase in age after age 18.

Odds ratio demonstrates change in odds associated with 1 percentage point increase for population living below the federal poverty line within the patient's zip code.

Odds ratio demonstrates change in odds associated with 1 point increase in Charlson comorbidity index score (range 1–29).

AOR, adjusted odds ratio; CI, confidence interval.

Discussion

Among a majority black patient population with OUD receiving a COVID-19 test at a large public safety net health system in Richmond, Virginia, ∼5% of both males and females tested positive for COVID-19. Sixty-day all-cause mortality rates from the time of the COVID-19 test were similar across sex. However, racial differences emerged that varied by sex, with black females and males identifying as a race other than black or white, having higher odds of COVID-19 test positivity than their white counterparts. Overall, more patients, especially males, presented with an opioid overdose than a positive COVID-19 test. Findings highlight the importance of incorporating an intersectionality framework into health investigations, such as the ongoing COVID-19 pandemic and overdose crisis.^29,30

There is strong evidence of a male bias for COVID-19 outcomes, including COVID-19 test positivity^12,31 and COVID-related mortality, hospitalization, and ICU submissions.¹⁰ Simultaneously, individuals with certain health conditions, such as mental health diagnoses¹⁹ and substance use disorders,²⁰ also demonstrate higher risk of COVID-19 morbidity and mortality. However, how these comorbid conditions intersect with sex and gender-related factors has not been elucidated.³² In the current study, a male disadvantage was not found for COVID-19 test positivity, with males and females demonstrating similar rates. Reasons for these findings, which are contrary to those consistently found in the general population, are not known but likely are multifactorial in nature. For example, the risks associated with OUD itself for COVID-19 (e.g., lack of social distancing in drug using environments)³³ may be significant enough to outweigh the physiologic and immunological COVID-19 risks associated with male sex,^13,14 leading to similar rates for males and females.

Considering gender effects, women may be burdening more of the sociocultural impacts of the pandemic than men.^15,17 This gendered disadvantage is likely amplified for populations, like individuals with OUD, who at baseline face more negative social determinants of health¹¹ due to factors such as poverty, stigma, and discrimination. Thus, the overlapping intersections of these sex and gender informed biopsychosocial factors may be contributing to the similar rates of COVID-19 test positivity among this OUD patient population and warrant further investigation in larger samples.

Important sex-related variations in racial disparities for the outcome of COVID-19 test positivity emerged in the multivariable analysis. Specifically, black females and males identifying as a race other than white or black demonstrated higher odds of COVID-19 test positivity than their white counterparts. In the general population, as well as among individuals with substance use disorders, significant disparities for nonwhite individuals have emerged across studies for COVID-19 outcomes.^6,7,20 Our findings also reflect recently published data highlighting how disparities by sex for COVID-19 outcomes vary across racial groups.³⁴ Factors related to structural racism are likely major drivers of these racial disparities in the COVID-19 pandemic.³⁵

In the current study, by using sex-disaggregated data, further detail about how sex and race intersect were highlighted. Prior studies focused on racial disparities in the COVID-19 pandemic have largely focused on individual-level factors, which can perpetuate stereotypes that minority individuals are “to blame” for their negative outcomes.³⁵ Also, prior studies have not utilized an intersectionality approach to prioritize investigations of how sex, gender, and socioecological factors (such as those aligned with race) intersect to heighten risk or protection.³⁶ Future studies using sex-stratified study designs and data analyses to investigate how intersecting identities impact outcomes are urgently needed to inform effective, culturally tailored responses in the COVID-19 pandemic.³⁷

As an intersectionality approach is undertaken to address disparities by sex, gender, race, and socioeconomic status in the COVID-19 pandemic, multilevel frameworks that prioritize equal investigations of factors across different levels of influence (e.g., individual to society) as well as domains (e.g., biological to health system) should be incorporated into upcoming research.³⁸ A crucial component within these investigations is to prioritize how social determinants of health drive outcomes, and how these linkages differ by gender.³⁹ Due to the retrospective design using electronic medical record data, the current study was limited in its ability to assess these important variables, such as those indicative of structural racism.

However, patients presented with significant medical comorbidities and low socioeconomic status as demonstrated at both the individual (i.e., majority being uninsured or with public insurance) and community levels (i.e., substantial proportion of patients' neighborhoods comprising of people living below the federal poverty line). It is widely accepted that addressing social determinants of health may improve health outcomes and save lives more than focused medical interventions.⁴⁰ Translations of emerging findings from data in the COVID-19 pandemic into culturally informed public health actions are needed to ensure that proper attention is paid to social determinants of health through a gender-informed lens.

Notably, study results indicate that current smoking confers a protective effect for females regarding COVID-19 test positivity, even after adjustment. Although smoking has been associated with increased likelihood for COVID-19 disease severity, there have been conflicting reports in the literature about the role of current smoking in testing positive for COVID-19.^41

–45 Overall, recent systematic reviews and meta-analyses indicate that smoking is a risk factor for COVID-19.^44
–46 A special form of selection bias known as collider bias is the most plausible explanation for our study's apparent result.^47
–49 As Tattan-Birch et al.⁴⁸ explain, collider bias can occur when assessing smoking in a sample selected based on whether participants were tested for COVID-19.

Since coughing can result from both smoking and COVID-19, people who smoke may unnecessarily get tested for COVID-19 more often than people who do not smoke. This would lead to an overrepresentation of people who currently smoke in the sample and higher percentages of negative COVID-19 tests in people who smoke compared to people who do not smoke. Our sample selection criteria—all individuals tested for COVID-19 at VCUHS—and results are consistent with this explanation. Future studies could test COVID-19 positivity and/or seroprevalence in the community to reduce or eliminate collider bias.

Finally, more individuals with OUD presented with an opioid overdose than with a positive COVID-19 test during the study time frame. While COVID-19 is a significant public health crisis, the overdose epidemic continues.⁵⁰ In years leading up to the COVID-19 pandemic, the gap between males and females for morbidity and mortality due to substance use disorders was narrowing.²⁴ Simultaneously, black individuals started to outpace their white counterparts in opioid overdose mortality rates.²³ Study findings highlight how these trends by sex and race in the overdose crisis may be merging now in the COVID-19 pandemic, leading to significant disparities in morbidity and mortality at their intersection.²⁵ More work is needed to further detail these epidemiologic phenomena followed by swift data-driven actions at the provider, policy and population health levels with responses tailored to the needs of individuals with substance use disorders.⁵¹

Limitations

This study is not without limitations. First, data were abstracted from a single hospital system in Richmond, Virginia, and has limited external generalizability to other health care settings and other locations across the state and the country. Given the retrospective design and use of electronic medical record data originally collected for nonresearch purposes, information bias was possible, and available data were limited by provider documentation and patient report. For example, OUD is underdiagnosed and under documented due to multiple reasons, such as patient nondisclosure due to fear of discrimination and criminalizing drug policies as well as provider underuse of evidence-based, nonjudgmental screening practices. Thus, this differential identification of OUD patients in the medical record could have led to inflated outcome rates, especially for black patients in our study sample.

Also, gender identity was not reported, which excluded our ability to assess gender minorities among this highly vulnerable population. Selection bias, specifically collider bias, due to sample selection was likely, especially for females who smoke. In addition, the lack of a comparison group precluded our ability to compare COVID outcomes across OUD and non-OUD groups. Further, information about OUD treatment receipt was constrained by the electronic medical record abstraction (i.e., methadone receipt was not captured by medication prescription data), which precluded the ability to assess its impact on study outcomes. Finally, individuals of Hispanic/Latino ethnicity were included within the three racial groups for analyses due to small numbers in this sample. Despite these limitations, study findings nonetheless provide unique insights into the intersection between OUD and disparities by race and sex within the COVID-19 pandemic.

Conclusions

The COVID-19 pandemic has coincided with the overdose crisis. Sex may modify COVID-19 and OUD outcomes, yet, how sex- and gender-related variables impact COVID-19 risk among individuals with OUD is not well understood. In a cohort of patients with OUD receiving a COVID-19 test, COVID-19 susceptibility was similar across sex, contrary to findings from the general population indicating a female advantage for COVID-19 outcomes. Importantly, sex-related variations in COVID-19 susceptibility by race were elucidated, with black females and male individuals in the Other racial group demonstrating higher odds of having a positive COVID-19 test than their white counterparts.

Finally, opioid overdose was overall more prevalent within the sample during the study timeframe, especially for males, than a positive COVID-19 test. The interplay of structural, sociocultural, and biological variables, such as those related to gender and sex, modifies health outcomes, including OUD and COVID-19. Findings highlight the importance of incorporating an intersectionality framework into health investigations and the urgent need for effective, culturally tailored responses in the COVID-19 pandemic and ongoing overdose crisis.

Footnotes

Authors' Contributions

Dr. C.E.M. had full access to all the data in the study and takes responsibility for the integrity of the data. Concept and design: C.E.M. and B.T. Acquisition, analysis, or interpretation of data: C.E.M., D.D.H.T., and D.A.C. Initial Draft of the article: C.E.M. and B.T. Critical revision of the article for important intellectual content: All the authors. Statistical analysis: D.D.H.T. and D.A.C. Supervision: C.E.M. and D.A.C.

Author Disclosure Statement

No competing financial interests exist.

Funding Information

Dr. C.E.M. is supported by award No. K23 DA053507 from the National Institute on Drug Abuse and was partially supported by Award No. KL2 TR002648 from the National Center for Advancing Translational Sciences, National Institutes of Health. Dr. D.A.C has partially supported funding from grant UL1TR002649 from the National Center for Advancing Translational Sciences, National Institutes of Health.

Supplementary Material

Supplementary Appendix A1

References

Woolf

, Chapman

, Lee

. COVID-19 as the leading cause of death in the United States. JAMA, 2021; 325:123–124.

Birkmeyer

, Barnato

, Birkmeyer

, Bessler

, Skinner

. The impact of the COVID-19 pandemic on hospital admissions in the United States. Health Affairs, 2020; 39:2010–2017.

Vindegaard

, Benros

. COVID-19 pandemic and mental health consequences: Systematic review of the current evidence. Brain Behav Immunity, 2020; 89:531–542.

Kaye

, Okeagu

, Pham

, et al. Economic impact of COVID-19 pandemic on healthcare facilities and systems: International perspectives. Best Pract Res Clin Anaesthesiol, 2021; 35:293–306.

Polyakova

, Udalova

, Kocks

, Genadek

, Finlay

, Finkelstein

. Racial disparities in excess all-cause mortality during the early COVID-19 pandemic varied substantially across states. Health Affairs, 2021; 40:307–316.

Muñoz-Price

, Nattinger

, Rivera

, et al. Racial disparities in incidence and outcomes among patients with COVID-19. JAMA Netw Open, 2020; 3:e2021892.

Price-Haywood

, Burton

, Fort

, Seoane

. Hospitalization and mortality among Black patients and White patients with Covid-19. N Engl J Med, 2020; 382:2534–2543.

Tejpal

, Gianos

, Cerise

, et al. Sex-based differences in COVID-19 outcomes. J Womens Health (2002), 2021; 30:492–501.

Newman

. Mortality in COVID-19: Further evidence for a sex-based difference in the OpenSAFELY study. J Womens Health (2002), 2021; 30:61–63.

10.

Gomez

JMD

, Du-Fay-de-Lavallaz

, Fugar

, et al. Sex differences in COVID-19 hospitalization and mortality. J Womens Health (2002), 2021; 30:646–653.

11.

Gausman

, Langer

. Sex and gender disparities in the COVID-19 pandemic. J Womens Health (2002), 2020; 29:465–466.

12.

Abate

, Kassie

, Kassaw

, Aragie

, Masresha

. Sex difference in coronavirus disease (COVID-19): A systematic review and meta-analysis. BMJ Open, 2020; 10:e040129.

13.

Wray

, Arrowsmith

. The physiological mechanisms of the sex-based difference in outcomes of COVID19 infection. Front Physiol, 2021; 12:627260.

14.

Takahashi

, Ellingson

, Wong

, et al. Sex differences in immune responses that underlie COVID-19 disease outcomes. Nature, 2020; 588:315–320.

15.

Spagnolo

, Manson

, Joffe

. Sex and gender differences in health: What the COVID-19 pandemic can teach us. Ann Intern Med, 2020; 173:385–386.

16.

Burki

. The indirect impact of COVID-19 on women. Lancet Infect Dis, 2020; 20:904–905.

17.

Jacques-Aviñó C, López-Jiménez T, Medina-Perucha L, et al.. Gender-based approach on the social impact and mental health in Spain during COVID-19 lockdown: A cross-sectional study. BMJ Open, 2020; 10:e044617.

18.

UN. COVID-19: Emerging gender data and why it matters, 2020. Available at: https://data.unwomen.org/resources/covid-19-emerging-gender-data-and-why-it-matters Accessed March 11, 2021 .

19.

Wang

, Xu

, Volkow

. Increased risk of COVID-19 infection and mortality in people with mental disorders: Analysis from electronic health records in the United States. World Psychiatry, 2021; 20:124–130.

20.

Wang

, Kaelber

, Xu

, Volkow

. COVID-19 risk and outcomes in patients with substance use disorders: Analyses from electronic health records in the United States. Mol Psychiatry, 2021; 26:30–39.

21.

Ahmad

, Rossen

, Sutton

. Provisional Drug Overdose Death Counts. National Center for Health Statistics, Hyattsville, Maryland, U.S, 2021.

22.

Centers for Disease Control and Prevention USDoHaHS. Products—Vital Statistics Rapid Release—Provisional Drug Overdose Data, 2021. Available at: https://www.cdc.gov/nchs/nvss/vsrr/drug-overdose-data.htm. Accessed May 17, 2021 .

23.

Furr-Holden

, Milam

, Wang

, Sadler

. African Americans now outpace whites in opioid-involved overdose deaths: A comparison of temporal trends from 1999 to 2018. Addiction (Abingdon, England), 2021; 116:677–683.

24.

Woolf

, Schoomaker

. Life expectancy and mortality rates in the United States, 1959–2017. JAMA, 2019; 322:1996–2016.

25.

Enns

, Pinto

, Venugopal

, et al. Substance use and related harms in the context of COVID-19: A conceptual model. Health Promot Chronic Dis Prevent Canada, 2020; 40:342–349.

26.

WHO. A minimal common outcome measure set for COVID-19 clinical research. Lancet Infect Dis, 2020; 20:e192–e197.

27.

U.S. Census Bureau. American Community Survey 5-Year Estimates, Table S1701. In: Bureau USC, ed. American Community Survey, 2019.

28.

Quan

, Sundararajan

, Halfon

, et al. Coding algorithms for defining comorbidities in ICD-9-CM and ICD-10 administrative data. Med Care, 2005; 43:1130–1139.

29.

Crenshaw

. Mapping the margins: Intersectionality, identity politics, and violence against women of color. Stanford Law Revew, 1991; 43:1241–1299.

30.

Laster Pirtle

, Wright

. Structural gendered racism revealed in pandemic times: Intersectional approaches to understanding race and gender health inequities in COVID-19. Gender Soc, 2021; 35:168–179.

31.

Mohamed

, Gale

, Kontopantelis

, et al. Sex differences in mortality rates and underlying conditions for COVID-19 deaths in England and Wales. Mayo Clin Proc, 2020; 95:2110–2124.

32.

Haitao

, Vermunt

, Abeykoon

, et al. COVID-19 and sex differences: Mechanisms and biomarkers. Mayo Clin Proc, 2020; 95:2189–2203.

33.

Volkow

. COVID-19: Potential implications for individuals with substance use disorders, 2020. Available at: https://www.drugabuse.gov/about-nida/noras-blog/2020/04/covid-19-potential-implications-individuals-substance-use-disorders Accessed April 25, 2020 .

34.

Rushovich

, Boulicault

, Chen

, et al. Sex disparities in COVID-19 mortality vary across US racial groups. J Gen Intern Med, 2021; 36:1696–1701.

35.

Tan

, deSouza

, Raifman

. Structural racism and COVID-19 in the USA: A county-level empirical analysis. J Racial and Ethnic Health Disparities, 2022; 9:236–246.

36.

Poteat

. Navigating the storm: How to apply intersectionality to public health in times of crisis. Am J Public Health, 2021; 111:91–92.

37.

Scully

. Hidden in plain sight: Sex and gender in global pandemics. Curr Opin HIV AIDS, 2021; 16:48–53.

38.

NIMHD. NIMHD research framework, 2018. Available at: https://www.nimhd.nih.gov/about/overview/research-framework.html Accessed October 10, 2020 .

39.

Bowleg

. We're not all in this together: On COVID-19, intersectionality, and structural inequality. Am J Public Health, 2020; 110:917.

40.

Woolf

, Johnson

, Phillips

, Philipsen

. Giving everyone the health of the educated: An examination of whether social change would save more lives than medical advances. Am J Public Health, 2007; 97:679–683.

41.

Guo

. Smoking links to the severity of COVID-19: An update of a meta-analysis. J Med Virol, 2020; 92:2304–2305.

42.

Simons

, Shahab

, Brown

, Perski

. The association of smoking status with SARS-CoV-2 infection, hospitalization and mortality from COVID-19: A living rapid evidence review with Bayesian meta-analyses (version 7). Addiction (Abingdon, England), 2021; 116:1319–1368.

43.

, Narasimhan

, Sheehan

, Wu

, Fung

. Controversy over smoking in COVID-19-A real world experience in New York city. J Med Virol, 2021; 93:4537–4543.

44.

Zhao

, Meng

, Kumar

, et al. The impact of COPD and smoking history on the severity of COVID-19: A systemic review and meta-analysis. J Med Virol, 2020; 92:1915–1921.

45.

Reddy

, Charles

, Sklavounos

, Dutt

, Seed

, Khajuria

. The effect of smoking on COVID-19 severity: A systematic review and meta-analysis. J Med Virol, 2021; 93:1045–1056.

46.

Gulsen

, Yigitbas

, Uslu

, Dromann

, Kilinc

. The effect of smoking on COVID-19 symptom severity: Systematic review and meta-analysis. Pulm Med, 2020; 2020:7590207.

47.

Griffith

, Morris

, Tudball

, et al. Collider bias undermines our understanding of COVID-19 disease risk and severity. Nat Commun, 2020; 11:5749.

48.

Tattan-Birch

, Marsden

, West

, Gage

. Assessing and addressing collider bias in addiction research: The curious case of smoking and COVID-19. Addiction (Abingdon, England), 2021; 116:982–984.

49.

Greenland

. Quantifying biases in causal models: Classical confounding vs collider-stratification bias. Epidemiology (Cambridge, Mass), 2003; 14:300–306.

50.

Volkow

, Blanco

. Interventions to address the opioid crisis-modeling predictions and consequences of inaction. JAMA Netw Open, 2021; 4:e2037385.

51.

Evans

, Bufka

. The critical need for a population health approach: Addressing the nation's behavioral health during the COVID-19 pandemic and beyond. Prevent Chronic Dis, 2020; 17:E79.