A Comparative Analysis on the Social Determinants of COVID-19 Vaccination Coverage in Fragile and Conflict Affected Settings and Non-fragile and Conflict Affected Settings

The COVID-19 pandemic has strained health and economic systems worldwide, with fragile and conflict-affected settings (FCS) facing compounded challenges such as disrupted supply chains, insecurity, weak infrastructure, and limited government effectiveness. Vaccine coverage has varied markedly between FCS and non-FCS countries due to differences in access, delivery capacity, and hesitancy. Guided by the WHO’s Commission on Social Determinants of Health (CSDH) framework, this study examines how structural social, economic, political, and health system factors shape COVID-19 vaccination coverage across countries. Research question: What are the significant social determinants of COVID-19 vaccination coverage in FCS and non-FCS countries? The study’s purpose is to identify and compare macro-level determinants to inform policy and resource allocation to improve vaccine equity, particularly in FCS.

The study is grounded in the WHO’s Social Determinants of Health and the CSDH conceptual framework, which posit that non-medical factors (eg, socioeconomic position, governance, political context) account for a substantial share of health outcomes. Prior literature links vaccine inequity to economic factors correlated with poverty, literacy, and demographics, and highlights interactions between economic development, governance, and democratic institutions. Evidence on vaccine hesitancy underscores its role as a demand-side factor influenced by trust, transparency, religion, and perceived risk. The study adapts the CSDH framework to COVID-19 vaccination inequity by grouping determinants into socioeconomic (eg, GDP per capita, a composite socioeconomic resilience index), political and governance (eg, government effectiveness, stability, equality before law, regional government power, exclusion of groups), and health system factors (eg, human resources for health, health spending, overall system strength), and incorporates global vaccine confidence data as a proxy for demand-side behavior. This literature foundation supports analyzing both separate and combined effects of these dimensions on vaccine coverage in FCS and non-FCS contexts.

Design: Cross-country comparative analysis using multivariate log-linear regression models to assess associations between macro-level determinants and COVID-19 vaccination coverage, estimated separately for non-FCS and FCS country samples. Data sources: World Bank classification of Fragile and Conflict-Affected Situations (FY2022) to define FCS (39 countries; cumulative population ~930 million). COVID-19 epidemiologic and vaccination data from Our World in Data (OWID; ~223 countries). Additional determinants from WHO Global Health Observatory, Global Health Security (GHS) Index, United Nations, World Bank World Development Indicators, Worldwide Governance Indicators (WGI), and the Varieties of Democracy (V-Dem) datasets. Vaccine confidence (general, pre-COVID) from de Figueiredo et al. Outcome variable: Share of population receiving at least one COVID-19 vaccine dose by end of July 2021; log-transformed for regression. Sensitivity analyses using alternative vaccination coverage definitions showed robustness. Independent variables: - Socioeconomic: GDP per capita (PPP), composite socioeconomic resilience index (adult literacy, UNDP Gender Inequality Index, extreme poverty rates at PPP $1.90/day, public confidence in government, strength of domestic media), demographics (population size, density, urban share, share aged 65+). - Health system: GHS Index indicators; WHO data on hospital beds per 1,000, medical doctors per 10,000, nurses and midwives per 10,000; domestic government health expenditure (% of GDP) and per capita PPP. - Political/governance: WGI indicators (government effectiveness, political stability/absence of violence, regulatory quality, rule of law, control of corruption, voice and accountability); select V-Dem indices (equality before the law and individual liberty, regional government power, measures of exclusion of groups). - Demand-side: Share strongly agreeing vaccines are effective (global vaccine confidence proxy). Samples and estimation: Two samples—non-FCS (n up to ~184; varying by data completeness) and FCS (n=39; regressions often based on smaller n due to missingness). Multivariate log-linear regressions estimated in STATA 16.1 with 5% significance threshold. Model diagnostics included tests for omitted variable bias, linearity, multicollinearity, and heteroskedasticity; heteroskedasticity-robust standard errors used. Aggregate models combined socioeconomic, political, and health system determinants. Supplementary files provide detailed variable lists, robustness checks, and sensitivity analyses.

- Descriptive differences (Table 1): Non-FCS vs FCS means: at least one dose coverage 38.33% vs 5.46%; fully vaccinated 28.30% vs 2.90%; average CFR 2.49% vs 3.02%. FCS have substantially lower health system capacity (eg, doctors per 10,000: 6.33 vs 22.80; domestic health spending per capita PPP: 105.03 vs 1198.41) and lower governance scores (eg, government effectiveness −1.20 vs 0.27). - Socioeconomic determinants: Economic and demographic blocks explain 66% of variance in non-FCS and 48% in FCS vaccine coverage. A 1% increase in GDP per capita is associated with a 0.64% increase in coverage in non-FCS and 0.9% in FCS. A 1-unit increase in socioeconomic resilience is associated with a 2.7% increase (non-FCS) and 2.1% increase (FCS). Demographic variables (density, 65+ share, urbanization) are not strong or consistent predictors; population size positively associated in non-FCS but negatively in FCS when considered alone, and loses significance alongside economic variables. - Health system determinants: Health system indicators explain R² ≈ 0.43 (non-FCS) and 0.48 (FCS). In non-FCS, each additional doctor per 10,000 is associated with a 1.7% increase in coverage; each additional nurse/midwife per 10,000 with a 0.04% increase. In FCS, doctor density is negatively associated while nurse/midwife density is positively associated. A 1 percentage point increase in domestic government health expenditure (% of GDP) is associated with a 13% increase (non-FCS) and 8.3% increase (FCS) in coverage. - Political and governance determinants: Greater government effectiveness is positively associated with coverage and appears statistically significant in FCS, indicating strong implementation capacity matters more where conflict/fragility exists. Political stability and absence of violence, and more power to regional governments show positive associations, particularly in FCS by coefficient magnitude. After controlling for other factors, equality before the law and individual liberty index shows a negative association with coverage (stronger and statistically significant in FCS), indicating complex structural dynamics that warrant further investigation. Exclusion of socioeconomic groups from public spaces shows mixed associations, underscoring context-specific effects. - Aggregate model (Table 2): Combined determinants explain R² ≈ 0.68 (non-FCS; 103 observations) and 0.63 (FCS; 19 observations). GDP per capita and socioeconomic resilience remain the strongest predictors overall (significant in non-FCS). Government effectiveness, regional government power, and political stability show larger positive coefficients in FCS than non-FCS, though significance varies with small samples. Health workforce variables show inconclusive effects when combined with other determinants. Vaccine confidence proxy: in non-FCS, higher shares strongly believing vaccines are effective associate with lower COVID-19 coverage; in FCS, the association is positive. - Overall: Socioeconomic capacity (income, resilience), governance effectiveness, political stability/violence absence, health financing, and specific HRH densities are key correlates of vaccine coverage, with patterns differing between FCS and non-FCS.

The study addresses the research question by demonstrating that structural social determinants—economic capacity, governance quality, and health system financing and workforce—are central to explaining cross-country disparities in COVID-19 vaccination coverage, with distinct patterns between FCS and non-FCS contexts. Economic strength (GDP per capita) and broader socioeconomic resilience most consistently predict higher coverage, reflecting greater fiscal space, literacy, lower poverty, and stronger institutions. In FCS, effective government machinery and political stability appear particularly salient, suggesting that implementation capacity under conditions of conflict or fragility is critical for vaccine rollout. Health financing shows robust positive associations in both groups, emphasizing the role of domestic investment in enabling delivery. The counterintuitive negative association of equality before the law and individual liberty indices after adjustment, and the divergent associations of vaccine confidence between FCS and non-FCS, indicate complex interactions among structural, societal, and behavioral factors. These findings highlight the need to contextualize demand-side interventions within broader governance and conflict dynamics. The results reinforce global policy calls for equitable access and underscore that success depends not only on supply but also on governance capacity, stability, and social conditions—especially in FCS—thereby informing targeted strategies for improving vaccine equity.

This study contributes comparative, cross-country evidence that socioeconomic capacity (GDP per capita, socioeconomic resilience), governance effectiveness, political stability, and health financing are pivotal determinants of COVID-19 vaccination coverage, with distinct effect patterns in fragile and conflict-affected settings. It confirms the critical roles of government effectiveness and stability in FCS and underscores that enhancing domestic health spending and appropriate health workforce configurations can improve coverage. Future research should probe the unexpected negative adjusted association of equality before the law and individual liberty with coverage, unpack the mechanisms of group exclusion and regional power structures, and refine demand-side measures using COVID-19-specific hesitancy indicators. Policy efforts should prioritize strengthening governance and implementation capacity in FCS, increasing health financing, and developing context-sensitive strategies to address vaccine hesitancy and access barriers to mitigate vaccine inequity.

- Ecological design using country-level averages limits causal inference and may mask within-country heterogeneity. - Data merged from multiple public sources with variable completeness; small FCS sample size (n≈39; fewer with complete data) reduces statistical power and precision. - Missing observations in both FCS and non-FCS may bias estimates; sample adequacy for regressions is limited for some models. - Outcome measured at a single time point (end of July 2021); coefficients are sensitive to the research date and additional data. - Vaccine confidence proxy reflects general vaccine attitudes, not COVID-19-specific perceptions, potentially attenuating or distorting demand-side associations. - Potential unobserved confounding despite tests for omitted variable bias and other diagnostics.