Abstract
Introduction:
American Indian/Alaska Native (AI/AN) individuals were disproportionately affected by the COVID-19 pandemic. We aimed to evaluate whether the presence of chronic health conditions (CHC) impacted COVID-19 testing and vaccination for AI/AN people in Oklahoma.
Methods:
We pooled survey data that included adult AI/AN participants weighted using the 2022 American Community Survey. We used modified Poisson regression to estimate prevalence proportion ratios (PPR) and 95% confidence intervals (CI) accounting for weighting, multiple imputation, and confounders to determine whether access to COVID-19 testing and vaccine uptake differed by CHC status.
Results:
Among the 1,139 participants, 62.6% reported being diagnosed with a CHC. The majority reported being tested for COVID-19 (80.6%) and that test access was easy (81.7%). We observed no association between CHCs and being tested (Adjusted PPR: 1.08, 95% CI: 0.97, 1.19) or testing access (Adjusted PPR: 1.08, 95% CI: 0.95, 1.21). Fifty-seven percent of participants received a COVID-19 vaccine. Among unvaccinated individuals, 9.7% reported they were likely to get a vaccine. While we observed no association between CHCs and vaccine status (Adjusted PPR: 1.08, 95% CI: 0.90, 1.27), unvaccinated individuals with a CHC reported they were more likely to get a vaccine compared to individuals with no CHC (Adjusted PPR: 1.81, 95% CI: 1.01, 2.62).
Discussion:
The AI/AN population is at higher risk of developing CHC, which increases the risk for severe COVID-19. Thus, evaluating barriers to accessing testing and vaccines is important to improve outcomes. This analysis is important in supporting programs for seasonal epidemics and future pandemics.
Trial registration numbers:
Keywords: COVID-19 testing, COVID-19 vaccination, chronic health conditions, American Indians/Alaska Natives
Introduction
American Indian/Alaska Native (AI/AN) populations were disproportionately affected by the COVID-19 pandemic compared to other racial/ethnic populations, with age-adjusted COVID-19-related hospitalization rates consistently higher compared to non-Hispanic (NH) White populations (Cumulative incidence for 2024–2025 respiratory virus season: 113.3 per 100,000 for NH AI/AN and 96.1 per 100,000 for NH White), though rates may vary regionally [1, 2, 3, 4, 5, 6]. Furthermore, as of July 27, 2024 (most recent available), NH AI/AN people have the lowest percent ever vaccinated against COVID-19 (≥1 dose, 69.5%) compared to other race/ethnic groups (NH Asian: 94.5%, Hispanic: 84.6%, NH Black: 82.0%, NH White: 79.4%) and lower coverage of the 2024–2025 COVID-19 vaccine compared to NH Whites (NH AI/AN: 14.4%, NH White: 26.3%) [7, 8]. Based on the National Immunization Survey, approximately one-quarter (26.2%) of NH AI/AN respondents were concerned about COVID-19 disease, and approximately one-third (36.9%) had confidence in COVID-19 vaccine safety. However, just over half of NH AI/AN respondents (56.4%) had confidence that the COVID-19 vaccine is important to protect the respondent and only 29.3% of respondents reported that their healthcare provider recommended the vaccine [7].
A recent analysis by Kompaniyets et al. [9], including all race/ethnic groups, identified that the risk of admission to the intensive care unit (ICU), invasive mechanical ventilation, and death due to COVID-19 increased as the number of chronic health conditions (CHC) increased after accounting for age, sex, race/ethnicity, payer type, and hospital urbanicity in a hospital-based, all-payer database. Those with one condition had a 1.32 times higher risk of ICU admission, 1.57 times higher risk of invasive mechanical ventilation, and 1.53 times higher risk of death compared to those without CHCs, with risks increasing as the number of comorbidities increased [9]. Compared to other racial groups, AI/AN adults were 10.5 times more likely to be hospitalized for COVID-19 [5]. Nationally, over one-third (34%) of AI/AN adults aged 18–64 years are at risk of serious illness from COVID-19 due to comorbidities, including CHCs such as cancer, diabetes, chronic obstructive pulmonary disease, and asthma [10, 11, 12, 13, 14, 15, 16]. Even when accounting for comorbidities, NH AI/ANs have a higher odds of severe COVID-19 disease compared to NH Whites [17]. Increased risks included hospitalization (Odds Ratio [OR]: 1.21, 95% confidence interval [CI]: 1.03, 1.43), maximum length of stay (OR: 1.32, 95% CI: 1.16, 1.51), invasive ventilator dependence (OR: 3.49, 95% CI: 2.87, 4.25), and death (OR: 2.06, 95% CI: 1.70, 2.50), which were the highest ORs of all the race/ethnic groups evaluated (NH Asian/Pacific Islander, NH Black, Hispanic, and NH mixed or unknown race) [17].
Despite healthcare being a treaty-guaranteed right for AI/AN people enrolled in a tribal nation or Alaska Village at no personal financial cost, AI/AN populations continue to experience persistent disparities in access to health services [18]. This was problematic during the pandemic due to the overwhelming number of AI/AN people testing positive for COVID-19, which further strained tribal health, Indian Health Services (IHS), and urban Indian healthcare facilities trying to meet the typical healthcare needs of AI/AN populations along with managing the immediate and emergency care needed by COVID-19 patients. Further, AI/AN people report additional barriers to accessing healthcare, prior to and during the pandemic, including historical trauma, discrimination, provider mistrust, and cultural misunderstandings [19, 20].
Testing for COVID-19 early in the disease process minimizes severe outcomes by identifying the need for antiviral treatment and enhanced monitoring for the development of severe symptoms [21]. Being up-to-date on COVID-19 vaccines also reduces the risk of developing severe disease and death [21]. The presence of CHCs increases the risk of severe disease and AI/ANs are at higher risk for poor outcomes due to COVID-19, enhanced in part due to the higher prevalence of CHCs among this population. Thus, understanding whether there are differences in testing and vaccine uptake among AI/ANs with and without CHCs can help identify areas of improvement in encouraging testing and vaccine uptake among this high-risk population with unique healthcare access in order to inform healthcare practitioners working with AI/AN populations. Thus, we aimed to analyze data collected from cross-sectional surveys that included AI/AN participants from Oklahoma as part of the Rapid Acceleration of Diagnostics Underserved Populations (RADx-UP) consortium to better understand how CHCs impacted COVID-19 testing and vaccination uptake for AI/AN RADx-UP participants in Oklahoma. We used the Health Belief Model as a conceptual framework for this analysis, which has been used in recent COVID-19 studies to understand factors related to vaccine hesitancy [22, 23, 24].
Methods
Study Design and Data Sources.
The RADx-UP program aimed to improve access to testing for COVID-19 early in the pandemic, with a particular focus among underserved populations in the U.S. RADx-UP projects varied in their implementation of testing programs based on community partner input, but all projects required collection of common data elements (CDEs) to allow pooling of data for analyses [25, 26, 27, 28]. We pooled data from three cross-sectional surveys from RADx-UP programs in Oklahoma that collected data from self-reported AI/AN participants aged 18 years of age or older (n=1,139). The projects include: Cherokee PROTECT (Cherokee Nation Community-Driven Program for Testing and Contact Tracing, surveys collected from June 2021 – December 2022), CATCH-UP (Community-engaged Approaches to Testing in Community and Healthcare settings for Underserved Populations, surveys collected from February 2021 – May 2023), and CATCH-UP Vaccines (surveys collected from March 2022 – June 2023) (Figure 1). Participants from all three projects received an incentive to complete the survey ($15-$20 depending on the project). COVID-19 vaccines were available throughout the survey period, though high demand for vaccines in early 2021 may have limited access to some survey respondents.
Figure 1.
Map displaying the Cherokee Nation Reservation (Cherokee PROTECT) and number of survey respondents per county, locations of Project CATCH-UP and CATCH-UP Vaccines data collection sites.
The Cherokee PROTECT survey was a mailed survey, with options to return completed surveys by mail or through a REDCap link/QR code on the survey. Eligibility requirements for that survey included being 18 years and older, living in a household with at least one AI/AN individual, and residing in a county located within the Cherokee Nation Reservation. The respondent themselves did not have to be of AI/AN descent to participate in the survey. We used an addressed-based sampling method to randomly select a sample of addresses (n=10,000) from the 14 counties of the Cherokee Nation Reservation using address data from the Marketing Systems Group (MSG). In order to increase the probability of selecting an AI/AN household, only census tracts with the highest density of AI/AN population based on the 2019 American Community Survey estimates were included in the sampling frame. Our team, including Cherokee Nation staff, identified a cutpoint of greater than 38.7% AI/AN population based on using the natural breaks classification method in ArcGIS [29]. We received a total of 976 responses (10% response proportion) from multiple race/ethnic groups, which was restricted to AI/AN respondents (n=578) residing in four counties for this analysis.
Project CATCH-UP included surveys collected through COVID-19 testing events (including both viral and antibody testing) and staff working at participating primary care practices across Oklahoma using a convenience sample (n=3,242 total participants, n=534 AI/AN participants). CATCH-UP Vaccines was a supplemental project in partnership with Project CATCH-UP that included an intervention to address COVID-19 vaccine hesitancy and collected CDEs across Oklahoma (n=312 total participants, n=27 AI/AN participants).
Outcome Variables.
Our primary outcomes were COVID-19 testing and COVID-19 vaccine uptake. Testing access was measured through two questions 1) “Have you ever been tested for COVID-19?” (Response Options: yes, no) and 2) “It is easy to get tested for COVID-19” (Response options: easy [includes agree and strongly agree], not easy [includes neither disagree or agree, disagree, and strongly disagree]. We analyzed these questions as dichotomous variables. Vaccine uptake was measured through the questions 1) “Have you received a COVID-19 vaccine?” (Response options: yes, no) and 2) among those who reported not receiving a COVID-19 vaccine (hereafter described as “unvaccinated”, “How likely are you to get an approved COVID-19 vaccine when it becomes available?” (Response options: likely [includes fairly likely and very likely], not likely [includes not too likely, not at all likely, and definitely not]). We conducted analyses with dichotomous variables and the original ordinal outcome variables, as described below.
Primary Independent Variable.
Participants answered questions about current CHCs, including having an immunocompromising condition, autoimmune disease, hypertension, diabetes, chronic kidney disease, cancer diagnosis and/or treatment in the past 12 months, cardiovascular disease, asthma, chronic obstructive pulmonary disease, other chronic lung disease, sickle cell anemia, depression, alcohol or substance use disorder, intravenous drug use, other mental health disorder, or other CHC. We combined these individual variables into a single indicator of CHCs due to few participants reporting many of the listed individual chronic health conditions, with participants reporting any of the listed CHCs as having a CHC. Those without any of the listed CHCs were classified as not having a CHC. In addition, we calculated the number of CHCs for each participant (0, 1, 2, ≥3 conditions). Evaluating CHCs combined among this study population will be informative for healthcare practitioners in discussing COVID-related issues like testing and vaccine uptake with their high-risk patients.
Covariates.
Sociodemographic factors included age (continuous), gender (man or woman), education (<high school, high school graduate/GED, some college, Bachelor’s degree, other advanced degree), family annual income (<$15,000; $15,000-$19,999; $20,000-$24,999; $25,000-$34,999; $35,000-$49,999; $50,000-$74,999; $75,000-$99,999; ≥$100,000), employment (currently working, not working, retired), household composition (respondent only; living with spouse, no kids; family including kids; family with three generations), health insurance (no health insurance, private [purchased directly or through employment], public [Medicare, Medicaid, Tricare, IHS for Cherokee PROTECT]), self-reported health assessment (5-point Likert response options ranging from excellent to poor), presence of a disability (yes, no). Regarding healthcare access, participants answered questions on whether they experienced challenges in getting the healthcare they needed (including for mental health), getting the medicine they needed, and getting to where they needed to go during the pandemic (not a challenge, minor challenge, or major challenge). We combined this into a single variable of healthcare/transportation challenges (any challenges, no challenges). Behavioral factors included lifetime alcohol use (yes, no), current smoking status (current, not current smoker) and electronic cigarette (e-cigarette) status (current, not-current e-cigarette user).
Statistical Analysis.
Since the existing samples were obtained by using nonprobability sampling designs, which may introduce selection bias, we weighted the survey data with the 2022 American Community Survey estimates containing respondents who were AI/AN, Oklahoma residents, and ≥18 years of age to improve the representativeness and reduce selection bias [30, 31, 32]. Sociodemographic variables used in weighting include age, gender, education, and household income. Missing values in both the sample and ACS were imputed with a single imputation prior to weighting, which was necessary to ensure complete data for the statistical weighting process. Base weights were raked and trimmed using the iterative proportional fitting procedure to produce the final weights. To address missing data among variables not included in the weighting process, we used the sequential multiple imputation procedure to generate 50 imputations per participant for reducing the nonresponse bias in the analysis [33, 34].
For descriptive statistics, we calculated the weighted percent and 95% CI for each covariate using the MIANALYZE procedures in SAS to pool estimates. For multivariable associations of dichotomous dependent variables, we used modified Poisson regression accounting for statistical weighting and imputation to estimate the prevalence proportion ratio (PPR) with 95% CIs to compare factors associated with testing and vaccine uptake in separate models [35, 36]. The PPR and 95% CI for each imputed regression model was averaged into a pooled final estimate of the PPR and 95% CI [37]. To perform the hypothesis testing, first the pooled standard error (SE) was calculated by 1) calculating the variance within and between using SE’s from each imputed model, 2) summing the variance within and between across imputed models, and 3) taking the square root of the total variance.
To evaluate the ordinal variables of testing access and likelihood of vaccine uptake among those who were unvaccinated, we used ordinal logistic regression to estimate OR and 95% CI. The proportional odds assumption could not be evaluated in procedures that account for statistical weighting due to limitations of the statistical analysis programs, thus we evaluated models with and without consideration of statistical weighting. These analyses are exploratory since we could not fully evaluate the statistical assumptions. In the ordinal logistic regression model ignoring survey weighting, the proportional odds assumption was not violated (pooled p-value >0.05) [38, 39, 40, 41].
To evaluate potential confounding, we developed a directed acyclic graph to minimize the risk of overadjustment (adjusting for mediating factors), unnecessary adjustment (which may impact precision), and collider stratification bias (a form of selection bias) (Figure S1) [42, 43, 44]. For both dependent variables, we identified age (dichotomized at the weighted median value ≤40 and >40), gender, education (dichotomized as high school or less and at least a high school diploma), family annual income (dichotomized as <$35,000 and ≥$35,000), healthcare/transportation challenges, lifetime alcohol use, current smoking status, and current e-cigarette status as the minimally sufficient set for inclusion in multivariable models. Covariates were dichotomized due to model convergence issues. All analyses assumed an alpha of 0.05 and were performed with SAS (v9.4; SAS Institute, Cary, NC), R Software (v. 4.4.1; R Foundation for Statistical Computing, Vienna, Austria), and ArcGIS Pro (v. 3.4.0, ESRI Inc.). This study was reviewed and approved by the Institutional Review Boards (IRB) of the University of Oklahoma Health Sciences and Cherokee Nation following the principles of the Declaration of Helsinki. The manuscript was approved for dissemination by the Cherokee Nation IRB. Because this is a secondary data analysis, we received a waiver of informed consent from both IRBs for this study.
Results
Among the 1,139 participants, 62.6% reported being diagnosed with a CHC (Table 1). Those with CHCs were older (≥65 years: 21.1% vs. 5.7%), had lower income (<$15,000 per year: 13.0% vs. 6.5%), and were more frequently unemployed (38.5% vs. 11.1%) or retired (5.6% vs. 2.9%) compared to those without CHCs. Furthermore, those with CHCs were less frequently living in a household with children (46.2% vs. 61.8%), more commonly had public insurance (36.7% vs. 20.7%), more frequently reported having a disability (20.3% vs. 2.2%), and more frequently reported having challenges with accessing healthcare or transportation (36.2% vs. 18.3%). In addition, those with CHCs more commonly reported they were current smokers (19.3% vs. 10.3%) and current e-cigarette users (9.5% vs. 5.7%). Among those with CHCs, nearly half had one reported CHC (47.4%), followed by two CHCs (27.5%) and three or more CHCs (25.1%).
Table 1.
Distribution of survey participant characteristics stratified by the presence of chronic health conditions (CHC) among American Indian/Alaska Native respondents in Oklahoma.
| Any Chronic Health Condition (Pooled n=785, 62.6%) | No Chronic Health Conditions (Pooled n=354, 37.4%) | |||
|---|---|---|---|---|
| Pooled na | Weighted Percent (95% CI) | Pooled na | Weighted Percent (95% CI) | |
| Age | ||||
| 18–24 | 35 | 12.5 (7.7, 17.4) | 32 | 20.2 (11.8, 28.7) |
| 25–34 | 76 | 16.5 (11.9, 21.2) | 65 | 23.8 (16.1, 31.5) |
| 35–44 | 127 | 20.4 (15.6, 25.3) | 71 | 19.6 (13.2, 26.0) |
| 45–54 | 140 | 15.6 (11.7, 19.5) | 71 | 16.7 (8.4, 24.9) |
| 55–64 | 170 | 13.7 (10.3, 17.2) | 68 | 13.9 (8.5, 19.3) |
| ≥65 | 237 | 21.1 (16.8, 25.4) | 47 | 5.7 (2.5, 8.9) |
| Gender | ||||
| Man | 237 | 46.2 (40.4, 52.1) | 103 | 48.2 (39.1, 57.3) |
| Woman | 548 | 53.8 (47.9, 59.6) | 251 | 51.8 (42.7, 60.9) |
| Education | ||||
| Less than High School Diploma | 55 | 12.5 (8.7, 16.3) | 15 | 13.4 (4.9, 21.9) |
| High school graduate or GED completed | 171 | 38.9 (32.9, 44.9) | 65 | 36.5 (27.3, 45.7) |
| Some college level/Technical/Vocational degree | 291 | 31.5 (26.2, 36.8) | 138 | 30.1 (22.5, 37.7) |
| Bachelor’s degree | 154 | 10.7 (7.7, 13.6) | 80 | 13.2 (8.6, 17.8) |
| Other advanced degree (Masters, Doctoral degree) | 114 | 6.5 (4.2, 8.8) | 56 | 6.7 (4.0, 9.4) |
| Family Annual Income | ||||
| Less than $15,000 | 134 | 13.0 (8.7, 17.4) | 37 | 6.5 (1.9, 11.1) |
| $15,000 – $19,999 | 67 | 4.0 (1.2, 6.9) | 19 | 2.9 (0.0, 7.0) |
| $20,000 – $24,999 | 61 | 7.2 (3.7, 10.7) | 25 | 6.7 (1.5, 12.0) |
| $25,000 – $34,999 | 77 | 9.7 (5.9, 13.5) | 37 | 8.5 (3.1, 13.9) |
| $35,000 – $49,999 | 107 | 12.3 (8.3, 16.4) | 48 | 10.0 (4.2, 15.9) |
| $50,000 – $74,999 | 146 | 17.6 (13.0, 22.2) | 77 | 22.1 (14.6, 29.5) |
| $75,000 – $99,999 | 89 | 12.7 (8.7, 16.7) | 43 | 12.9 (6.5, 19.4) |
| $100,000 and above | 104 | 23.4 (18.0, 28.7) | 68 | 30.4 (21.0, 39.8) |
| Employment | ||||
| Currently working | 405 | 55.9 (50.0, 61.8) | 273 | 86.0 (81.1, 90.9) |
| Not currently working | 360 | 38.5 (32.8, 44.2) | 72 | 11.1 (6.8, 15.4) |
| Retired | 20 | 5.6 (2.1, 9.1) | 9 | 2.9 (0.4, 5.5) |
| Household Composition | ||||
| Respondent only | 174 | 12.5 (8.8, 16.2) | 61 | 9.4 (5.2, 13.6) |
| Living with spouse, no kids | 257 | 32.1 (26.5, 37.6) | 88 | 20.8 (14.1, 27.4) |
| Family including kids | 287 | 46.2 (40.1, 52.2) | 173 | 61.8 (53.4, 70.3) |
| Family with 3 generations (parents, children, grandchildren) | 67 | 9.2 (5.6, 12.9) | 32 | 8.0 (3.6, 12.4) |
| Health Insurance | ||||
| No health insurance | 97 | 15.5 (10.9, 20.0) | 70 | 22.4 (14.2, 30.7) |
| Private (Purchased Directly or Through Employment) | 349 | 47.8 (41.9, 53.8) | 193 | 56.9 (47.6, 66.1) |
| Public (Medicare, Medicaid, Tricare) | 339 | 36.7 (31.0, 42.4) | 91 | 20.7 (13.9, 27.4) |
| Disability | ||||
| No | 616 | 79.7 (74.8, 84.7) | 343 | 97.8 (95.2, 100.3) |
| Yes | 169 | 20.3 (15.3, 25.2) | 11 | 2.2 (0.0, 4.8) |
| Healthcare/Transportation Challenges | ||||
| No | 501 | 63.8 (58.2, 69.5) | 261 | 81.7 (75.6, 87.9) |
| Yes | 284 | 36.2 (30.5, 41.8) | 93 | 18.3 (12.1, 24.4) |
| Lifetime Alcohol Use | ||||
| No | 299 | 28.9 (23.3, 34.5) | 92 | 27.1 (18.5, 35.7) |
| Yes | 486 | 71.1 (65.5, 76.7) | 262 | 72.9 (64.3, 81.5) |
| Smoking Status | ||||
| Not Current Smoker | 618 | 80.7 (75.5, 85.9) | 311 | 89.7 (84.3, 95.1) |
| Current Smoker | 167 | 19.3 (14.1, 24.5) | 43 | 10.3 (4.9, 15.7) |
| Vaping Status | ||||
| Not Current E-cigarette User | 728 | 90.5 (86.6, 94.5) | 326 | 94.3 (90.0, 98.6) |
| Current E-cigarette User | 57 | 9.5 (5.5, 13.4) | 28 | 5.7 (1.4, 10.0) |
| Number of CHCs per Participant | ||||
| 0 | - | - | 354 | 100.0 |
| 1 | 332 | 47.4 (41.5, 53.3) | - | - |
| 2 | 215 | 27.5 (22.4, 32.7) | - | - |
| ≥3 | 238 | 25.1 (20.0, 30.2) | - | - |
Counts are raw pooled counts with weighted percentages
Most respondents reported that they had been tested for COVID-19 (80.6%) and that accessing COVID-19 testing was easy (81.7%). Although a higher percentage of those with CHCs reported being tested for COVID-19 compared to those without CHCs, this difference was not statistically significant in the multivariable models (Adjusted PPR: 1.08, 95% CI: 0.97, 1.19) (Table 2). We observed similar results when evaluating ease of testing access by CHC status (Adjusted PPR: 1.08, 95% CI: 0.95, 1.21). In our exploratory analyses with testing access as an ordinal variable, we observed that those with CHCs had a higher odds of reporting access to testing was easy compared to those without CHCs, which was statistically significant when accounting for statistical weighting (Adjusted OR: 1.68, 95% CI: 1.09, 2.59) (Table S1).
Table 2.
Prevalence proportion ratios (PPR) and 95% confidence intervals (CI) evaluating the relationship with the presence of chronic health conditions (CHC) and COVID-19 testing and vaccination among American Indian/Alaska Native survey respondents in Oklahoma.
| Variable | CHC Present N (W%)a |
CHC Absent N (W%)a |
Unadjusted PPR (95% CI) |
Adjustedb PPR (95% CI) |
|---|---|---|---|---|
| Ever Tested for COVID-19 | ||||
| Yes | 653 (82.7) | 281 (77.2) | 1.07 (0.96, 1.18) | 1.08 (0.97, 1.19) |
| No | 132 (17.3) | 73 (22.8) | Referent | Referent |
| Ease of Testing Access | ||||
| Easy | 650 (83.5) | 293 (78.6) | 1.06 (0.94, 1.19) | 1.08 (0.95, 1.21) |
| Not Easy | 135 (16.5) | 61 (21.4) | Referent | Referent |
| Vaccine Uptake | ||||
| Yes | 565 (60.9) | 200 (52.4) | 1.16 (0.96, 1.37) | 1.08 (0.90, 1.27) |
| No | 220 (39.1) | 154 (47.6) | Referent | Referent |
| Likelihood of Getting a COVID-19 Vaccinec | ||||
| Likely | 54 (28.2) | 31 (15.5) | 1.85 (1.01, 2.69) | 1.81 (1.01, 2.62) |
| Not Likely | 167 (71.8) | 123 (84.5) | Referent | Referent |
Counts are raw pooled counts with weighted percentages (W%)
Based on DAG: age, gender, education, family income, healthcare/transportation challenges, lifetime alcohol use, current smoking status, and current e-cigarette status
Among respondents with no previous vaccine doses (n=221 with CHC, n=154 without CHC)
Among the respondents, 57.7% of participants reported having received a COVID-19 vaccine. Of those who had not received at least one dose of a COVID-19 vaccine (n=375), 9.7% reported they were likely to get a vaccine. We did not observe a statistically significant difference in ever getting a COVID-19 vaccine by CHC status, though those with CHCs more frequently received a vaccine (Adjusted PPR: 1.08, 95% CI: 0.90, 1.27). However, those who were unvaccinated with CHCs were nearly two times more likely to report willingness to get a COVID-19 vaccine in the future (Adjusted PPR: 1.81, 95% CI: 1.01, 2.62) than those without CHCs. While models evaluating willingness to get a COVID-19 vaccine using ordinal logistic regression accounting for statistical weighting were also increased for those with a CHC, the results were not statistically significant (Adjusted OR: 1.38, 95% CI: 0.72, 2.66).
Discussion
Our analyses did not identify differences by CHC status for having ever been tested for COVID-19 or ease of accessing COVID-19 testing. Few studies have evaluated barriers to COVID-19 testing among AI/AN populations with none evaluating differences in those with CHCs or other comorbidities. One study among tribal members living in Great Plains Reservations, found three main factors that influenced the receipt of a COVID-19 test: an ability to isolate if positive for COVID-19, being employed full-time, and of knowing the importance of testing even when asymptomatic [45]. A qualitative study including participants from the Flathead Reservation in Montana identified both facilitators and barriers to testing for COVID-19 [46]. Facilitators included social factors such as role modeling, acknowledgment of testing early in the disease process to improve outcomes, and the impact of testing in reducing infections and allowing community events to restart. Barriers included employment requirements to quarantine after testing positive, resulting in potential lost wages, inconsistent practices related to returning to work, and fear of losing employment. Distrust of the government and medical community based on injustices of AI/AN people, including the introduction of infectious diseases to the community, scientific misconduct, and poor quality of medical care were also identified barriers [46]. In another qualitative study of AI/AN people in the Southwest US, primary barriers to testing included confidentiality concerns, risk of contracting COVID-19 at testing facilities, and discomfort with the nasal swab [47]. Facilitators identified in this study included easy access to testing facilities (including mobile clinics), work and school related testing requirements, and protection for family/friends [47].
After the initial shortage of viral testing supplies, many labs, pharmacies, health departments, and tribes, including Cherokee Nation, offered drive-thru testing. Cherokee Nation also offered COVID-19 testing to non-AI/ANs residing within the Cherokee Nation Reservation throughout the pandemic and continues to do so in the interest of community well-being. At-home COVID-19 testing became more readily available as the pandemic progressed, including the availability of free testing kits from the U.S. federal government [48]. This open access to testing across Oklahoma and the Cherokee Nation Reservation reduced barriers to testing, thus likely reduced testing disparities among survey respondents.
We observed no association between CHC status and ever having received a COVID-19 vaccine. However, those with CHCs who were unvaccinated reported they were more likely to receive a vaccine in the future compared to those without CHCs. This suggests that those with CHCs may be more responsive to healthcare providers’ encouragement to receive a COVID-19 vaccine or stay up-to-date with seasonal COVID-19 vaccines. An alternate explanation is that those who were unvaccinated with CHCs have barriers to accessing vaccines due to transportation or other healthcare-related barriers. These participants may also have been more concerned about their health than those without CHCs. Future studies should explore these issues in more depth. In March 2021, Cherokee Nation opened up vaccine status to anyone regardless of AI/AN tribal membership and has continued to offer seasonal COVID-19 vaccine opportunities to the general public as of this writing [49]. The surveys included in this analysis were conducted after initial doses of COVID-19 vaccines were made available, from January 2021 through June 2023, thus reflect vaccine perceptions as vaccines were becoming available across the state.
Few studies have evaluated barriers and facilitators to obtaining a COVID-19 vaccine among AI/AN populations. Among AI/AN college students at the University of Alaska, Gonzalez et al. [50] identified that experiencing racism and healthcare system distrust resulted in increased vaccine hesitancy. Among AI/AN people in the Southwest US, barriers to obtaining a COVID-19 vaccine included concerns about adverse side effects of the vaccine, conspiracy theories, and distrust of the vaccine’s ingredients and effectiveness [47]. Facilitators to obtaining a vaccine included protection of themselves and friends/family, requirements for social activities, financial incentives, and community members having positive experiences with the vaccine [47]. In analyzing RADx-UP data from the Great Plains AI/AN populations, Purvis et al. [51] reported that motivating factors for receiving a COVID-19 vaccine included keeping their family safe (OR: 4.31, 95% CI: 2.29, 8.11) and themselves safe (OR: 2.47, 95% CI: 1.10, 5.51). While those with multiple CHCs had an increased frequency of receiving a vaccine, this was not statistically significant (OR: 2.92, 95% CI: 0.98, 8.68) but is consistent with our study results.
When considering the impact of CHCs on vaccine uptake, a survey in Australia identified high COVID-19 vaccine uptake among survey participants with serious health conditions, including cancer, diabetes mellitus, and multiple sclerosis (81.5% received at least one dose), which was similar to the general Australian population (79.9% in 2021) [52]. In a longitudinal study of the US population, Na et al. [53] reported an association between CHCs and time to obtain a COVID-19 vaccine from the baseline survey in December 2020, when vaccines were first available. As the number of CHCs increased, the time to receive a vaccine also increased (Time to first dose hazard ratio [HR]: 1.07, 95% CI: 1.03, 1.11; Time to second dose HR: 1.09, 95% CI: 1.06, 1.13). While our study indicated that participants who were unvaccinated with CHCs were more likely to receive a future COVID-19 vaccine, results are mixed [53, 54, 55]. Further evaluation of this association in the national RADx-UP data is the focus of future work. One unique aspect of Cherokee Nation’s vaccine distribution was prioritized access to COVID-19 vaccines to those at high risk for severe disease, which may have influenced our results. Furthermore, due to the unique nature of healthcare access for AI/AN individuals, barriers to obtain vaccines for those with CHCs may have been reduced compared to the non-AI/AN population.
One strength of our study included the availability of survey responses from AI/AN participants from across Oklahoma, including those residing in the Cherokee Nation Reservation. Oklahoma is home to 38 federally recognized tribes, thus representing a diversity of AI/AN populations. However, our study also had limitations, which included lack of information on IHS as a payer of healthcare services (though not a form of health insurance) from the RADx-UP data and the cross-sectional nature of the survey. Additional limitations were potential selection bias due to convenience sampling of participants and missing data. However, we minimized these limitations by applying statistical weighting of the survey responses to the Oklahoma population and use of multiple imputation to address missing data. We were limited in evaluating specific types of CHCs due to small samples sizes, thus, we combined all CHCs for this analysis. This may mask potential differences in associations for specific types of CHCs. Our primary analyses used regression models that required a binary dependent variable, though we also evaluated ordinal logistic regression models with similar results to our primary analyses. Further research is needed to allow for evaluating the proportional odds assumption with weighted data.
Although the emergency phase of the COVID-19 pandemic has ended, this analysis provides support for addressing testing and vaccination for future seasonal viral epidemics and pandemics through the domains of the Health Belief Model. Domains of this model include perceived susceptibility, perceived severity, perceived benefits, cues to action, and self-efficacy [22], which may all influence uptake of testing and vaccines. These domains can be incorporated into future programs that aim to reduce transmission of infectious diseases. Future programs should focus on early and open access to viral testing and vaccines to reduce disparities in viral transmission. Furthermore, early identification of infection is important to minimize the impacts of severe respiratory distress caused by COVID-19 and other viral infections. Identifying those at high risk of developing severe disease, including those with CHCs, and education on the benefits of vaccination will also help reduce disparities in severe illness and death from viral infections. This is particularly important to evaluate among AI/AN populations as they have COVID-19-related health and healthcare access disparities despite the unique nature of healthcare access opportunities [5, 17, 18].
Supplementary Material
Funding:
Research reported in this publication was supported by grant number R21MD019385 from the National Institute on Minority Health and Health Disparities and grant number S06GM127983-03S1 from the National Institute of General Medical Sciences. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.
Footnotes
Competing Interests: The authors have no relevant financial or non-financial interests to disclose.
Ethics Approval: This study was reviewed and approved by the Institutional Review Boards (IRB) of the University of Oklahoma and Cherokee Nation following the principles of the Declaration of Helsinki. The manuscript was approved for dissemination by the Cherokee Nation IRB. Because this is a secondary data analysis, we received a waiver of informed consent from both IRBs for this study.
References
- 1.Centers for Disease Control & Prevention. Respiratory Virus Hospitalization Surveillance Network (RESP-NET) 2024. [Available from: https://www.cdc.gov/resp-net/dashboard/.
- 2.Arrazola J, Masiello MM, Joshi S, Dominguez AE, Poel A, Wilkie CM, et al. COVID-19 Mortality Among American Indian and Alaska Native Persons - 14 States, January-June 2020. MMWR Morb Mortal Wkly Rep. 2020;69(49):1853–6. 10.15585/mmwr.mm6949a3 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Bassett MT, Chen JT, Krieger N. Variation in racial/ethnic disparities in COVID-19 mortality by age in the United States: A cross-sectional study. PLoS Med. 2020;17(10):e1003402. 10.1371/journal.pmed.1003402 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Hatcher SM, Agnew-Brune C, Anderson M, Zambrano LD, Rose CE, Jim MA, et al. COVID-19 Among American Indian and Alaska Native Persons - 23 States, January 31-July 3, 2020. MMWR Morb Mortal Wkly Rep. 2020;69(34):1166–9. 10.15585/mmwr.mm6934e1 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Hicks JT, Burnett E, Matanock A, Khalil G, English K, Doman B, et al. Hospitalizations for COVID-19 Among American Indian and Alaska Native Adults (≥ 18 Years Old) - New Mexico, March-September 2020. J Racial Ethn Health Disparities. 2023;10(1):56–63. 10.1007/s40615-021-01196-0 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Rodriguez-Lonebear D, Barceló NE, Akee R, Carroll SR. American Indian Reservations and COVID-19: Correlates of Early Infection Rates in the Pandemic. J Public Health Manag Pract. 2020;26(4):371–7. 10.1097/phh.0000000000001206 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Centers for Disease Control & Prevention. COVID-19 Vaccination Coverage and Vaccine Confidence Among Adults 2024. [Available from: https://www.cdc.gov/vaccines/imz-managers/coverage/covidvaxview/interactive/adults.html.
- 8.Centers for Disease Control & Prevention. COVID-19 Vaccination Coverage and Intent for Vaccination, Adults 18 Years and Older, United States 2025. [Available from: https://www.cdc.gov/covidvaxview/weekly-dashboard/adult-vaccination-coverage.html.
- 9.Kompaniyets L, Pennington AF, Goodman AB, Rosenblum HG, Belay B, Ko JY, et al. Underlying Medical Conditions and Severe Illness Among 540,667 Adults Hospitalized With COVID-19, March 2020-March 2021. Preventing Chronic Disease. 2021;18:E66. 10.5888/pcd18.210123 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Raifman MA, Raifman JR. Disparities in the Population at Risk of Severe Illness From COVID-19 by Race/Ethnicity and Income. Am J Prev Med. 2020;59(1):137–9. 10.1016/j.amepre.2020.04.003 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Artiga S, Neuman T, Claxton G, Rae M, Kates J, Michaud J. Low-Income and Communities of Color at Higher Risk of Serious Illness if Infected with Coronavirus. Kaiser Family Foundation; 2020. [Google Scholar]
- 12.Centers for Disease Control & Prevention. Estimates of diabetes and its burden in the United States. Atlanta, GA; 2017. [Google Scholar]
- 13.Centers for Disease Control & Prevention. Summary Health Statistics: National Health Interview Survey: 2018. Table A-1a. 2021. [Available from: http://www.cdc.gov/nchs/nhis/shs/tables.htm.
- 14.Melkonian SC, Weir HK, Jim MA, Preikschat B, Haverkamp D, White MC. Incidence of and Trends in the Leading Cancers With Elevated Incidence Among American Indian and Alaska Native Populations, 2012–2016. Am J Epidemiol. 2021;190(4):528–38. 10.1093/aje/kwaa222 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Oklahoma State Department of Health. Behavioral Risk Factor Surveillance System 2017, on Oklahoma Statistics on Health Available for Everyone (OK2SHARE) 2017. [Available from: http://www.health.ok.gov/ok2share
- 16.Wheaton AG, Cunningham TJ, Ford ES, Croft JB. Employment and activity limitations among adults with chronic obstructive pulmonary disease--United States, 2013. MMWR Morb Mortal Wkly Rep. 2015;64(11):289–95. https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4584881/pdf/289-295.pdf [PMC free article] [PubMed] [Google Scholar]
- 17.Qeadan F, VanSant-Webb E, Tingey B, Rogers TN, Brooks E, Mensah NA, et al. Racial disparities in COVID-19 outcomes exist despite comparable Elixhauser comorbidity indices between Blacks, Hispanics, Native Americans, and Whites. Sci Rep. 2021;11(1):8738. 10.1038/s41598-021-88308-2 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Kim G, Bryant AN, Goins RT, Worley CB, Chiriboga DA. Disparities in health status and health care access and use among older American Indians and Alaska Natives and non-Hispanic Whites in California. J Aging Health. 2012;24(5):799–811. 10.1177/0898264312444309 [DOI] [PubMed] [Google Scholar]
- 19.Call KT, McAlpine DD, Johnson PJ, Beebe TJ, McRae JA, Song Y. Barriers to care among American Indians in public health care programs. Med Care. 2006;44(6):595–600. 10.1097/01.mlr.0000215901.37144.94 [DOI] [PubMed] [Google Scholar]
- 20.Garcia AN, Venegas-Murrillo A, Martinez-Hollingsworth A, Smith LV, Wells K, Heilemann MV, et al. Patterns of Health Care Access and Use in an Urban American Indian and Alaska Native Population. J Racial Ethn Health Disparities. 2023:1–11. 10.1007/s40615-023-01624-3 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Centers for Disease Control & Prevention. Types of COVID-19 Treatment 2024. [Available from: https://www.cdc.gov/covid/treatment/index.html#:~:text=If%20you%20think%20you%20could,health%20center%20or%20health%20department.
- 22.Coe AB, Gatewood SB, Moczygemba LR, Goode JV, Beckner JO. The use of the health belief model to assess predictors of intent to receive the novel (2009) H1N1 influenza vaccine. Innov Pharm. 2012;3(2):1–11. 10.24926/iip.v3i2.257 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Mercadante AR, Law AV. Will they, or Won’t they? Examining patients’ vaccine intention for flu and COVID-19 using the Health Belief Model. Res Social Adm Pharm. 2020. 10.1016/j.sapharm.2020.12.012 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Wong LP, Alias H, Wong P-F, Lee HY, AbuBakar S. The use of the health belief model to assess predictors of intent to receive the COVID-19 vaccine and willingness to pay. Hum Vaccin Immunother. 2020;16(9):2204–14. 10.1080/21645515.2020.1790279 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.RADx-UP. NIH RADx-UP Common Data Elements 2024. [Available from: https://radx-up.org/research/cdes/.
- 26.Carrillo GA, Cohen-Wolkowiez M, D’Agostino EM, Marsolo K, Wruck LM, Johnson L, et al. Standardizing, harmonizing, and protecting data collection to broaden the impact of COVID-19 research: the rapid acceleration of diagnostics-underserved populations (RADx-UP) initiative. Journal of the American Medical Informatics Association : JAMIA. 2022;29(9):1480–8. 10.1093/jamia/ocac097 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Corbie G, D’Agostino EM, Knox S, Richmond A, Woods CW, Dave G, et al. RADx-UP Coordination and Data Collection: An Infrastructure for COVID-19 Testing Disparities Research. Am J Public Health. 2022;112(S9):S858–s63. 10.2105/ajph.2022.306953 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Narayanasamy S, Veldman TH, Lee MJ, Glover WA 2nd, Tillekeratne LG, Neighbors CE, et al. RADx-UP Testing Core: Access to COVID-19 Diagnostics in Community-Engaged Research with Underserved Populations. Journal of clinical microbiology. 2023;61(8):e0036723. 10.1128/jcm.00367-23 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.ESRI. Data classification methods N.D. [Available from: https://pro.arcgis.com/en/pro-app/latest/help/mapping/layer-properties/data-classification-methods.htm.
- 30.Valliant R. Comparing Alternatives for Estimation from Nonprobability Samples. Journal of Survey Statistics and Methodology. 2019;8(2):231–63. 10.1093/jssam/smz003 [DOI] [Google Scholar]
- 31.Chen JKT, Valliant R, Elliott MR, editors. Model-assisted calibration of non-probability sample survey data using adaptive LASSO 2018.
- 32.Chen S, Campbell J, Spain E, Woodruff A, Snider C. Improving the representativeness of the tribal behavioral risk factor surveillance system through data integration. BMC Public Health. 2023;23(1):273. 10.1186/s12889-023-15159-z [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33.Schafer JL, Olsen MK. Multiple Imputation for Multivariate Missing-Data Problems: A Data Analyst’s Perspective. Multivariate Behav Res. 1998;33(4):545–71. 10.1207/s15327906mbr3304_5 [DOI] [PubMed] [Google Scholar]
- 34.Zhu J, Raghunathan TE. Convergence Properties of a Sequential Regression Multiple Imputation Algorithm. Journal of the American Statistical Association. 2015;110(511):1112–24. http://www.jstor.org/stable/24739710 [Google Scholar]
- 35.Hale JJ, Thompson DM, Darden PM. Calculating subset weighted analysis using PROC SURVEYFREQ and GENMOD (Paper 272–2013). 2013.
- 36.Zou G. A modified poisson regression approach to prospective studies with binary data. American Journal of Epidemiology. 2003;159(7):702–6. 10.1093/aje/kwh090 [DOI] [PubMed] [Google Scholar]
- 37.Heymans MW, Eekhout I. Chapter 9 Rubin’s Rules. Applied Missing Data Analysis with SPSS and (R)Studio. Amsterdam: 2019. [Google Scholar]
- 38.Rubin DB. Multiple Imputation for Nonresponse in Surveys. New York: John Wiley & Sons, Inc; 1987. [Google Scholar]
- 39.Meng X-L, Rubin DB. Performing Likelihood Ratio Tests with Multiply-Imputed Data Sets. Biometrika. 1992;79(1):103–11. 10.2307/2337151 [DOI] [Google Scholar]
- 40.Barnard J, Rubin DB. Small-Sample Degrees of Freedom with Multiple Imputation. Biometrika. 1999;86(4):948–55. http://www.jstor.org/stable/2673599 [Google Scholar]
- 41.Brant R. Assessing Proportionality in the Proportional Odds Model for Ordinal Logistic Regression. Biometrics. 1990;46(4):1171–8. 10.2307/2532457 [DOI] [PubMed] [Google Scholar]
- 42.Greenland S, Pearl J, Robins JM. Causal diagrams for epidemiologic research. Epidemiology. 1999;10(1):37–48. http://journals.lww.com/epidem/Abstract/1999/01000/Causal_Diagrams_for_Epidemiologic_Research_.8.aspx [PubMed] [Google Scholar]
- 43.Schisterman EF, Cole SR, Platt RW. Overadjustment bias and unnecessary adjustment in epidemiologic studies. Epidemiology. 2009;20(4):488–95. http://webproxy.ouhsc.edu/login?url=http://ovidsp.ovid.com/ovidweb.cgi?T=JS&CSC=Y&NEWS=N&PAGE=fulltext&D=medl&AN=19525685 http://jg2fw3rn2d.search.serialssolutions.com/?url_ver=Z39.88-2004&rft_val_fmt=info:ofi/fmt:kev:mtx:journal&rfr_id=info:sid/Ovid:medl&rft.genre=article&rft_id=info:doi/&rft_id=info:pmid/19525685&rft.issn=1044-3983&rft.volume=20&rft.issue=4&rft.spage=488&rft.pages=488-95&rft.date=2009&rft.jtitle=Epidemiology&rft.atitle=Overadjustment+bias+and+unnecessary+adjustment+in+epidemiologic+studies.&rft.aulast=Schisterman [DOI] [PMC free article] [PubMed] [Google Scholar]
- 44.Greenland S. Quantifying biases in causal models: classical confounding vs collider-stratification bias. Epidemiology. 2003;14(3):300–6. [PubMed] [Google Scholar]
- 45.Purvis SJ, Soltoff A, Isaacson MJ, Duran T, Johnson G, LaPlante JR, et al. COVID-19 Testing Factors Among Great Plains American Indians. J Racial Ethn Health Disparities. 2022:1–12. 10.1007/s40615-022-01433-0 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 46.Van Rensburg D, Adams AK, Perez G, Bishop S, Warne T, Hassell L, et al. Factors influencing COVID-19 testing among Native Americans and Latinos in two rural agricultural communities: a qualitative study. Frontiers in Public Health. 2023;11. 10.3389/fpubh.2023.1220052 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 47.Oguntade HA, Pontes M, Pablo K, Pablo E, Goklish N, Stifter M, et al. Factors Influencing Behaviors to Reduce the Spread of COVID-19 Among Indigenous Families in the Southwest, United States. Int J Environ Res Public Health. 2024;21(11). 10.3390/ijerph21111407 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 48.The White House. FACT SHEET: Biden-Harris Administration Increases COVID-19 Testing in Schools to Keep Students Safe and Schools Open 2022. [Available from: https://www.whitehouse.gov/briefing-room/statements-releases/2022/01/12/fact-sheet-biden-harris-administration-increases-covid-19-testing-in-schools-to-keep-students-safe-and-schools-open/..
- 49.Polansky C. Cherokee Nation Opens Vaccine Eligibility To Non-Natives Living Within Reservation Boundaries [Available from: https://www.publicradiotulsa.org/local-regional/2021-03-10/cherokee-nation-opens-vaccine-eligibility-to-non-natives-living-within-reservation-boundaries.
- 50.Gonzalez VM, Stewart TJ. COVID-19 vaccine hesitancy among American Indian and Alaska native college students: the roles of discrimination, historical trauma, and healthcare system distrust. Journal of behavioral medicine. 2024;47(1):123–34. 10.1007/s10865-023-00443-5 [DOI] [PubMed] [Google Scholar]
- 51.Purvis SJ, Armstrong K, Isaacson MJ, Soltoff A, Duran T, Johnson G, et al. Factors Associated with COVID-19 Vaccination Uptake in Great Plains American Indian Communities. J Racial Ethn Health Disparities. 2024;11(6):3690–703. 10.1007/s40615-023-01818-9 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 52.Day D, Grech L, Nguyen M, Bain N, Kwok A, Harris S, et al. Serious Underlying Medical Conditions and COVID-19 Vaccine Hesitancy: A Large Cross-Sectional Analysis from Australia. Vaccines (Basel). 2022;10(6). 10.3390/vaccines10060851 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 53.Na L, Banks S, Wang PP. Racial and ethnic disparities in COVID-19 vaccine uptake: A mediation framework. Vaccine. 2023;41(14):2404–11. 10.1016/j.vaccine.2023.02.079 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 54.Moon I, Han J, Kim K. Determinants of COVID-19 vaccine Hesitancy: 2020 California Health Interview Survey. Prev Med Rep. 2023;33:102200. 10.1016/j.pmedr.2023.102200 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 55.Warren AM, Perrin PB, Elliott TR, Powers MB. Reasons for COVID-19 vaccine hesitancy in individuals with chronic health conditions. Health Science Reports. 2022;5(2):e485. 10.1002/hsr2.485 [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.