Social determinants of health and COVID-19: An evaluation of racial and ethnic disparities in attitudes, practices, and mental health

The study addresses how social determinants of health (SDOH) intersect with race and ethnicity to influence COVID-19-related attitudes, preventative practices, and mental health outcomes. The context is the disproportionate burden of COVID-19 cases, hospitalizations, and deaths among Black and Hispanic populations in the United States. Contributing factors include structural racism, lower insurance coverage, inconsistent management of comorbid conditions, chronic stress from discrimination, and overrepresentation in high-exposure occupations. Prior work has shown racial minorities may perceive higher risk yet engage in safer preventative practices. Arkansas, with historically poor health outcomes and pronounced racial disparities, provides a setting to evaluate how SDOH may differentially affect attitudes, practices, and mental health across racial/ethnic groups. The study’s purpose is to quantify associations between SDOH, race/ethnicity, and COVID-19 attitudes/practices and mental health symptoms using a statewide random digit dialing pulse poll.

Prior research indicates that Black, Hispanic, and Asian American individuals often perceive COVID-19 as a greater health threat than White individuals and report more financial and healthcare access worries. Some studies found Black and Hispanic respondents had lower knowledge of COVID-19 recommendations but engaged in safer practices (e.g., mask-wearing, avoiding travel/large gatherings). Other work noted mixed perceptions of risk among Black respondents. Broad literature identifies structural racism as a root cause of disparities in infection, hospitalization, and mortality, via inequities in housing, education, neighborhood resources, employment, insurance coverage, testing, diagnosis, and treatment access. Reports highlight unequal testing access and higher mortality in Black communities despite lower test volumes. Pandemic-era studies also show worse mental health outcomes among racial/ethnic minorities and essential workers. For Hispanic communities, additional factors such as immigration-related fears and lack of legal protections may influence mental health and access to services.

Design and data source: Cross-sectional analysis of data from the UAMS Fay W. Boozman College of Public Health Pandemic Pulse Poll, a random digit dialing (landlines and cell phones) survey conducted statewide in Arkansas from May–December 2020. For this study, analyses were limited to surveys completed September–December 2020 due to changes in key variables. Telephone numbers were purchased from a national vendor; cell numbers were targeted based on usage patterns and likely in-state use. Trained research assistants conducted standardized interviews in English or Spanish; responses were entered into REDCap. Eligibility and sample: Adults aged 18+ residing in Arkansas who provided verbal consent. Of 13,057 total respondents, 9,470 were from September–December 2020. Exclusions: respondents not identifying as White, Black, or Hispanic (n=196), and those with missing information (n=692), yielding a final analytic sample of 8,582. The study was deemed non-human subjects research by the UAMS IRB (IRB#262928). Variables: - Demographics: sex (male/female), age (<60, ≥60), race/ethnicity (White, Black, Hispanic), rurality (rural vs urban per RUCA codes). - Attitudes score (0–8, higher = safer attitudes): sum of eight binary-transformed items including: desire for COVID-19 testing; perceived importance of testing; recognition that shopping/eating in restaurants may be unsafe; non-optimism about the pandemic ending soon; disagreement with opening public schools; disagreement with allowing large gatherings; agreement that masks stop the spread; agreement that the state needs a mask order. - Practices score (0–3, higher = less risky practices): sum of three items: regularly wearing masks; not attending religious gatherings in last two weeks; not attending family/community events with ≥10 people in last two weeks. - SDOH score (0–4, higher = better SDOH): sum of four binary-coded items reflecting: not being worried about ability to see a doctor if sick; having enough money for food in last two weeks; not being worried about paying for food in last two weeks; not being worried about paying rent/mortgage in last two weeks. Coding accounted for directionality of responses. - Mental health outcomes (past 7 days): depression symptoms via PHQ-2 (little interest/pleasure; feeling down/depressed/hopeless) and anxiety symptoms via GAD-2 (nervous/anxious/on edge; inability to control worrying). Statistical analysis: Descriptive statistics used chi-square tests (categorical) and Wald tests (continuous). Multivariable linear regression assessed associations of sex, age, rurality, and race/ethnicity with SDOH scores, overall and stratified by race/ethnicity (controlling for sex, age, rurality). Linear regression modeled associations of SDOH with attitudes and practices. Logistic regression modeled associations of SDOH with odds of screening negative for anxiety and depression. Stratified analyses by race/ethnicity examined differential associations. Analyses were weighted by age, race, and sex to represent the Arkansas population. Software: STATA v15.

- Sample: N=8,582 (76.4% White, 18.2% Black, 5.5% Hispanic). Females 51.6%. Age <60: 70.2%. Rural: 55.9%. - Mean scores by race/ethnicity: • Attitudes (0–8, higher=safe): Overall 3.75 (SE 0.03); White 3.45; Black 4.90; Hispanic 4.26. • Practices (0–3, higher=less risky): Overall 2.46 (SE 0.01); White 2.41; Black 2.63; Hispanic 2.50. • SDOH (0–4, higher=better): Overall 3.57 (SE 0.01); White 3.65; Black 3.33; Hispanic 3.22. In adjusted models, Black (−0.30; 95% CI −0.36, −0.25) and Hispanic (−0.40; 95% CI −0.51, −0.28) respondents had lower SDOH scores than Whites. - Associations of SDOH with attitudes and practices (adjusted): For each 1-point increase in SDOH, attitude score decreased by 0.35 and practices score decreased by 0.09 (i.e., less safe attitudes/practices). By race: attitude decreases were significant among White (−0.39; 95% CI −0.49, −0.29), Black (−0.28; 95% CI −0.41, −0.16), and Hispanic (−0.27; 95% CI −0.53, −0.01) respondents. Practice decreases were significant among Whites (−0.12; 95% CI −0.14, −0.10) and Hispanics (−0.08; 95% CI −0.15, −0.00), but not Blacks (−0.02; 95% CI −0.05, 0.01). - Mental health prevalence: Screening negative for generalized anxiety disorder ranged ~82.5%–83.2% across races; negative for depressive symptoms ranged ~85.7%–87.5%; differences by race were not statistically significant. - Associations of SDOH with mental health (adjusted): Overall, each 1-point higher SDOH score was associated with higher odds of screening negative for anxiety (~76% increase) and depression (reported as ~66% increase in results text; abstract reports ~85% increase). Stratified analyses indicated increased odds among Whites and Blacks; improvements were less pronounced among Hispanics. - Overall pattern: Black and Hispanic respondents had lower SDOH scores yet demonstrated safer attitudes and practices than White respondents.

Findings indicate that Black and Hispanic Arkansans faced worse social determinants of health (lower SDOH scores) but reported safer COVID-19-related attitudes and practices than White respondents. Better SDOH was paradoxically associated with less safe attitudes and practices across groups, possibly reflecting that those with fewer resources adopt stricter protective behaviors due to higher perceived vulnerability and exposure risks. Improved SDOH correlated with lower odds of anxiety and depression symptoms overall; however, mental health disparities persisted, with Black respondents experiencing worse mental health outcomes relative to Whites regardless of SDOH, implicating structural factors beyond measured SDOH. The results align with prior literature documenting structural racism’s role in unequal exposure, healthcare access, and outcomes during COVID-19, including disparities in testing and treatment access. The study underscores the need for equity-focused public health strategies addressing SDOH and structural barriers, particularly in rural settings like Arkansas, to improve both behavioral responses to pandemics and mental health outcomes.

This study contributes evidence that racial and ethnic disparities in social determinants of health are linked to differences in COVID-19 attitudes, practices, and mental health. Black and Hispanic respondents exhibited safer pandemic-related behaviors despite worse SDOH, while better SDOH was associated with reduced anxiety and depression symptoms but less safe behaviors. Policymakers should prioritize comprehensive, culturally tailored interventions addressing SDOH—such as healthcare access, economic supports, housing and food security, and community trust-building—especially in rural states. Future research should include broader representation of underrepresented groups, refine SDOH measures to capture pandemic-specific and immigration-related factors, and assess temporal changes as policies, vaccines, and variants evolve.

- Limited representation of underrepresented racial/ethnic groups beyond White, Black, and Hispanic; no oversampling of smaller populations (e.g., Marshallese), restricting generalizability and subgroup analyses. - Relatively small sample for some subgroups and large standard errors in adjusted, race-stratified mental health analyses, limiting precision and comparisons. - Changes in survey questions beginning in September prevented assessment of temporal trends in attitudes, practices, and mental health over the broader study period. - SDOH score may not capture all relevant pandemic-specific social risks (e.g., immigration concerns, deportation fears, legal protections), potentially underestimating impacts in Hispanic communities. - Study timeframe ended before vaccine rollout, emergence of new variants in the U.S., and administrative changes, limiting applicability to subsequent pandemic phases. - Cross-sectional design precludes causal inference.