AbstractVaccine hesitancy or refusal poses a significant public health challenge resulting in the resurgence of preventable diseases and undermining the effectiveness of national and global health initiatives. This study investigates multilevel determinants of COVID-19 vaccine hesitancy in the Dominican Republic (DR) shortly after the launch of the national COVID-19 vaccination campaign in February 2021. Participants aged 18 years and older were enrolled through a national multistage cluster survey conducted from June-October 2021. The Health Belief Model guided the selection of potential factors contributing to vaccine hesitancy. Hierarchical mixed-effect logistic regression models were used to examine individual, household, and community factors associated with vaccine hesitancy. COVID-19 vaccine hesitancy was observed in 12.6% (95% CI: 11.7–13.5%) of participants (n = 5,566), with spatial variations at the cluster level. Individual factors associated with lower odds of vaccine hesitancy included older age, higher education levels, mulatto ethnicity, and perceiving vaccination as crucial for health. In contrast, factors significantly associated with hesitancy included being born in the DR and concerns about COVID-19 vaccine side effects. For factors at the household level, differential trust in health information sources significantly influenced vaccine hesitancy, with certain sources correlating with increased hesitancy and others with reduced vaccine hesitancy. Better access to healthcare, as indicated by a higher number of hospitals per population, was paradoxically associated with increased vaccine hesitancy. Future strategies to reduce vaccine hesitancy in the DR might consider these multifaceted factors.
IntroductionThe COVID-19 pandemic posed unprecedented challenges to global healthcare systems and communities. The rapid development and distribution of safe and effective vaccines became crucial tools in controlling the spread of the virus1. However, vaccine hesitancy during the COVID-19 pandemic resulted in reduced vaccine uptake in certain settings. COVID-19 vaccines significantly reduced disease severity, hospitalizations, and deaths, but vaccine hesitancy remains a persistent challenge2,3.Vaccine hesitancy encompasses both delayed and outright refusal of vaccination that manifests differently across locations and social contexts4. Individuals’ motivations for vaccination, shaped by beliefs and attitudes towards health, vary widely5. Evaluating the risk-benefit perspective requires careful consideration of perceived risks derived from information, such as that circulating through public media and social channels, including concerns about vaccine side effects and exposure to anti-vaccination narratives6. Personal experiences such as financial hardship, political affiliations and having chronic medical conditions also contribute to this assessment7,8. Importantly, vaccine hesitancy is shaped not only by individual factors but also by household dynamics and the socio-cultural context of communities9,10. While an individual’s general opposition to vaccines may increase hesitancy towards COVID-19 vaccines specifically, new vaccines have the potential to introduce unique beliefs about hesitancy that warrant thorough investigation11.Various conceptual frameworks for identifying possible factors associated with vaccine hesitancy elucidate the diverse components influencing health behavior decision-making, including the Health Belief Model (HBM)12, theory of planned behavior13, Social-Ecological Model14, 5As Framework (Access, Affordability, Awareness, Acceptance, and Activation Framework)15, COM-B Model (Capability, Opportunity, Motivation, and Behavior Model)16, Three Cs Model (Confidence, Complacency, and Convenience Model)17, and the 5 C psychological antecedents (confidence, constraints, complacency, calculation, and collective responsibility)18. These frameworks highlight the complexity of vaccine hesitancy, encompassing the interplay between psychological, social, cultural, and contextual factors2,13,19. For instance, the socio-ecological model was used to examine factors influencing community engagement for general vaccination in India, identifying both enablers and barriers across individual, community, organizational, and policy levels. While supportive policies and social mobilization promoted community engagement, challenges such as limited formal strategies, power imbalances, and insufficient institutional support hindered progress20. In Bangladesh, researchers employed the HBM, the Theory of Planned Behaviour, and the 5 C framework of psychological antecedents to examine a range of psychological factors driving COVID-19 vaccine hesitancy. Their findings revealed that the Theory of Planned Behaviour provided the highest predictive accuracy in this context18.The HBM, a psychological framework, has been widely utilized to analyze COVID-19 vaccine hesitancy and and associated determinants2,18. The HBM encompasses components such as the perceived severity of and susceptibility to COVID-19, perceived benefits of and barriers to receiving COVID-19 vaccines, and cues to action. These cues can include implicit or explicit incentives or situations that motivate vaccination, such as information from mass media2.Multiple factors shape individuals’ decisions to accept or refuse COVID-19 vaccination, including employment status (e.g., whether a person is employed, unemployed, or retired), religiosity, political affiliation, gender, age, education, ethnicity, and income2,19,21. According to recent systematic reviews, primary reasons for vaccine refusal included a general opposition to vaccines, concerns about safety, the perception of COVID-19 as benign, distrust of health authorities, doubts about scientific research and vaccine efficacy, belief in pre-existing immunity, and uncertainty about the vaccine’s origin8,22. It is important to note that behavior refers to observable actions or responses, such as deciding whether or not to get vaccinated. In contrast, attitude encompasses an individual’s internal feelings, beliefs, and evaluations regarding a subject. While attitudes can influence behavior, they do not always result in action. For instance, someone who is vaccine hesitant might still need to get vaccinated due to travel restrictions or employment requirements.While much of the existing research on COVID-19 vaccine hesitancy has focused on high-income countries8,19, our study shifts the focus to the Dominican Republic (DR), a middle-income nation with a Gross National Income (GNI) per capita of USD 9,710 in 202323. The DR’s unique sociocultural landscape is shaped by a rich blend of indigenous, African, and European influences, with spirituality and religion playing a central role in shaping societal values and healthcare practices, creating a distinctive approach to health-seeking behaviors and community support systems. Such sociocultural factors, including historical inequalities and varied access to healthcare services, are critical to understanding the context of vaccine hesitancy in the country24. Additionally, previous studies have predominantly explored vaccine hesitancy in relation to individual, household, or community-level factors, but rarely all three simultaneously19. Previous research conducted in the DR which used a part of this study’s dataset focused narrowly on trust in information sources25, while our study addresses this gap by exploring the association of individual, household, and community-level factors with COVID-19 vaccine hesitancy in the DR. Through this approach, we aim to enhance the understanding of COVID-19 vaccine hesitancy in the DR, broaden existing knowledge, and lay the groundwork for identifying contextually relevant factors to reduce vaccine hesitancy for future pandemics.MethodsTheoretical backgroundWe selected the HBM2 as the most suitable theoretical framework because it aligned well with our 3-level dataset (namely individual, household and community), and research question. Furthermore, the decision to use the HBM in this study was driven by its robustness in capturing the cognitive factors influencing health behaviours. The HBM’s emphasis on individual perceptions aligns well with the goal of understanding personal decision-making processes regarding vaccination. It allows for a nuanced analysis of how perceived risks and benefits, combined with external cues, influence vaccine acceptance or hesitancy. Table 1 in supplementary materials outlines the categorization of our factors based on the HBM model.Study locationThe DR, located in the Caribbean on the island of Hispaniola alongside Haiti (Fig. 1), has a population of approximately 10.5 million, making it the second most populous country in the Caribbean. Administrative divisions include 31 provinces and the Santo Domingo National District, comprising 155 municipalities, 386 district municipalities, 1565 sections, and 12,565 barrios/parajes26. Barrios refer to neighborhoods or urban communities, while parajes are small rural settlements, similar to unincorporated communities in the United States. Despite 80% of the population residing in urban and semi-urban areas, only about 20% of the total barrios/parajes are designated as urban. Over the past two decades, the DR has witnessed consistent economic growth, leading to an overall reduction in poverty. However, persistent social inequities endure, with higher poverty rates prevalent in urban slums and rural regions, particularly those close to the Haitian border27.Data sources and samplingWe utilized data collected from a field study in 2021 for the identification of the most common causes of acute febrile illnesses in the DR, which used a spatial random sampling method to select specific clusters (barrios/parajes), across the country (see Fig. 1). In brief, the 31 provinces and the Santo Domingo National District were categorized into five areas for logistical efficiency. Within each area, a predetermined number of urban and rural clusters were chosen utilizing a spatially representative sampling method. In urban clusters, a grid approach was employed for household selection28. Conversely, in rural clusters, households were selected using a spatially representative sampling method designed to optimize spatial dispersion, thereby preventing oversampling in both densely populated and sparsely populated regions29.Fig. 1The study area map, including the sampled clusters (outlined by red lines) across the Dominican Republic. The figure was created by the authors using ArcGIS Pro (version 3.4), available at https://www.esri.com/en-us/arcgis/products/arcgis-pro/overview.Full size image Sampled households and individuals were nested within primary sampled clusters across the country, and we collected detailed information at both the individual and household levels30 using a two-part questionnaire. The first section, administered to all participants, included individual-level questions covering areas such as demographics, personal health status and vaccination attitudes. Household-level questions addressed shared circumstances, such as household environment (electricity, water, toilets, etc.) and whether any household member had died from COVID-19. Household representatives were selected by the household, generally one of the adults capable of responding to questions on their behalf. The structure of the questionnaire was therefore well suited to an analysis at the individual and household levels. A total of 23 households per cluster were chosen in 132 clusters, with an exception for two provinces, which were associated with a study on clinical surveillance of acute febrile illnesses31. In these two provinces, each cluster was oversampled with 60 households. All household members aged ≥ 5 years were invited to participate in the national survey. However, for the purposes of the current study, we only included individuals aged 18 years or older, i.e. adults. Individuals less than 18 years of age were excluded as they would typically not play a decisive role in vaccine uptake decision-making.Additionally, we used open-source datasets, such as motorized travel time to the nearest health facility and poverty index, both available in a raster format32 and census data presented in Table 1 in the supplementary file, to obtain information at the community level. To extract the values, we matched the geographic coordinates of households with the corresponding pixels in the raster files.OutcomesThe primary outcome of interest was COVID-19 vaccine hesitancy. To capture nuanced attitudes towards COVID-19 vaccination, we adopted a two-pronged approach in our analysis, including both binary and ordinal variables as outcomes. By employing both models, we aimed to gain a comprehensive understanding of vaccine hesitancy, accounting for both a simplified binary categorization and the spectrum of hesitancy levels.We created a binary vaccine hesitancy variable by categorizing participants’ responses to the question, “If it were available to you right now, would you accept a COVID-19 vaccine?” with “No, definitely not,” “Probably not,” “Don’t know/Not sure,” and “Probably yes,” as indicating hesitancy, and “Yes, definitely” or “Already vaccinated” as not hesitant. We included “Probably yes” in the hesitancy category because every degree of uncertainty or reluctance reflects a form of vaccine hesitancy, which we believe should be captured in the binary model. This approach ensures that even the mildest forms of hesitancy are accounted for, as they still represent potential barriers to vaccination.Additionally, we assessed vaccine hesitancy as an ordinal variable to acknowledge varying degrees of hesitancy. Responses ranged from strong hesitancy (“No, definitely not”, “Probably not” and " Don’t know/Not sure”) to mild hesitation (“Probably yes”), and no hesitancy (“Yes, definitely” or “Already vaccinated”). This approach allowed us to capture the spectrum of participants’ attitudes towards the vaccine.While this study focuses on findings related to the binary outcome classification, results for the ordinal outcome have been provided in Supplementary Table 2.Data analysisCluster analysisSpatial Scan Statistics, using SaTScan version 10.1.3, were employed to identify clusters with high or low levels of vaccine hesitancy33. The Bernoulli model, implemented in SaTScan, allows for the detection of clusters by comparing the observed distribution of binary outcomes (vaccine hesitancy, in this case) to what would be expected under the assumption of spatial randomness.Variables selection for modellingTo determine the appropriateness of each predictor for inclusion in our 3-level hierarchical mixed effect models, we calculated bivariate associations between each pair of potential predictors and both outcomes. Variables with a p-value < 0.05 were included in our models. To ensure the robustness of our multilevel modeling approach, we conducted a comprehensive assessment of potential associations among the predictor variables as follows.Variance Inflation Factors (VIFs) were computed to investigate potential correlations among continuous/ordinal variables, and a value above 7 was considered indicative of strong correlation. These variables were categorized based on their conceptual meaning. If high VIF values were observed within each group, a more general variable was selected based on the researchers’ judgment. For instance, the variables “the degree of concern regarding the impact of COVID-19 on an individual’s health” and “the degree of concern regarding the impact of COVID-19 on family health” exhibited elevated VIFs, each having a value equal to or greater than 7, which is considered notably high. Consequently, only the former variable was included in the modeling process.For categorical variables, we calculated Cramér’s V, which measures chi-squared effect size, to determine the strength of association between different categorical predictors. We did not observe any strong associations, as all values were less than 0.3.Three-level modellingTo quantify the role of each variable in vaccine hesitancy in the DR, we employed a three-level mixed-effect multivariable logistic regression modelling approach due to the hierarchical structure of our data. Specifically, individuals were nested within households, and households within clusters. We utilized a generalized linear modelling approach since our outcomes were ordinal and binary. StataSE (version 18, College Station, etc.) was used for statistical modeling, and R 4.3.2 for data preparation.The regressions incorporated weights derived from the survey design, considering the sampling process in three stages. Initially, the likelihood of a cluster being selected was computed by taking into account the total number of clusters within each region, and the corresponding weight was defined as the inverse of the selection probability for that category. Subsequently, the probability of a household being selected was determined based on the total number of households in each cluster, and the associated weight was the inverse of the household selection probability. In the final stage, the weights from the first two steps were multiplied and adjusted for a finite population. These sampling weights at each level were incorporated into our models. A comprehensive explanation of the weight calculation is available in a recently published work26.We initially included all the predictors selected based on VIF and bivariate associations with the outcome in the hierarchical model. The final model included only the variables that were significant in the initial model at p-value < 0.05.ResultsDescriptive statisticsTable 1 shows the descriptive statistics for independent variables included in our models. The mean age of the participants was 46 years (standard deviation 17.85). Most participants were born in the DR (95.49%). Ethnicity varied, with mulatto (51.68%) and mestizo (33.44%) groups being the predominant categories. Most participants reported having completed primary or secondary school (36.67% and 36.20% respectively). Participants primarily reported a mixed work environment including both indoor and outdoor (15.32%).Table 1 Descriptive statistics of characteristics of adults living in the Dominican Republic (surveyed June-October 2021).Full size tableRegarding health-related factors, the majority (56.63%) expressed little concern about the serious adverse effects of the COVID-19 vaccine. Attitudes towards vaccination revealed nuanced perspectives: a significant proportion strongly agreed that vaccinating against COVID-19 is important for their health (69.44%) and that its benefits outweigh the risks (63.55%). Trust in various sources of health information exhibited similar patterns, with a notable skepticism towards information obtained from social media.Cluster analysisCluster analysis using SaTScan revealed 16 statistically significant clusters of lower (n = 4) and higher (n = 12) vaccine hesitancy across the DR. Notably, the study identifies clusters 2 (North-central, near the coast) and 9 (Southeast, near the coast) as areas where all individuals were hesitant to COVID-19 vaccination (Fig. 2).Fig. 2Geographic clustering of COVID-19 vaccine hesitancy and readiness among adults in the Dominican Republic (surveyed June-October 2021). All clusters are statistically significant at P < 0.05. The figure was created by the authors using ArcGIS Pro (version 3.4), available at https://www.esri.com/en-us/arcgis/products/arcgis-pro/overview.Full size imageModeling results Among 5566 adults surveyed in DR, vaccine hesitancy was 12.6% (n = 699 participants) (95% CI: 11.7–13.5%). The multilevel model identified several significant factors associated with COVID-19 vaccine hesitancy. Older age was associated with decreased odds of being vaccine hesitant, with odds ratio (OR) = 0.47 (95% CI: 0.28–0.80) for each year of age increase. Mulatto ethnicity was associated with a significantly reduced odds of being vaccine hesitant when compared to Caucasians, (OR = 0.11, 95% CI: 0.02–0.66). Notably, those born outside the DR exhibited higher levels of vaccine hesitancy, (OR = 43.67, 95% CI: 6.01-317.22). Education played a crucial role, with higher educational attainment associated with reduced hesitancy, particularly among those who completed secondary school, (OR = 0.17, 95% CI: 0.05–0.61) or university, (OR = 0.18, 95% CI: 0.03–0.95) compared to those did not have any formal education. Health-related factors, such as the perceived importance of vaccinating against COVID-19 for personal health, were found to be inversely associated with COVID-19 vaccine hesitancy (OR = 0.21, 95% CI: 0.09–0.49). However, higher concern regarding the adverse effects of the COVID-19 vaccine was positively associated with COVID-19 vaccine hesitancy (OR = 4.34 (95% CI: 2.15 8.78). The sources of health information were also associated with COVID-19 vaccine hesitancy. For instance, individuals who relied on guidance from health professionals and schools demonstrated lower levels of vaccine hesitancy compared to those who reported that they obtained information from social media.Trust in local government appeared to be associated with less COVID-19 vaccine hesitancy (OR = 0.68, 95% CI: 0.47–0.98), while trust in religious leaders (OR = 2.86, 95%CI: 1.04–7.85) and the media (OR = 1.93, 95% CI: 1.28–2.90) associated with higher COVID-19 vaccine hesitancy.Regarding community characteristics, the number of hospitals per population was associated with higher COVID-19 vaccine hesitancy (OR = 1.10, 95% CI: 1.06–1.14).Finally, the random effects analysis revealed variance of COVID-19 vaccine hesitancy at both the household and community levels, emphasizing the importance of considering these contextual factors using hierarchical modelling. Table 2 in the supplementary file also shows the results of modelling using ordinal variables as the outcome.Table 2 Factors associated with COVID-19 vaccine hesitancy based on the 3-level hierarchical model in the Dominican Republic (surveyed June-October 2021).Full size tableDiscussionThis study suggests a low level of vaccine hesitancy in the DR and identified multifaceted factors associated with COVID-19 vaccine hesitancy among adults in the DR. Our analysis reveals a complex interplay of demographic, socio-economic, health-related, and informational determinants linked to COVID-19 vaccine hesitancy. These factors underscore the critical need for targeted public health strategies to reduce vaccine hesitancy that takes into account factors such as age, ethnicity, educational attainment, perceived health benefits, concerns about adverse effects, sources of health information, and trust in authorities. Interestingly, our analysis revealed that better access to healthcare, indicated by a higher number of hospitals per population, was paradoxically associated with increased COVID-19 vaccine hesitancy. This counterintuitive finding suggests that in areas with better healthcare access, other factors, such as trust in the healthcare system or exposure to misinformation, might play a more significant role in influencing vaccine decisions. Additionally, spatial clustering, as well as variability observed within household and community contexts highlights the necessity of localized and context-sensitive approaches in tackling COVID-19 vaccine hesitancy and specific regions where targeted intervention efforts may be particularly warranted.Spatial analysis revealed significant heterogeneity as well as clustering of vaccine hesitancy (or lower levels thereof) across different regions in the DR. Higher vaccine hesitancy was observed predominantly in the north-central, east-central, and southeast coastal areas, with some very small clusters where all sampled individuals were hesitant. Conversely, lower vaccine hesitancy was concentrated in the northwest and south-central regions. This geographical variation underscores the need for location-specific strategies to reduce vaccine hesitancy. Factors contributing to these patterns may include social selection, where individuals with similar sociodemographic traits cluster together, and social influence, where vaccine attitudes spread within communities34. Targeted interventions in high-hesitancy areas could involve localized education campaigns and community engagement efforts to foster vaccine acceptance.Our analysis revealed several key findings on COVID-19 vaccine hesitancy in the DR. Sociodemographic factors, particularly older age and mulatto ethnicity, were associated with lower COVID-19 vaccine hesitancy. Conversely, younger people, appeared to exhibit higher hesitancy, potentially influenced by negative social media narratives or a belief in their own resilience35. Individuals born outside the DR also exhibited higher COVID-19 vaccine hesitancy, suggesting that this group might not fully trust the country’s health system36. Mistrust may stem from negative past experiences with healthcare, language barriers, and cultural differences, leading to poor communication with providers37,38. Limited access to reliable vaccine information and opportunities, potentially due to systemic inequities, further exacerbates the issue39. Solutions could include community outreach, culturally sensitive education, and policies to reduce access barriers for immigrants.Educational attainment played a crucial role, with lower hesitancy observed among those with secondary or university education compared to those with no formal education. Previous research has shown higher education to be a significant predictor of COVID-19 vaccine acceptance across multiple countries, highlighting the need for tailored public health messaging and interventions that consider diverse educational backgrounds40. Future research should investigate contextual factors to deepen our understanding of the influences on vaccine hesitancy in the Dominican Republic and other countries in the region.Consistent with the HBM, health-related beliefs appear to be significantly associated with COVID-19 vaccine hesitancy. Participants who perceived vaccination as crucial for personal health were less likely to be hesitant, whereas those with concerns about serious adverse effects of the COVID-19 vaccine exhibited higher hesitancy. This highlights the importance of addressing misconceptions and fears about vaccine safety through effective public health communication. Previous research has shown that vaccinated individuals often expressed strongly positive views regarding the safety and effectiveness of COVID-19 vaccines and considered community benefits, whereas unvaccinated individuals often expressed neutral or negative views on vaccine safety and effectiveness40.The role of information sources and trust also appeared to be contextually important correlates of vaccine hesitancy. Reliance on health information from social media was associated with higher COVID-19 vaccine hesitancy, which aligns with previous research41. Our analysis revealed that trust in local government and health professionals was associated with lower vaccine hesitancy, while higher trust in religious leaders and the media correlated with increased hesitancy. These findings are consistent with the only previous study conducted in the DR25, which also utilized a portion of this dataset. However, even with the introduction of additional variables in our study—spanning individual, household, and community levels—trust remained a significant predictor of vaccine hesitancy. This reinforces the critical role of trust in shaping public health outcomes, particularly in the context of the DR. The robustness of these trust-related factors across different analytical models highlights their importance, suggesting that effective public health campaigns must prioritize building and maintaining trust in reliable sources while actively countering misinformation from less trustworthy channels. A randomized controlled trial conducted in the United States found that even a brief exposure to an infographic explaining a scientific process can enhance trust in science, potentially influencing individuals’ willingness to adopt COVID-19 preventive measures42. Furthermore, engaging religious health leaders to address their concerns about vaccines could help reduce vaccine hesitancy, both among the leaders themselves and the people who trust them.A significant strength of this study is the application of the HBM model as the theoretical framework, which guided the selection of variables most relevant to understanding vaccine hesitancy. By focusing on perceptions of vaccine side effects and beliefs about the severity and susceptibility to COVID-19—factors often overlooked in prior research in the DR—we provide a more nuanced understanding of vaccine hesitancy. Additionally, our inclusion of community-level factors, such as motorized travel time to the nearest health facility and the number of hospitals per community household, adds an important layer of analysis that has been largely absent in earlier studies. This multi-level approach, combined with the broader range of individual and household-level variables, offers a more comprehensive view of the factors influencing vaccine hesitancy.LimitationsThis study has several limitations. Eligible individuals who refused to participate may have had higher levels of COVID-19 vaccine hesitancy than study participants. Moreover, the data were collected through self-reporting, which may have introduced social desirability bias. While based in existing theory, the scale used to measure COVID-19 vaccine hesitancy was not a validated tool, as it was developed in response to the rapidly changing circumstances of the pandemic and evolving vaccine availability. We considered individuals who had already received the COVID-19 vaccine as non-hesitant. This operational definition may overlook cases where individuals harbor vaccine hesitancy yet got vaccinated due to external pressures, such as employment or travel requirements. Although the Dominican Republic did not implement a universal mandate for COVID-19 vaccination, specific circumstances might have influenced some to vaccinate despite personal reservations. As a result, our findings should be interpreted with caution, acknowledging that vaccination behaviour may not always perfectly reflect underlying attitudes.Another limitation involves the use of different geographical scales for data extraction at the community level; for example, illiteracy rates were obtained at the municipality level, while poverty levels were derived from a raster dataset intersected with household locations, which may introduce variability in our community-level findings. Furthermore, in the extracted raster data of Poverty, Development Threat Index, and Motorized Travel Time to the nearest healthcare facility, some null values were identified in the exact household locations. To estimate these null values, we utilized the average value of the four nearest pixels with non-null values.Additionally, the definition and reporting of ethnicity in the DR can vary due to differences in how individuals self-identify and how ethnicity is categorized in different datasets. While this variability could introduce some degree of uncertainty, we believe that including ethnicity in our model still provides valuable insights. The potential inconsistencies are unlikely to significantly affect the overall findings, but future research could benefit from more standardized and reliable measures of ethnicity to further enhance the accuracy of such analyses.ConclusionThis study highlights the complex and multifaceted landscape of COVID-19 vaccine hesitancy among adults in the DR, identifying significant associations with demographic, socio-economic, health-related, and informational factors. The findings underscore the critical need for targeted public health strategies such as enhancing vaccine education, fostering trust in healthcare providers, and addressing socio-economic disparities. The observed spatial clustering and community level factors emphasizes the importance of localized interventions tailored to regional contexts. Our study suggests that future efforts should focus on these areas to effectively address and reduce COVID-19 vaccine hesitancy in DR and serve as a model for other countries in the region or with similar local context. This study can also inform future public health interventions for pandemic response, beyond COVID-19.
Data availability
The data presented in this study are available on request from the corresponding author. The data are not publicly available due to preserving participants’ privacy.
ReferencesFontanet, A. & Cauchemez, S. COVID-19 herd immunity: where are we? Nat. Rev. Immunol. 20 (10), 583–584 (2020).PubMed
PubMed Central
Google Scholar
Limbu, Y. B., Gautam, R. K. & Pham, L. The health belief model applied to COVID-19 vaccine hesitancy: A systematic review. Vaccines (Basel)10(6), 973. https://doi.org/10.3390/vaccines10060973 (2022).de Albuquerque Veloso Machado, M., Roberts, B., Wong, B. L. H., van Kessel, R. & Mossialos, E. The relationship between the COVID-19 pandemic and vaccine hesitancy: A scoping review of literature until August 2021. Front. Public. Health. 9, 747787 (2021).PubMed
PubMed Central
Google Scholar
MacDonald, N. E. Vaccine hesitancy: definition, scope and determinants. Vaccine 33 (34), 4161–4164 (2015).PubMed
MATH
Google Scholar
Michie, S. Encouraging vaccine uptake: lessons from behavioural science. Nat. Rev. Immunol. 22 (9), 527–528 (2022).CAS
PubMed
PubMed Central
MATH
Google Scholar
Skafle, I., Nordahl-Hansen, A., Quintana, D. S., Wynn, R. & Gabarron, E. Misinformation about COVID-19 vaccines on social media: rapid review. J. Med. Internet Res. 24 (8), e37367 (2022).PubMed
PubMed Central
Google Scholar
Piltch-Loeb, R. et al. Determinants of the COVID-19 vaccine hesitancy spectrum. PLoS One. 17 (6), e0267734 (2022).CAS
PubMed
PubMed Central
MATH
Google Scholar
Aw, J., Seng, J. J. B., Seah, S. S. Y. & Low, L. L. COVID-19 vaccine Hesitancy-A scoping review of literature in High-Income countries. Vaccines (Basel) 9(8), 900. https://doi.org/10.3390/vaccines9080900 (2021).Mollalo, A. & Tatar, M. Spatial modeling of COVID-19 vaccine hesitancy in the united States. Int. J. Environ. Res. Public Health. 18 (18), 9488 (2021).CAS
PubMed
PubMed Central
MATH
Google Scholar
Bucyibaruta, G., Blangiardo, M. & Konstantinoudis, G. Community-level characteristics of COVID-19 vaccine hesitancy in England: A nationwide cross-sectional study. Eur. J. Epidemiol. 37 (10), 1071–1081 (2022).PubMed
PubMed Central
Google Scholar
Wiysonge, C. S. et al. Vaccine hesitancy in the era of COVID-19: could lessons from the past help in divining the future? Hum. Vaccines Immunotherapeutics. 18 (1), 1–3 (2022).CAS
Google Scholar
Anuar, H., Shah, S., Gafor, H., Mahmood, M. & Ghazi, H. F. Usage of health belief model (HBM) in health behavior: A systematic review. Malaysian J. Med. Health Sci. 16 (11), 2636–9346 (2020).
Google Scholar
Conner, M. & Armitage, C. J. Extending the theory of planned behavior: A review and avenues for further research. J. Appl. Soc. Psychol. 28 (15), 1429–1464 (1998).MATH
Google Scholar
Kolff, C. A., Scott, V. P. & Stockwell, M. S. The use of technology to promote vaccination: A social ecological model based framework. Hum. Vaccines Immunotherapeutics. 14 (7), 1636–1646 (2018).MATH
Google Scholar
Thomson, A., Robinson, K. & Vallée-Tourangeau, G. The 5As: A practical taxonomy for the determinants of vaccine uptake. Vaccine 34 (8), 1018–1024 (2016).PubMed
MATH
Google Scholar
Mondera, F. et al. Covid-19 vaccination adherence and correlated factors in patients with substance use disorders. Eur. J. Pub. Health 32(Suppl 3), ckac131.351. https://doi.org/10.1093/eurpub/ckac131.351 (2022).Gerretsen, P. et al. Individual determinants of COVID-19 vaccine hesitancy. PloS One. 16 (11), e0258462 (2021).CAS
PubMed
PubMed Central
MATH
Google Scholar
Hossain, M. B. et al. Health belief model, theory of planned behavior, or psychological antecedents: what predicts COVID-19 vaccine hesitancy better among the Bangladeshi adults?? Front. Public. Health. 9, 711066 (2021).PubMed
PubMed Central
Google Scholar
Al-Jayyousi, G. F. et al. Factors influencing public attitudes towards COVID-19 vaccination: A scoping review informed by the Socio-Ecological model. Vaccines (Basel) 9(6), 548. https://doi.org/10.3390/vaccines9060548 (2021).Dutta, T. et al. Perceived enablers and barriers of community engagement for vaccination in India: using socioecological analysis. PLoS One. 16 (6), e0253318 (2021).CAS
PubMed
PubMed Central
Google Scholar
Dutta, T. et al. COVID-19 vaccine behaviors, intentions, and beliefs at a US native American-Serving nontribal institution (NASNTI). BMC Res. Notes. 16 (1), 175 (2023).PubMed
PubMed Central
MATH
Google Scholar
Troiano, G. & Nardi, A. Vaccine hesitancy in the era of COVID-19. Public. Health. 194, 245–251 (2021).CAS
PubMed
MATH
Google Scholar
World Bank national accounts data aONAdf. GNI per capita, Atlas method (current US$) - Dominican Republic: World Bank Group; 2023 [cited 2025. Available from: https://data.worldbank.org/indicator/NY.GNP.PCAP.CD?locations=DOLópez-Sierra, H. E. Cultural diversity and spiritual/religious health care of patients with cancer at the Dominican Republic. Asia Pac. J. Oncol. Nurs. 6 (2), 130–136 (2019).PubMed
PubMed Central
MATH
Google Scholar
Garnier, S. et al. Trust and COVID-19 vaccine hesitancy in the Dominican Republic: a National cross-sectional household survey, June-October 2021. BMJ Open. 14 (5), e081523 (2024).PubMed
PubMed Central
MATH
Google Scholar
Mario Martin, B. et al. Using regional Sero-Epidemiology SARS-CoV-2 Anti-S antibodies in the Dominican Republic to inform targeted public health response. Trop. Med. Infect. Dis. 8(11), 493. https://doi.org/10.3390/tropicalmed8110493 (2023).Morales, D. & Rodríguez, C. Migration in the Dominican Republic: Context, Challenges and Opportunities. (2022).Wang, J-F., Stein, A., Gao, B-B. & Ge, Y. A review of Spatial sampling. Spat. Stat. 2, 1–14 (2012).MATH
Google Scholar
DIGGLE, P. Bayesian Geostatistical design. Scand. J. Stat. 33 (1), 53–64 (2006).MathSciNet
MATH
Google Scholar
Nilles, E. J. et al. SARS-CoV-2 Seroprevalence, cumulative infections, and immunity to symptomatic infection – A multistage National household survey and modelling study, Dominican Republic, June–October 2021. Lancet Reg. Health - Americas. 16, 100390 (2022).PubMed
MATH
Google Scholar
Nilles, E. J. et al. Integrated SARS-CoV-2 serological and virological screening across an acute fever surveillance platform to monitor temporal changes in anti-spike antibody levels and risk of infection during sequential waves of variant transmission — Dominican Republic, March 2021 to August 2022. medRxiv. 2022:10.24.22281399. (2022).Weiss, D. J. et al. Global maps of travel time to healthcare facilities. Nat. Med. 26 (12), 1835–1838 (2020).CAS
PubMed
MATH
Google Scholar
Kiani, B., Fatima, M., Amin, N. H. & Hesami, A. Comparing Geospatial clustering methods to study Spatial patterns of lung cancer rates in urban areas: A case study in Mashhad. Iran. Geoj. 88 (2), 1659–1669 (2023).
Google Scholar
Alvarez-Zuzek, L. G., Zipfel, C. M. & Bansal, S. Spatial clustering in vaccination hesitancy: the role of social influence and social selection. PLoS Comput. Biol. 18 (10), e1010437 (2022).ADS
CAS
PubMed
PubMed Central
MATH
Google Scholar
Hudson, A. & Montelpare, W. J. Predictors of vaccine hesitancy: implications for COVID-19 public health messaging. Int. J. Environ. Res. Public Health. 18 (15), 8054 (2021).CAS
PubMed
PubMed Central
MATH
Google Scholar
Razai, M. S., Osama, T., McKechnie, D. G. J. & Majeed, A. Covid-19 vaccine hesitancy among ethnic minority groups. BMJ 372, n513 (2021).PubMed
Google Scholar
Paul, E., Fancourt, D. & Razai, M. Racial discrimination, low trust in the health system and COVID-19 vaccine uptake: a longitudinal observational study of 633 UK adults from ethnic minority groups. J. R. Soc. Med. 115 (11), 439–447 (2022).PubMed
PubMed Central
Google Scholar
Sapienza, A. & Falcone, R. The role of trust in COVID-19 vaccine acceptance: considerations from a systematic review. Int. J. Environ. Res. Public Health. 20 (1), 665 (2023).MATH
Google Scholar
Demeke, J. et al. Strategies that promote equity in COVID-19 vaccine uptake for undocumented immigrants: A review. J. Community Health. 47 (3), 554–562 (2022).PubMed
PubMed Central
MATH
Google Scholar
Bennett, M. M. et al. Attitudes and personal beliefs about the COVID-19 vaccine among people with COVID-19: a mixed-methods analysis. BMC Public. Health. 22 (1), 1936 (2022).PubMed
PubMed Central
MATH
Google Scholar
Wilson, S. L. & Wiysonge, C. Social media and vaccine hesitancy. BMJ Global Health. 5 (10), e004206 (2020).PubMed
PubMed Central
Google Scholar
Agley, J., Xiao, Y., Thompson, E. E., Chen, X. & Golzarri-Arroyo, L. Intervening on trust in science to reduce belief in COVID-19 misinformation and increase COVID-19 preventive behavioral intentions: randomized controlled trial. J. Med. Internet Res. 23 (10), e32425 (2021).PubMed
PubMed Central
Google Scholar
Download referencesAcknowledgementsWe would like to thank the many study participants that volunteered to participate in this study. We would also like to thank the study staff that collected the field data, the Dominican Republic Ministry of Health and Social Assistance, and the Pedro Henriquez Ureña National University, for their commitment and support for the study.FundingThis study did not receive any fundings. However, CLL was supported by an Australian National Health and Medical Research Council Fellowship (Grant number 1193826). The funder had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.Author informationAuthors and AffiliationsUniversity of Queensland Centre for Clinical Research (UQCCR), Faculty of Health, Medicine, and Behavioural Sciences, The University of Queensland, Brisbane, QLD 4029, AustraliaBehzad Kiani, Benn Sartorius, Beatris Mario Martin, Helen J. Mayfield & Colleen L. LauSchool of Public Health, Faculty of Health, Medicine, and Behavioural Sciences, The University of Queensland, Brisbane, AustraliaAngela Cadavid RestrepoMinistry of Health and Social Assistance, Santo Domingo, Dominican RepublicCecilia Then Paulino, Ronald Skews Ramm, Farah Peña, Bernarda Henríquez & Lucia de la CruzBrigham and Women’s Hospital, Boston, MA, USAPetr Jarolim, Micheal De St Aubin, Devan Dumas, Salome Garnier, Marie Caroline Etienne, Gabriela Abdalla, Margaret Baldwin & Eric J. NillesHarvard Medical School, Boston, MA, USAPetr Jarolim & Eric J. NillesInfectious Diseases and Epidemics Program, Harvard Humanitarian Initiative, Cambridge, MA, USAMicheal De St Aubin, Devan Dumas, Salome Garnier, Margaret Baldwin & Eric J. NillesDepartment of Infectious Disease Epidemiology and Dynamics, Faculty of Epidemiology and Population Health, London School of Hygiene & Tropical Medicine, London, UKAdam KucharskiFaculty of Health Sciences, Pedro Henriquez Urena National University, Santo Domingo, Dominican RepublicWilliam DukeAuthorsBehzad KianiView author publicationsYou can also search for this author inPubMed Google ScholarBenn SartoriusView author publicationsYou can also search for this author inPubMed Google ScholarBeatris Mario MartinView author publicationsYou can also search for this author inPubMed Google ScholarAngela Cadavid RestrepoView author publicationsYou can also search for this author inPubMed Google ScholarHelen J. MayfieldView author publicationsYou can also search for this author inPubMed Google ScholarCecilia Then PaulinoView author publicationsYou can also search for this author inPubMed Google ScholarPetr JarolimView author publicationsYou can also search for this author inPubMed Google ScholarMicheal De St AubinView author publicationsYou can also search for this author inPubMed Google ScholarRonald Skews RammView author publicationsYou can also search for this author inPubMed Google ScholarDevan DumasView author publicationsYou can also search for this author inPubMed Google ScholarSalome GarnierView author publicationsYou can also search for this author inPubMed Google ScholarMarie Caroline EtienneView author publicationsYou can also search for this author inPubMed Google ScholarFarah PeñaView author publicationsYou can also search for this author inPubMed Google ScholarGabriela AbdallaView author publicationsYou can also search for this author inPubMed Google ScholarAdam KucharskiView author publicationsYou can also search for this author inPubMed Google ScholarWilliam DukeView author publicationsYou can also search for this author inPubMed Google ScholarMargaret BaldwinView author publicationsYou can also search for this author inPubMed Google ScholarBernarda HenríquezView author publicationsYou can also search for this author inPubMed Google ScholarLucia de la CruzView author publicationsYou can also search for this author inPubMed Google ScholarEric J. NillesView author publicationsYou can also search for this author inPubMed Google ScholarColleen L. LauView author publicationsYou can also search for this author inPubMed Google ScholarContributionsB.K. analyzed the data, wrote the first draft of the manuscript, reviewed the paper, and led the project. C.L.L. supervised the study. B.S. contributed to the data analysis. A.C.R., H.J.M., C.T.P., P.J., M.D.S.A., R.S.R., D.D., S.G., M.C.E., F.P., G.A., A.K., W.D., M.B., B.H., L.D.C., and E.J.N. contributed to data collection, analysis, interpretation, or manuscript preparation. All authors reviewed and approved the final version of the manuscript.Corresponding authorCorrespondence to
Behzad Kiani.Ethics declarations
Competing interests
The authors declare no competing interests.
Ethical approval
The study was conducted in accordance with the Declaration of Helsinki and approved by the National Council of Bioethics in Health, Santo Domingo (013-2019), the Institutional Review Board of Pedro Henríquez Ureña National University, Santo Domingo, the Mass General Brigham Human Research Committee, Boston, USA (2019P000094) and the Research Ethics and Integrity from the University of Queensland (2022/HE001475).
Informed consent
Written informed consent was obtained from all subjects involved in the study.
Additional informationPublisher’s noteSpringer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.Electronic supplementary materialBelow is the link to the electronic supplementary material.Supplementary Material 1Rights and permissions
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.
Reprints and permissionsAbout this articleCite this articleKiani, B., Sartorius, B., Martin, B.M. et al. Spatial multilevel analysis of individual, household, and community factors associated with COVID-19 vaccine hesitancy in the Dominican Republic.
Sci Rep 15, 11203 (2025). https://doi.org/10.1038/s41598-025-94653-3Download citationReceived: 17 December 2024Accepted: 17 March 2025Published: 02 April 2025DOI: https://doi.org/10.1038/s41598-025-94653-3Share this articleAnyone you share the following link with will be able to read this content:Get shareable linkSorry, a shareable link is not currently available for this article.Copy to clipboard
Provided by the Springer Nature SharedIt content-sharing initiative
KeywordsCOVID-19Vaccine hesitancyVaccine acceptanceHealth belief modelSocioeconomicHousehold dynamicsThe Dominican Republic