COVID-19 Vaccine Hesitancy: Trust in Local Clinics Predict Perceptions of Vaccine Safety and Vaccination Status

Abstract

Background.

Despite widespread COVID-19 vaccine availability, vaccine hesitancy remains a barrier to achieving public health goals, especially in underserved communities. Trust in healthcare providers, scientists, and government institutions plays a critical role in shaping vaccine perceptions and uptake.

Purpose.

To inform targeted interventions, we tested a mediational path model examining predictors of perceived vaccine safety, effectiveness, and vaccination status among public housing residents.

Methods.

We surveyed 120 public housing residents in Central and Southeast Virginia. Participants completed questionnaires assessing trust in various people and institutions, a sense of community responsibility, beliefs about vaccination as a personal choice, and vaccination-related perceptions and status. A mediational path model was tested to examine direct and indirect associations.

Results.

Among 10 trust domains examined, only trust in the federal government and trust in local clinics and pharmacies significantly predicted vaccination-related perceptions and behavior. Trust in local clinics was directly associated with perceived vaccine safety (β = .42, 95% CI [.18, .66]) and effectiveness (β = .36, 95% CI [.11, .62]) and indirectly predicted vaccination status through perceived safety (β = .11, 95% CI [.01, .22]). Trust in the federal government directly predicted perceived effectiveness (β = .35, 95% CI [.04, .66]) but not safety. The final model explained 62% of the variance in vaccination status.

Discussion.

An indirect effect emerged where trust in local clinics predicted vaccination status via perceived safety, underscoring the critical role of local health centers in fostering vaccine confidence and uptake. Strengthening relationships between local providers and their communities may be a key strategy for addressing vaccine hesitancy. These findings highlight the importance of building trust to support vaccine uptake efforts.

Vaccine hesitancy, defined as the delay in acceptance or refusal of safe vaccines despite the availability of vaccination services, has increased globally and poses a significant threat to public health.¹ Although vaccines have been instrumental in preventing infectious diseases for more than two centuries, concerns regarding adverse effects, distrust in medical institutions, and sociocultural factors continue to fuel skepticism, particularly among racial and ethnic minority groups.² Vaccine acceptance, or the willingness to receive a vaccine deemed safe and effective, is shaped by social, cultural, and historical factors, leading to significant differences in vaccine uptake across populations and contexts.^3,4

The COVID-19 pandemic created an urgent need for a safe and effective vaccine to reduce illness and death. While the rapid development and distribution of COVID-19 vaccines marked an important milestone, rollout challenges including misinformation, evolving public health guidance, and temporary safety-related pauses such as the Johnson & Johnson vaccine suspension complicated public perception and trust.^5,6 More recently, the pharmaceutical company AstraZeneca’s request to withdraw its COVID-19 vaccine’s European approval due to safety concerns, further intensified public skepticism.⁷ As a result, COVID-19 vaccine hesitancy has become a significant barrier to achieving herd immunity and protecting vulnerable populations.⁵

COVID-19 vaccine hesitancy is particularly pronounced among racial and ethnic minority groups, including Black Americans, who experience disproportionate infection risk and severe outcomes.² Reported acceptance rates range widely, from 12% to 91.4%, highlighting the complexity of vaccine decision-making and the need for targeted public health strategies.^3,4 Despite widespread vaccine availability, COVID-19 continues to shape health and social outcomes across communities worldwide, with sustained patterns of seasonal surges in the United States. According to the Centers for Disease Control and Prevention (CDC), increases in COVID-19 cases and hospitalizations consistently occur in late summer (August-September) and peak during the winter months (December-January).^8–10 COVID-19 hospitalizations rose by 147% during the final 5 weeks of 2024, with the rate climbing from 1.7 to 4.2 per 100,000 between November 2024 and January 2025.⁸ These trends reflect an ongoing public health burden that persists despite the widespread availability of vaccines.

Yet, vaccine uptake in the U.S. remains suboptimal. As of May 2024, 69.5% of the population had completed a primary COVID-19 vaccine series, but only 40% have received a booster dose, and just 22.5% received the updated booster.⁹ These low rates raise urgent questions about the barriers to vaccine acceptance among subpopulations with historically lower uptake.

Black Americans have consistently reported some of the highest levels of COVID-19 vaccine hesitancy, with research documenting deeply rooted medical mistrust, stemming from historical injustices, structural inequities, and prior negative healthcare experiences, as major determinants of hesitancy.^4,11–13 For example, Willis et al. found that non-Hispanic Black adults had 2.42 times greater odds of COVID-19 vaccine hesitancy compared to White adults, and low trust in vaccines increased hesitancy more than elevenfold.¹² Similarly, Bogart et al. identified healthcare institutional mistrust and vaccine safety concerns as significant drivers of hesitancy among Black Americans.¹³

Taken together, these patterns highlight the central role of medical mistrust in shaping vaccine-related behaviors and support the need for models that treat mistrust not as a by-product of low information or access, but as a distinct and measurable factor. Our study builds on this premise by examining how trust and mistrust influence vaccine-related decision-making within marginalized populations.

Trust in healthcare is broadly defined as a psychological state in which individuals accept vulnerability based on positive expectations of healthcare providers and healthcare institutions.¹⁴ It reflects confidence that healthcare entities act in the patient’s best interest, demonstrate competence, maintain honesty, and uphold ethical standards.^15,16 In contrast, medical mistrust refers to skepticism or concerns about potential harm, often rooted in historical and systemic inequities, discrimination, and previous negative healthcare experiences, particularly among marginalized communities.^17–19

Prior research consistently links trust in healthcare institutions, providers, and government agencies to vaccine confidence and uptake.^13,20–22 Global evidence further demonstrates that confidence in organizations such as the World Health Organization (WHO), as well as trust in domestic scientists, and healthcare professionals drives vaccine acceptance globally, and across diverse populations.²³ Among Black Americans, strengthening trust in healthcare providers and addressing systematic inequities are critical for improving vaccine uptake.^12,24

Our study distinguishes between two related but conceptually distinct forms of trust: trust in personal doctors and trust in healthcare clinics. Trust in personal doctors refers to confidence in the competence, intentions, and communication of an individual provider. This form of trust may be built through ongoing relationships and personalized care.^15,25 In contrast, trust in clinics reflects perceptions of the broader healthcare setting, including its accessibility, policies, staff behavior, and overall safety.^26,27 It is possible for an individual to trust their doctor but harbor skepticism about the broader clinic or healthcare organization, particularly if they have experienced discrimination, administrative barriers, or poor treatment in the clinical setting.^28,29 Conversely, someone might trust a clinic’s reputation but feel uncertain about their specific provider.³⁰ Exploring these dimensions separately provides a more nuanced understanding of how different types of trust influence vaccine attitudes.^31,32

Vaccine uptake may also be influenced by an individual’s sense of responsibility to their community. Social Identity Theory provides a useful framework for understanding this process by proposing that people who strongly identity with a social group, such as their local community, are more likely to adopt behaviors that align with group norms and serve collective well-being.³³ Empirical work by Wakefield and Khauser support this relationship, showing that community identification predicted willingness to engage in protective behaviors, including vaccination, with communal responsibility fully mediating this association even after accounting for socioeconomic factors and age.³⁴ Thus, although Social Identity Theory is not formally tested in the current study, the theory informed out inclusion of community responsibility as a predictor of vaccine perceptions, illustrating how group-based motivations and a shared sense of duty to safeguard one’s community may contribute to vaccine acceptance, particularly in populations with historically lower vaccination uptake.

Present Study

The current study investigates how multiple forms of trust in individuals and institutions, along with beliefs about community responsibility, shape perceptions about vaccine safety, effectiveness, and ultimately vaccine uptake among predominantly Black Americans residing in income-based public housing in Virginia. Public housing residents represent a critically important yet understudied population, as they experience heightened exposure to structural inequities, have limited access to healthcare resources, and report disproportionately higher rates of medical mistrust and lower rates of vaccination.^11,13,35 Extant research suggests trust in healthcare providers, scientists, and governmental bodies may influence individuals’ perceptions of vaccine safety and effectiveness, key beliefs that, in turn, shape willingness to get vaccinated.^{12,13,17–23} Additionally, both a sense of community responsibility to get vaccinated and the belief that vaccination is an individual choice have been linked to vaccination attitudes.²⁴ However, less is known about how multiple forms of trust function together in a model that includes both cognitive beliefs about vaccines and actual vaccination behavior, particularly within high-risk, historically marginalized populations.

To address this gap, we tested a mediational path model in which trust in various people and institutions, along with beliefs about community responsibility and individual choice, predicted perceptions of vaccine safety and effectiveness, which in turn predicted COVID-19 vaccination status. This study contributes to a deeper understanding of how trust at both institutional and local levels may shape vaccine uptake. While we focused on current vaccination status, findings may inform strategies to support ongoing and future booster campaigns, addressing the need for research on COVID-19 vaccine hesitancy.³⁶

Method

Participants

Participants were 120 residents from public housing in Central and Southeast Virginia. Inclusion criteria for the study included being at least 18 years of age or older and being a resident of one of nine agencies, two of which provide housing choice vouchers, seven of which are PHAs. The majority of participants identified as Black (n = 111, 92.5%), and female (n = 81, 67.5%), with a mean age of 52.64 years (SD = 15.44, range 18–75). Approximately two-thirds of the participants (n = 76, 65%) had received an initial COVID-19 vaccination at the time of data collection. See Table 1 for participant demographics.

Table 1.

Demographics of Study Participants.

Variable	Frequency	Percentage
Race
African American or Black	111	92.5%
White	5	4.2%
Biracial or multiracial	2	1.7%
Other racial background	1	0.8%
Don’t know or unsure	1	0.8%
Sex
Male	15	12.5%
Female	81	67.5%
Missing or unknown	24	20.0%
Age group
18–30	9	7.5%
31–40	17	14.2%
41–50	11	9.2%
51–60	18	15.0%
61–70	22	18.3%
71 +	43	35.8%
Vaccinated
Yes	76	63.3%
No	41	34.2%
Missing	3	2.5%

Procedure

The study was supported by an established Community Advisory Board (CAB) comprised of residents from income-based housing communities. As part of a community-engaged research approach, the CAB played a key consultative role in ensuring cultural relevance and appropriateness throughout the study. CAB members reviewed and refined survey language to enhance clarity, contextual fit, and offered feedback that informed recruitment strategies.

Participants were recruited from a larger study pool which used multiple recruitment approaches including posting flyers in PHA buildings, word-of-mouth outreach through CAB members and participants, and door-to-door recruitment by the research team. Interested participants were consented and emailed the survey questionnaire via a REDCap link. Participants were compensated $20 for completing the survey. Data for the current study were collected between June through December 2021. The study received human subjects’ approval from the participating university Institutional Review Board.

Measures

Trust in People and Institutions.

Informed by previous literature, such as the COVID collaborative survey,³⁷ and refined through consultation with the CAB, participants rated their level of trust in 10 areas, using single-item measures. Single-item indicators were selected intentionally to minimize participant burden, given the study’s community-based context, and to ensure accessibility for respondents with varying levels of survey familiarity. Participants rated their trust with response options ranging from Not at All (1) to Completely Trust (5). Items included trust in the (1) federal government, (2) food and drug administration (FDA), (3) pharmaceutical companies that make the vaccines, (4) scientists who make the vaccines, (5) Dr Fauci, (6) Mr Trump, (7) Mr Biden, (8) pharmacies and clinics where people get vaccinated, (9) your personal doctor, and (10) Black elected officials in your community.

Sense of Community Responsibility.

Participants rated their agreement to the statement “People have a responsibility to get vaccinated to help stop the spread of the [COVID-19] virus in their community.” Response options ranged from Strongly Disagree (1) to Strongly Agree (7).

Sense of Individual Choice.

Participants rated their agreement to the statement “Getting [COVID-19] vaccinated is an individual choice and community considerations should not come into it.” Response options ranged from Strongly Disagree (1) to Strongly Agree (7).

Trust in the Safety of the Vaccine.

Participants indicated the degree to which they trusted that COVID-19 vaccines are safe, ranging from Not at All (1) to Completely Trust (5).

Trust in the Effectiveness of the Vaccine.

Participants indicated the degree to which they trusted that COVID-19 vaccines are effective, ranging from Not at All (1) to Completely Trust (5).

Vaccination Status.

Participants indicated whether they had (1) or had not (0) received the COVID-19 vaccine.

Data Analytic Strategy

The study variables and hypotheses were initially driven by variables identified in existing literature as important constructs related to vaccine hesitancy, and subsequent model exploration was data-driven to inform the most accurate and parsimonious model possible. The 10 trust items remained as distinct items in analyses. Potential path model variables were first screened using bivariate correlations and regression models. Variables associated with the mediators (trust in the safety of the vaccine, trust in the effectiveness of the vaccine) and the main endogenous variable (vaccination status) at p < .20 were included in subsequent path analysis to avoid excluding potentially meaningful predictors at the initial stages and to support the specification of an exploratory path model.³⁸ The tested path model can be found in Figure 1. The path model, including direct and indirect effects, was analyzed using Mplus version 8.4, using the weighted least squares mean and variance adjusted (WLSMV) estimator with 5000 bootstrap replications.³⁹ Missing data were handled using listwise deletion for categorical outcomes and pairwise deletion for exogenous variables, consistent with Mplus defaults for WLSMV estimation. Significance was determined using 95% bias-corrected (BC) confidence intervals (CIs) for both direct and indirect effects. Standardized direct path coefficient results can be found in Figure 2 with either dotted (non-significant) or solid (significant) lines.

Figure 1.

Tested path model based on bivariate and regression analyses.

Figure 2.

Results of the direct path model. Note. Path coefficients are standardized. Solid lines indicate significant paths and dashed lines indicate non-significant paths. *p < .05 **p < .01, ***p < .001.

Results

Preliminary Analyses

Prior to conducting analyses, data were examined for normality. No extreme outliers were found, and kurtosis and skewness values were found to be within normal range for all variables. Bivariate correlations can be found in Table 2 Vaccination status was related to nearly all of the variables at the bivariate level. Those who were vaccinated tended to be older and had greater levels of trust in all trust items except for trust in Mr. Trump (this relationship was not significant). Vaccination status also was not correlated with date of data collection. Vaccination status was positively correlated with the statement that participants have a community responsibility to get vaccinated and negatively correlated with the statement reflecting that vaccination was an individual’s choice. Vaccination status was also positively correlated with trust in the safety of the vaccine and trust in the effectiveness of the vaccine.

Table 2.

Bivariate Correlations of Study Variables.

Variable	1	2	3	4	5	6	7	8	9	10	11	12	13	14	15	16
1. Date	—
2. Age	−.01	—
3. Trust fed. gov.	.12	.33**	—
4. Trust FDA	.05	.23*	.57***	—
5. Trust pharma. companies	.05	.34**	.60***	.63***	—
6. Trust scientists	.08	.31**	.44***	.46***	.71***	—
7. Trust Fauci	.02	.43***	.59***	.63***	.58***	.65***	—
8. Trust Trump	.05	−.10	.23*	.23*	.30**	.10	.16	—
9. Trust Biden	.02	.31**	.60***	.53***	.46***	.50***	.63***	−.03	—
10. Trust local clinics	.15	.34**	.53***	.44***	.48***	.50***	.49***	.07	.57***	—
11. Trust my doctor	.04	.38***	.46***	.33***	.37***	.37***	.40***	−.02	.53***	.62***	—
12. Trust my black elected official	.14	.07	.55***	.56***	.48***	.39***	.54***	.14	.54***	.47***	.46***	—
13. Vaccination is community responsibility	.15	.43***	.39***	.43***	.40***	.44***	.44***	−.19*	.57***	.48***	.53***	.29**	—
14. Vaccination is individual choice	−.05	−.27*	−.30***	−.15	−.33***	−.46***	−.30**	.06	−.34***	−.30***	−.22*	−.23*	−.24*	—
15. Perceived vaccine safety	.01	.41***	.47***	.47***	.53***	.61***	.55***	−.06	.43***	.59***	.40***	.32**	.56***	−.41***	—
16. Perceived vaccine safety	.03	.43***	.59***	.52***	.56***	.54***	.53***	−.04	.48***	.54***	.39***	.38***	.48***	−.34***	.82***	—
17. Vaccination status	.01	.66***	.41***	.24*	.40***	.44***	.40***	−.02	.43***	.53***	.42***	.26**	.60***	−.39***	.61***	.52***

Note. Pearson correlations presented for all continuous variables; Spearman Rho correlations presented for the variable of Vaccination Status.

^*p < .05

^**p < .01,

^***p < .001.

Path Model

The data-driven model had excellent model fit, χ² (18) = 18.02, p = .45, RMSEA = .004, SRMR = .075, and excellent incremental fit, CFI = 1.00, and TLI = 1.00.⁴⁰ The model predicted over 37% of the variance in perceived COVID-19 vaccine effectiveness, nearly 44% of the variance in perceived COVID-19 vaccine safety, and 62% of the variance in vaccination status. Overall, trust in local clinics and trust in the federal government emerged as the most consistent predictors across mediators and the primary endogenous variable.

Direct Effects.

Significant direct effects were found between many of the predicted paths (see Table 3 and Figure 2).

Table 3.

Results of the Direct Paths With Bootstrapped CIs for the Final Model.

	β	p	95% CI
Vaccination status
Perceived vaccine safety	.27	.011	.06, .48
Trust pharma. companies	.03	.769	−.18, .24
Trust Dr. Fauci	−.14	.224	−.37, .09
Trust my doctor	−.08	.535	−.31, .16
Vaccination is community responsibility	.16	.160	−.06, .38
Vaccination is individual choice	−.15	.066	−.31, .01
Age	.61	.000	.40, .81
Perceived vaccine effectiveness
Trust local clinics	.36	.005	.11, .62
Trust fed. government	.35	.028	.04, .66
Perceived vaccine safety
Trust local clinics	.42	.001	.18, .66
Trust scientists	.26	.078	−.03, .56
Trust FDA	.04	.794	−.23, .30
Vaccination is individual choice	−.09	.288	−.26, .08
Perceived vaccine safety
Perceived vaccine effectiveness	.59	.000	.27, .91
Perceived vaccine effectiveness
Vaccination status	−.13	.195	−.33, .07

Note. β represents STDYX estimates.

Vaccine Safety.

Trust in the safety of the vaccine was strongly predicted by trust in local clinics (β = .42, 95% CI [.18, .66]), however the three other tested predictors were not significant, including trust in scientists (β = .26, 95% CI [−.03, .56]), trust in the FDA (β = .04, 95% CI [−.23, .30]), and the endorsement that getting vaccinated is an individual choice (β = −.09, 95% CI [−.26, .08]).

Vaccine Effectiveness.

Trust in the effectiveness of the vaccine was significantly predicted by both trust in the federal government (β = .35, 95% CI [.04, .66]), and trust in local clinics (β = .36, 95% CI [.10, .62]). The two mediators of trust in the effectiveness of the vaccine and trust in the safety of the vaccine were also strongly correlated with each other (β = .59, 95% CI [.27, .91]).

Vaccination Status.

Vaccination status was directly predicted by the mediator of trust in the safety of the vaccine (β = .27, 95% CI [.06, .48]), and age (β = .61, 95% CI [.40, .81]). The other five tested predictors however, did not significantly predict vaccination status, including, trust in pharmaceutical companies that make the vaccines (β = .03, 95% CI [−.18, .24]), trust in Dr. Fauci (β = −0.14, 95% CI [−.37, .09]), trust in your personal doctor (β = −.08, 95% CI [−.31, .16]), a community sense of responsibility to get vaccinated (β = .16, 95% CI [−.06, .38]), and the belief that choosing to get vaccinated is an individual choice (β = −0.15, 95% CI [−.31, .01]). Vaccination status was not significantly correlated with trust in the effectiveness of the vaccine (β = −0.13, 95% CI [−.33, .07]).

Indirect Effects.

Multiple indirect effects were tested from the exogenous variables of trust in the scientists who make the vaccines, trust in the local clinics, trust in the FDA, and vaccination being an individual choice, through the mediator of vaccine safety, to the endogenous variable of vaccination status (See Table 4). The only significant indirect path supported was trust in local clinics predicting vaccination status through its effects on trust in the safety of the vaccine (β = .11, 95% CI [.01, .22]).

Table 4.

Results of the Indirect Paths From Variables of Interest to Vaccination Status.

Path	β	p	95% CI
Trust scientists → perceived vaccine safety → vaccination status	.07	.17	(−.03, .17)
Trust FDA → perceived vaccine safety → vaccination status	.01	.81	(−.07, .09)
Trust clinics → perceived vaccine safety → vaccination status	.11	.03	(.01, .22)
Vaccination is individual choice → perceived vaccine safety → vaccination status	−.03	.37	(−.08, .03)

Note. FDA= Food and Drug Administration; Vaccination is Individual Choice = The decision to get vaccinated is an individual choice. Significance was inferred based on the bootstrapped 95% confidence intervals.

Discussion

The present study explored perceptions of trust in various individuals and organizations to identify salient predictors of beliefs about the COVID-19 vaccine; specifically, perceptions of safety and effectiveness, and how these beliefs may be associated with vaccination status. Across all the tested entities examined, only two of the ten trust items remained significant in the final path model: trust in the federal government and trust in local pharmacies and clinics. Trust in local clinics predicted greater perceived vaccine safety and effectiveness, while trust in the federal government predicted perceived vaccine effectiveness. In turn, perceived vaccine safety emerged as the most proximal predictor of vaccination status.

Study findings clarify the distinct roles of institutional trust versus interpersonal trust. The saliency of trust in the federal government aligns with prior research showing that trust in national agencies influences uptake of recommended preventive behaviors during outbreaks on both COVID-19 and earlier public health responses. For instance, vaccination uptake strategies often rely on transparent and timely communication, which necessitates trust in the institutions disseminating information.⁴¹ Reviews of previous virus outbreaks, such as the 2009/2010 H1N1 influenza pandemic or the 2014/2015 Ebola crisis, suggest that communities with higher trust in government agencies were more likely to engage in preventive behaviors like vaccination.^42,43 Although trust in government frequently predicts COVID-19 vaccine intentions,^44,45 cross-national patterns remain variable,⁴⁶ highlighting the importance of local context. Notably, trust in individual political or scientific figures (e.g., Dr. Fauci, Mr. Trump, Mr. Biden) did not predict vaccine beliefs or behavior in our sample. This suggest that for residents of income-based housing, institutional trust may outweigh trust in individual messengers, a pattern warranting further study in diverse populations.

Equally important is the absence of effects for trust in one’s personal doctor. Contrary to assumptions that clinicians strongly shape vaccine decisions, trust in one’s personal doctor did not predict vaccine beliefs or uptake in this population. In communities facing inconsistent provider relationships, high turnover, or fragmented care, institutional trust, particularly trust in the broader clinic, may play a stronger role than interpersonal trust. These results highlight the need to distinguish between interpersonal (doctor-level) and institutional (clinic-level) trust when examining vaccine hesitancy in underserved settings.

Our findings suggest that trust in local clinics where individuals can receive vaccinations may be associated with more favorable perceptions of vaccine safety and effectiveness. In particular, we observed a significant indirect association between trust in local clinics and vaccination status, mediated by vaccine safety beliefs. While this pathway does not establish directionality, it highlights the potential for local health institutions to serve as trusted sources of vaccine-related information and support. Given that many individuals may associate local providers with personalized care and safety assurances, whereas the federal government may be more closely associated with policy and distribution, these results may point to distinct but complementary roles for different trust agents. Future longitudinal or experimental studies are needed to better understand how these relationships unfold over time and whether efforts to build trust in local healthcare settings could contribute to improved vaccination rates. Additionally, although community responsibility beliefs did not significantly predict vaccine outcomes in the final model, they remain conceptually relevant and warrant further investigation in larger, more diverse samples.

Implications

The observed link between trust in local clinics and vaccine perceptions may inform community-level engagement strategies. Local pharmacies and health centers, due to their proximity, familiarity, and continuity, are well positioned to address vaccine concerns through personalized, culturally competent communication. Approaches such as motivational interviewing, clear explanations of vaccine risks and benefits, and proactive dissemination of updated vaccine information may strengthen trust and shift safety perceptions. Because federal-level messaging can be hindered by institutional distance or historical inequities, community-based clinics may hold a unique advantage in fostering trust among marginalized populations.

At the same time, building trust at the clinic level requires addressing persistent barriers. Community members frequently report experiences of discrimination, limited time with providers, and inadequate communication.^47–50 These barriers highlight the need for relational approaches grounded in empathy, cultural humility, and transparency. Local clinics can help address these challenges through nonjudgmental dialogue, culturally tailored communication, and value-based care models that emphasize ongoing relationships rather than transactional encounters.⁵¹ Additionally, updated COVID-19 vaccines targeting circulating variants (e.g., Moderna, Novavax, Pfizer-BioNTech) provide opportunities for clinics to engage in proactive, trust-building communication within their community.⁵² While the FDA provides educational material on these vaccines, local clinics could use the opportunity to support direct communication with community members and strengthen trust. Finally, broader systems-level strategies, such as Accountable Care Organizations and Patient-Centered Medical Homes, may complement clinic-level trust building by reinforcing transparency, coordination, and long-term relationship-building.^53,54

Taken together, these findings demonstrate the pivotal role of institutional trust, especially trust in local clinics, in shaping vaccine perceptions and uptake. Strengthening clinic-community relationships may represent a critical strategy to reduce vaccine hesitancy and advance health equity in historically marginalized communities.

Limitations

Several limitations should be noted. First, the sample was intentionally composed of predominantly Black Americans (92.5%) living in low-income, public housing, a population at higher risk for vaccine hesitancy and health inequities. While appropriate for the study aims, this homogeneity limits generalizability, and findings may not extend to other racial, socioeconomic, or geographic groups but provide valuable insights for similar underserved populations. Participants who chose to take part in a health-focused study may have been more trusting of healthcare systems and less vaccine-hesitant than the broader community, highlighting the potential for self-selection bias. Additionally, the use of single-item measures of trust and vaccine beliefs is an important limitation. While such items efficiently capture broad perceptions, they limit the ability to assess internal consistency, reliability, and the multidimensional nature of these constructs. This may reduce measurement precision and attenuate observed associations. Nonetheless, the items were theoretically grounded and demonstrated conceptually consistent associations with related variables, supporting their construct validity. Future research should create, test, and incorporate validated multi-item scales to better capture the complexity of trust and vaccine beliefs and allow for more robust psychometric evaluation. Additionally, although existing literature and theory were utilized to include relevant predictors, the study design was cross-sectional and thus no causal inference can be assumed. It is possible that paths are bidirectional, such that individuals who received the COVID-19 vaccine at their local clinic may then have felt a sense of trust and safety. Data were collected between June and December 2021, during a period of rapidly evolving vaccine policies and public messaging. These shifts may have influenced participants’ levels of trust and hesitancy, and findings should be interpreted within that temporal context. Further, while we examined key trust-related factors, vaccine hesitancy is complex and also shaped by psychosocial factors, health beliefs, and misinformation, areas future research should explore in more depth.⁵⁴ Lastly, although significant direct and indirect effects speak to the strength of the path model, the relatively small sample size may have limited statistical power, and some non-significant indirect effects might reach significance in a larger sample.

Conclusions

The present findings demonstrate that perceived vaccine safety functioned as the most proximal and influential predictor of COVID-19 vaccination status among residents of income-based housing. Importantly, trust in local clinics, rather than trust in individual providers or political figures, emerged as the most salient factor shaping these safety perceptions. This highlights the critical role that community-based healthcare institutions play in influencing vaccine-related decision-making within marginalized populations. As public health agencies confront ongoing challenges, including booster fatigue, low uptake of updated COVID-19 formulations, and declining confidence in routine immunizations such as the measles–mumps–rubella (MMR) vaccine, strengthening trust at the local institutional level remains essential.⁵⁵ Our findings suggest that local pharmacies and clinics may serve as especially effective and credible messengers for communicating vaccine safety, addressing concerns, and supporting informed decision-making. Enhancing the visibility, transparency, and cultural responsiveness of local clinics may improve institutional credibility and, in turn, bolster perceptions of vaccine safety, ultimately increasing vaccination uptake. Strategies such as motivational interviewing, providing clear and accessible communication, community-engaged outreach, and culturally humble care practices may be particularly impactful in communities historically affected by medical mistrust.⁴⁷ Taken together, these results highlight a promising pathway for public health interventions: investing in trust-building within local healthcare institutions may help strengthen safety beliefs, reduce vaccine hesitancy, and support long-term efforts to advance vaccine equity.

Biographies

Brynn E. Sheehan, PhD, is an Associate Professor of Psychiatry and Behavioral Sciences and Director of the Research & Infrastructure Service Enterprise at the Macon & Joan Brock Virginia Health Sciences of Old Dominion University in Norfolk, Virginia. Her research examines structural drivers of health inequities and psychosocial contributors of substance use and health disparities, with an emphasis on mental health and community-engaged health policy research in historically under-resourced communities.

Miasha O’Neal, MPH, is a Research Support Specialist for the Research & Infrastructure Service Enterprise at Macon & Joan Brock Virginia Health Sciences at Old Dominion University, where she is also pursuing her Doctorate of Health Sciences. Her primary research of interest is community-based public health, with particular emphasis on health equity and mental health.

Kapri Hannon, MPH, is a Program Coordinator at the Center for Public Health Initiatives at NSU in Norfolk, Virginia. Her core interests lie in community-based public health initiatives. Her work includes program development, capacity-building with local organizations, and applied qualitative research.

Andrew Plunk, PhD, is an ethicist and psychiatric epidemiologist whose work focuses on trust, perceived legitimacy, and the ethical design of community-engaged policy and research initiatives.

Data Availability Statement

Our data is linked to a larger study as part of the RADx-UP initiative and contains sensitive information. Due to privacy concerns and the nature of this ongoing research, we are unable to provide or share the data for open use. Access to the data is restricted to protect participant confidentiality and adhere to the guidelines established by the larger study.