Factors Associated With Racial/Ethnic Group–Based Medical Mistrust and Perspectives on COVID-19 Vaccine Trial Participation and Vaccine Uptake in the US

IMPORTANCE The impact of COVID-19 in the US has been far-reaching and devastating, especially in Black populations. Vaccination is a critical part of controlling community spread, but vaccine acceptance has varied, with some research reporting that Black individuals in the US are less willing to be vaccinated than other racial/ethnic groups. Medical mistrust informed by experiences of racism may be associated with this lower willingness. OBJECTIVE To examine the association between race/ethnicity and rejection of COVID-19 vaccine trial participation and vaccine uptake and to investigate whether racial/ethnic group–based medical mistrust is a potential mediator of this association. DESIGN, SETTING, AND PARTICIPANTS This cross-sectional survey study was conducted from June to December 2020 using a convenience sample of 1835 adults aged 18 years or older residing in Michigan. Participants were recruited through community-based organizations and hospital-academic networks.


Introduction
In the US, there have been more than 29.8 million confirmed COVID-19 cases and more than 542 400 deaths from this disease. 1 Nationwide, racial and ethnic disparities have been substantial. Data show that Black individuals represent 22% of cases to date while only representing 13% of the US population. 2 Trends have been similar in Michigan, where COVID-19 incidence has been higher among Black individuals (48 443 per 1 000 000 population) compared with White individuals (45 427 per 1 000 000 population) and all other racial/ethnic groups. 3 COVID-19-related mortality has also been higher among Black individuals (2328 per 1 000 000 population) compared with White individuals (1329 per 1 000 000 population) and other racial/ethnic groups. Of note, Black individuals account for 14% of the population in Michigan but have represented 23% of COVID-19related deaths to date. 4 Vaccination against COVID-19 has been widely viewed as an essential component of a strategic plan to control community spread of COVID-19. However, public acceptance of a COVID-19 vaccine has varied. For example, in 1 nationally representative sample of 991 adults, 5 participants were asked, "When a vaccine for the coronavirus becomes available, will you get vaccinated?" In this sample, 10.8% reported "no" and 31.6% reported "not sure"; Black participants were significantly more likely to report "no" or "not sure" compared with White participants. Among those who reported no intention to be vaccinated, 32.5% cited lack of trust as a reason, including distrust of vaccines, government, pharmaceutical companies, and vaccine development or testing processes. The study did not explicitly examine lack of trust across racial and ethnic groups. However, there are compelling reasons why Black individuals in the US in particular would distrust a COVID-19 vaccine, and these reasons are rooted in racism. These include implicit bias within health care systems and among health care professionals that may be associated with lower quality of care and worse health outcomes among Black individuals in the US 6 and the rapid development and promotion of a COVID-19 vaccine within a sociopolitical climate that many Black individuals in the US perceive as hostile to them. 7,8 This survey study focused on examining the acceptability of a hypothetical COVID-19 vaccine.
The study also assessed acceptability of a hypothetical COVID-19 vaccine research trial. Lackland et al 9 noted that the participation of underrepresented racial/ethnic groups in COVID-19 trials is essential for external validity and translatability of results. Furthermore, the enrollment of members of these groups in trials may potentially increase acceptability of the vaccine among similar other individuals by helping to establish vaccination as an emerging group norm. In the current study, we hypothesized that (1) self-reported Black race would be associated with greater rejection of both COVID-19 vaccine trials and vaccine uptake and (2) racial/ethnic group-based medical mistrust would mediate the association between Black race and greater vaccine rejection. Illuminating these associations may assist in the understanding of the layered vaccine-related apprehensions of a population particularly burdened by COVID-19 and in the promotion of equitable access to an effective vaccine.

Participants and Procedures
In this survey study, participants included 1835 adult Michigan residents aged 18 years or older who responded to at least 1 of 2 survey items assessing agreement to participate in a COVID-19 vaccine trial and to receive a COVID-19 vaccine. A purposive sampling strategy was used that relied on survey dissemination through 9 community-based organizations (CBOs) throughout Michigan that were already formal community research partners with the study's lead institution. These CBOs included social service agencies targeting parents, older individuals, and LGBTQ (lesbian, gay, bisexual, transgender, queer) populations as well as rural communities, faith-based organizations, and federally qualified health centers. A link to an online survey was disseminated through these CBOs and through communications and marketing networks of hospitals affiliated with the lead academic institution. Participants also had the option of completing the survey via telephone interview.
Participants were enrolled from June to December 2020 and received $10 for participating. This study was reviewed and approved by Wayne State University's institutional review board (IRB), which granted a waiver of oral and signed informed consent because survey dissemination online could not practicably be carried out without this alteration. All participants were provided with an IRB-approved study information sheet before survey administration, representing an unsigned form of passive consent in which the participant indicated consent by proceeding with this minimal-risk task. The study followed the American Association for Public Opinion Research (AAPOR) reporting guideline for survey studies and met the expectations for the reporting of recruitment and participation outcomes, institutional research standards and ethics, and best practices.

Assessment
Participants' level of agreement with participation in a vaccine trial was assessed with 1 item adapted from the work of Jacobsen et al 10 : "If you were asked today to participate in a research study to test a COVID-19 vaccine, would you agree to participate?" A similar item was asked regarding vaccine uptake: "If you were offered a coronavirus vaccine that had been approved by the US FDA [US Food and Drug Administration] today, would you agree to be vaccinated?" For both items, responses were based on a Likert-type scale with available responses of "definitely yes" (1), "probably yes" (2), "neither yes or no" (3), "probably no" (4), or "definitely no" (5). Group-based medical mistrust was measured using the 6-item suspicion subscale of the Group-Based Medical Mistrust Scale, 11 a 12-item scale assessing suspicion of mainstream health care systems and professionals and of the treatment provided to individuals of the respondent's racial/ethnic group. Responses were based on a Likerttype scale ranging from 1 (strongly disagree) to 5 (strongly agree), with higher scores representing greater mistrust. Reliability in the current sample was high (Cronbach α = .94).
Sociodemographic variables were also assessed, including age, gender, income level, essential worker status, and race/ethnicity. Participants were asked to self-identify based on categories provided by study investigators: White; Black or African American; Arab, Chaldean, Middle Eastern, or North African (MENA); Hispanic; Asian; and multiracial or other groups.

Statistical Analysis
Sociodemographic characteristics, including medical mistrust, were summarized by count and percentage for categorical variables and mean (SD) for continuous variables. The association between race/ethnicity and medical mistrust was assessed by 1-way analysis of variance.
The study hypotheses addressed the associations between race/ethnicity, medical mistrust, and outcomes as outlined in our path model ( Figure). As an analytic approach, path analysis offered an efficient strategy to test the fit of this proposed model by examining a covariance structure consistent with hypothesized associations among model constructs. We used a separate path analysis for each outcome, fit with the lavaan package 12,13 using R, version 3.6.3 (R Project for Statistical Computing). 14 We specified paths from the racial/ethnic group and covariates to outcomes (ie, substantive associations and covariate controls) and paths from the racial/ethnic group to medical mistrust (ie, a direct association between race/ethnicity and the mediator). Consequently, the model included estimates of direct associations between racial/ethnic groups and mistrust, separately, on the outcomes, as well as indirect associations between racial/ethnic groups and outcomes that were mediated by mistrust. Products of relevant coefficients ( Figure) and associated bootstrap SEs indicated whether the indirect associations were statistically significant.

JAMA Network Open | Public Health
Given this study's cross-sectional convenience sample, the indirect associations cannot be interpreted as causal paths; rather, they should be interpreted as evidence that a model construct (eg, racial/ethnic group) may simultaneously share a significant association with a separate model construct (eg, mistrust) and the substantive model outcome (eg, rejection). 15 The outcomes were coded such that higher scores indicated greater rejection of vaccine trial participation and vaccine uptake. Age was entered as a continuous variable. Racial/ethnic categories were effect coded so that the model's coefficients represented a group's deviation from the adjusted overall mean estimate for the outcomes. The multiracial or other group was used as the effectcoded reference group (ie, the coefficient did not appear in the model) to examine deviations for the racial/ethnic groups that were more commonly identified and investigated. However, effect-coded coefficients indicated deviation from the overall mean estimate and not from the reference group. All other categorical variables (gender, educational level, income level, and essential worker status) were dummy coded such that the coefficients represented deviation from a reference group's mean estimate. Statistical significance was set at a 2-tailed P < .05. In response to the item asking about trial participation (n = 1827), response rates were as follows: "definitely yes," 8%; "probably yes," 17%; "neither yes or no" (unsure), 11%; "probably no,"

Results
27%; and "definitely no," 37%. Table 3 shows the proportion of the total sample and of each racial/ ethnic group that responded "definitely or probably no or unsure" vs "definitely or probably yes" to whether they would participate in a COVID-19 vaccine trial. A total of 1376 participants (75%) had responses consistent with greater rejection; differences were observed across race/ethnicity. In response to the item regarding vaccination (n = 1815), response rates were as follows: "definitely yes," 20%; "probably yes," 28%; "neither yes or no," 12%; "probably no," 19%; and "definitely no," 21%. Table 3 also shows the proportion of the total sample and of each racial/ethnic group that responded "definitely or probably no or unsure" vs "definitely or probably yes" to the question of whether they would receive a COVID-19 vaccine. Participants were divided approximately evenly between the 2 categories.

Discussion
Broad vaccine acceptance is generally regarded as critical to the long-term containment of COVID-19 in the US. The findings of the current study may be useful and formative for researchers, clinicians, and public health leaders to better understand the associations between racial/ethnic group-based medical mistrust and the willingness of diverse racial/ethnic groups to participate in COVID-19 vaccine trials and to accept a federally approved COVID-19 vaccine.
Of note, most participants in this study indicated low willingness or refusal to participate in a COVID-19 clinical trial to test the efficacy of a vaccine, with rejection highest among Black participants, followed by those who identified as MENA. Vaccine uptake rejection was lower, with half of participants indicating responses consistent with rejection and Black participants reporting the most refusal, followed by individuals who identified as MENA and Hispanic. These findings are consistent with other recent studies documenting widespread COVID-19 vaccine rejection in the US. 5,16 Although those studies showed differences across race and ethnicity, the current study is, to our knowledge, the first to investigate the role of racial/ethnic group-based medical mistrust in acceptance and rejection of COVID-19 vaccines. Specifically, medical mistrust partially mediated the association between Black race/ethnicity and refusal to participate in vaccine trials or receive a vaccine, suggesting that general suspicion of mainstream health care professionals and systems may be associated with this group's rejection of the vaccine.
A recent commentary by Warren and colleagues 17 offered insight relevant to these findings, citing the "deep and justified lack of trust" that many Black individuals in the US have of health care systems and clinical research. These authors state, "This distrust is often traced to the legacy of the infamous syphilis study at Tuskegee, in which investigators withheld treatment from hundreds of Black men in order to study the natural history of the disease. But the distrust is far more deeply rooted, in centuries of well-documented examples of racist exploitation by American physicians and researchers." Of importance, mistrust may be rooted in contemporary health care experiences. Such experiences were revealed in the results of a 2020 Kaiser Family Foundation survey of 1700 adults that included nearly 800 Black individuals in the US. 18 The findings showed that 45% of Black respondents reported at least 1 of 6 negative experiences with a health care professional (eg, the health care professional assumed something about them without asking, talked down to them or did not treat them with respect, or did not believe they were telling the truth), and 36% reported believing that they would have received better medical care if they had belonged to a different race/ ethnicity. Such findings support the notion that mistrust is associated with perceptions of past injustices as well as present-day experiences.
The current study also offers some insight into vaccine rejection among other racial/ethnic groups. The results showed that medical mistrust within racial/ethnic groups was associated with vaccine rejection in the overall sample regardless of race/ethnicity. Furthermore, although direct associations were not found between identifying as MENA or Hispanic and vaccine uptake rejection,

Limitations
This study has limitations. A probability sampling strategy was not used, and data were not collected from a random sample of the Michigan population. However, we attempted to compensate for this through the use of a maximum variation sampling strategy to approximate the potential responses of a population-based sample. Another limitation is that despite this approach, those who identified as male only represented 20% of the sample, reducing the generalizability of the findings. In addition, the study only assessed mistrust and did not include other variables found to be associated with vaccine rejection in the extant literature, such as concerns about vaccine safety and efficacy, which may be particularly relevant to COVID-19 vaccines in light of the unprecedented speed at which  Figure]). e Indirect associations were calculated as the product of paths from racial/ethnic group to mistrust and mistrust to outcome (Figure).
Role of the Funder/Sponsor: The funder had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.