Northern–Southern Inequalities in COVID-19 Vaccination Uptake Among Women of Reproductive Age in Nigeria: A Multivariate Decomposition Analysis Authors: Michael Ekholuenetale, Oluwakemi Christie Ogidan, Victor A Ochagu, Amadou Barrow Published: 2026 ABSTRACT AbstractBackgroundThe vaccination of women against COVID-19 remains a priority in the post-pandemic era. Regional inequalities in COVID-19 vaccination uptake among Nigerian women are a key factor of interest in public health research. The objective of this study was to examine the northern–southern inequality in COVID-19 vaccination among Nigerian women of reproductive age.MethodsWe analysed a nationally representative sample of 36,555 women aged 15–49 years from the 2024 Nigeria Demographic and Health Survey. Nigeria’s six geopolitical zones were grouped into a northern region (North Central, North East, North West) and a southern region (South East, South, South West). Multivariate decomposition analysis for non-linear models was used to partition the northern–southern difference in COVID-19 vaccination into an endowment (compositional) component and a coefficient (unexplained) component.ResultsCOVID-19 vaccination was reported by 29.0% (95% CI: 27.2–33.0) of southern women and 24.1% (95% CI: 21.9–26.4) of northern women (p<0.05). Higher percentages of vaccination were observed among women aged 45–49 years and among those with higher education, health insurance coverage, employment, classification as empowered, the lowest level of socioeconomic disadvantage, media exposure, residence in the richest households, residence in communities with high willingness for vaccine uptake, and urban residence. In the decomposition, the endowment (compositional) component accounted for −34.5% and the coefficient (unexplained) component for 134.5% of the northern–southern difference. Importantly, the negative compositional contribution indicates that differences in the measured characteristics of women between regions did not explain the northern disadvantage. Rather, the gap was predominantly associated with regional differences in the effects of these characteristics. Thus, under a counterfactual scenario in which northern women had the same observed characteristics as southern women, the vaccination gap would not necessarily narrow and could instead widen in favour of the south, suggesting an important role for contextual factors not captured in the model. Educational attainment, religion, employment, media exposure, community-level vaccine willingness, community-level illiteracy, community-level nativity, and community-level urbanicity were the categories with the largest contributions.ConclusionCOVID-19 vaccination uptake among women of reproductive age was lower in northern than southern Nigeria. The decomposition indicates that the regional gap was driven predominantly by the coefficient component (134.5%) rather than the compositional component (−34.5%), suggesting that differences in how individual, socioeconomic, and community characteristics operate across regions are more important than differences in population composition alone. Therefore, equalizing measured characteristics between regions may not eliminate the disparity; region-specific interventions should address the contextual, sociocultural, behavioural, and health-system factors shaping vaccination decisions.Keywords: Nigeria, COVID-19, women’s health, infectious diseases, vaccination Introduction Since its emergence in late 2019, COVID-19 caused millions of deaths, overwhelmed health systems, disrupted routine healthcare services, and exposed longstanding inequalities in healthcare access and public health preparedness.1 Although the WHO officially declared the end of the COVID-19 Public Health Emergency of International Concern (PHEIC) in May 2023, the pandemic left enduring lessons regarding disease surveillance, risk communication, vaccination strategies, health system resilience, and preparedness for future public health emergencies.2 Among these lessons, equitable access to vaccination stood out as central to epidemic and pandemic control. Vaccination emerged as the most effective public health intervention for reducing COVID-19-related morbidity, mortality, hospitalizations, and transmission. Globally, COVID-19 vaccines substantially reduced severe disease outcomes and facilitated the gradual reopening of economies and social systems.3,4 However, vaccine distribution and uptake were characterized by significant inequities both between and within countries. Low- and middle-income countries experienced delayed access to vaccines, while socioeconomic, demographic, cultural, and geographical disparities influenced vaccine acceptance and uptake among populations.5 These inequalities not only undermined pandemic control efforts but also highlighted broader structural determinants of health that shape access to healthcare services during public health crises. Nigeria, the most populous country in Africa, experienced substantial challenges in its COVID-19 response. Although national vaccination campaigns achieved notable progress, vaccine uptake remained below optimal levels and varied considerably across population groups and geographical regions. Nigeria’s national COVID-19 vaccination programme, coordinated by the National Primary Health Care Development Agency (NPHCDA), aimed to fully vaccinate 70% of the eligible population (approximately 111.8 million people) by the end of 2022.6 By late 2023, administrative reports indicated that about 70% of eligible Nigerians had been fully vaccinated, although household survey estimates of uptake have generally been lower; a national survey conducted in 2022 estimated that only about one-third of respondents had received a COVID-19 vaccine.7 These national benchmarks provide context for interpreting the survey-based uptake estimates reported in this study for women of reproductive age. Studies conducted during and after the pandemic identified vaccine hesitancy, misinformation, concerns regarding vaccine safety, religious influences, low risk perception, and distrust in government institutions as important determinants of vaccine acceptance among Nigerians.8,9 Research further revealed that educational attainment, socioeconomic status, urban residence, access to healthcare services, and exposure to media were important predictors of COVID-19 vaccine uptake.10,11 Women of reproductive age constitute a particularly important population in COVID-19 vaccination research. Beyond their own health needs, women often influence household healthcare decisions, maternal and child health service utilization, and community health behaviours. Furthermore, evidence suggests that women may face unique barriers to vaccine access and acceptance, including concerns regarding fertility, pregnancy outcomes, vaccine safety, caregiving responsibilities, and differential access to health information.12 Ensuring equitable vaccine uptake among women therefore contributes not only to individual protection but also to broader community health resilience and preparedness for future disease outbreaks. Nigeria is characterized by substantial regional heterogeneity in socioeconomic development, educational attainment, healthcare access, cultural norms, and health outcomes. Historically, northern and southern regions have exhibited marked disparities in maternal health indicators, childhood immunization coverage, educational achievement, poverty levels, and utilization of healthcare services.13 These structural inequalities may similarly influence COVID-19 vaccination uptake. Emerging evidence suggests that regional differences in vaccine acceptance and coverage exist across Nigeria; however, limited research has systematically quantified the extent to which observed disparities are attributable to differences in population characteristics versus differences in how those characteristics influence vaccination behaviour.14 For this analysis, the six geopolitical zones were grouped into northern (north-central, north-east, north-west) and southern (south-west, south-east, south-south) regions to examine the broad regional inequality in vaccination uptake. The six geopolitical zones were collapsed into a binary northern–southern classification to improve statistical power and interpretability while capturing the broad and well-established regional socioeconomic, cultural, health-system, and epidemiological differences in Nigeria. This approach reduces model complexity and avoids sparse estimates that may arise from comparisons across six categories, particularly where some zones have smaller sample sizes. The North–South classification was therefore considered more appropriate for estimating an overall regional gradient rather than zone-specific differences.15,16 This classification reflects longstanding and well-documented differences between northern and southern Nigeria in socioeconomic development, educational attainment, health-system access, cultural and social contexts, and utilization of health services. The binary regional comparison was therefore intended to capture the overarching structural and contextual divide relevant to vaccination inequities rather than to imply homogeneity within either region. Although previous studies have examined determinants of COVID-19 vaccine acceptance and uptake in Nigeria, nationally representative evidence on northern–southern inequalities among women of reproductive age remains scarce. Furthermore, the pathways through which individual-, household-, and community-level factors contribute to regional differences in vaccine uptake have not been comprehensively investigated. Addressing this gap is important for informing equitable vaccination policies and strengthening preparedness for future public health emergencies. Therefore, this study aimed to examine the prevalence and determinants of COVID-19 vaccination uptake among women of reproductive age in Nigeria and to quantify the contributions of individual, household, and community-level factors to northern–southern inequalities in vaccination uptake. In contrast to earlier decomposition analyses of maternal and child health outcomes in Nigeria, the present study applies multivariate decomposition specifically to COVID-19 vaccination among women of reproductive age, using the most recent nationally representative data. Methods Data Source A cross-sectional study was conducted for 36,555 women of reproductive age (15–49 years) using the 2024 Nigeria Demographic and Health Survey (NDHS) data. The National Population Commission (NPC), acted on behalf of the Federal Ministry of Health and Social Welfare (FMoHSW) to support the data collection process. Data collection took place between December 1, 2023, and May 7, 2024. The United States Agency for International Development’s (USAID) funds Demographic and Health Surveys Program and offers technical support for population and health surveys across the globe. In addition, ICF, the World Health Organisation (WHO), the United Nations Population Fund (UNFPA), the United Nations Children’s Fund (UNICEF), and the Global Fund to Fight AIDS, Tuberculosis, and Malaria (Global Fund) were among the other organisations and agencies that provided financial or technical assistance to enable the survey’s successful implementation. Sample Design The upgraded cartographical frame created for Nigeria’s anticipated first fully digital Population and Housing Census served as the sample frame for the 2024 NDHS. Nigeria is separated into six administrative zones: North Central, North East, North West, South East, South, and South West. The Federal Capital Territory (FCT) and 36 states make up the zones, which result in 37 subnational entities. Local government areas (LGAs) make up each state, while communities make up each LGA. During the delineation process for the impending census, each locality is further divided into suitable areas known as enumeration areas (EAs). These EAs served as the basis for primary sampling units (PSUs), also known as clusters in the 2024 NDHS. The 2024 NDHS sample was created to provide representative data for all six zones, the 36 states and the Federal Capital Territory, urban and rural areas separately, and the nation as a whole. The 2024 NDHS used a two-stage stratified sample design. By dividing each of the 36 states and the Federal Capital Territory into urban and rural areas, stratification was accomplished. There were 74 sampling strata in all. In each stratum, samples were chosen separately using a two-stage selection process. The initial step was to choose sample locations, or clusters, made up of EAs. Within each sampling stratum, EAs were selected with a probability that was proportionate to their size. There were 701 urban and 699 rural clusters out of the 1400 total that were chosen. A systematic sample of households was used in the second step. All of the chosen clusters underwent a household listing procedure, and a systematic equal probability selection process was used to choose a set number of 30 households per cluster, for a total sample size of about 42,000 households. At the time of listing and during interviews, Global Positioning System (GPS) data were gathered for every household. The report of the sampling design, training of field assistants and data collection process has been published previously.17 Selection and Measurements of Variables Outcome Variable The response variable was COVID-19 vaccination and determined from the question: “S1112P: Received COVID-19 vaccination”. Women responded “yes” or “no”. COVID-19 vaccination was coded “1” if a woman has received the vaccine and “0” if otherwise. COVID-19 vaccination data was self-reported by the women based on their uptake status during and after the global COVID-19 epidemic. Responses were not verified against vaccination cards or facility records, and the 2024 NDHS did not collect a card-verified subsample for COVID-19 vaccination; a verification-based sensitivity analysis was therefore not possible. Explanatory Variables Age in 5-year groups between 15 through 49; education: none, primary, secondary, higher; religion: Christianity, Islam, traditional/no religion; covered by health insurance scheme: yes vs no; marital status: single, currently married, formerly married; employment status: not working vs working; cultural norms about wife-beating was created by aggregating responses from women in each community. Here, we used the items: “beating justified if wife goes out without telling husband”, “beating justified if wife neglects the children”, “beating justified if wife argues with husband”, “beating justified if wife refuses to have sex with husband” and “beating justified if wife burns the food”. A binary variable was created by aggregating responses from women for acceptance of wife beating, where “no” indicates empowered women and “yes” to wife beating justified, indicates not empowered women; socioeconomically disadvantaged status: neighborhood socioeconomic disadvantaged level used items such as rural residence, poorest household wealth, no formal education, and not working. Using principal component analysis (PCA), the standardized z-score was used to disentangle the overall assigned scores to low, medium, or high; exposed to media (read newspaper, listen to radio, watch tv, use internet): yes vs no. In addition, sex of household head: male vs female; household size: 1–2, 3–4, 5–6, 7+; household wealth: wealth indicator weights were assigned using PCA. The wealth indicator variables (bicycle, motorcycle/scooter, car/truck, main floor material, main wall material, main roof material, sanitation facilities, water source, radio, television, electricity, refrigerator, cooking fuel, furniture and number of people per room) were standardised and scores were assigned. Factor loadings, or factor coefficient scores, and z-scores were computed. To determine each household’s wealth index value, the indicator values were multiplied by the loadings and added together. The overall scores were divided into the poorest/poorer/middle/richer/richest groups using the standardised z-score. Since the DHS did not gather aggregate-level data at the community-level, we used EAs to represent communities. Therefore, the analysis’s community-level factors were centred on women’s characteristics, especially those that affect the coverage of COVID-19 vaccinations. Community-level women’s empowerment was created from the individual woman cultural norms about wife-beating justification by aggregating responses from women in each community; hence, community-level empowerment was created as the proportion of women who reported wife beating justified. Similarly, community-level willingness for COVID-19 vaccine uptake, nativity, illiteracy, media use, and poverty were created as the proportion of women who reported willingness to COVID-19 vaccine uptake, had lived 5 years or more in communities, cannot read at all, exposed to media use, and from the two lowest household wealth quintiles in communities respectively. The place of residence was classified as urban vs rural. Geographical region: North west, North east, North central, South east, South, South west. This approach was used in a previous study.18–20 Analytical Approach Stata software, version 17.0, was used to analyse the data (Stata Corporation, College Station, Texas, USA). The study used a multi-stage stratified cluster sample design data, so the survey module’s (“svy”) function was used to account for sampling design (weighting, clustering, and stratification). The percentage and chi-square test were employed in the univariate and bivariate analyses. Multivariate decomposition analysis was used to decompose the percentage difference of COVID-19 vaccination across northern–southern regions. Item non-response for the variables included in this study was minimal. Analyses were based on women with complete information on the outcome and all covariates (complete-case analysis), and no imputation of missing values was performed. Composite indices (the household wealth index and the neighbourhood socioeconomic disadvantage score) were derived from the observed indicator items, and community-level variables were computed by aggregating the responses of all women interviewed in each enumeration area. Multivariate decomposition for the non-linear model was used to quantify the contributions to group differences in average predictions from multivariate models. So, COVID-19 vaccination difference in northern and southern regions was explained by two factors, and it could be due to the composition difference in the population (endowment) or the change in characteristics (coefficient) in the explanatory variables. The northern–southern difference in the average predicted probability of COVID-19 vaccination was decomposed as follows: where and denote the average predicted probabilities of COVID-19 vaccination among southern and northern women, respectively; F(·) is the logistic cumulative distribution function; XS and XN are vectors of the observed characteristics of southern and northern women; and βS and βN are the corresponding vectors of estimated coefficients. The first bracketed term, E, is the endowment (explained or compositional) component: the part of the regional difference attributable to differences in the distribution of women’s characteristics between the regions. The second bracketed term, C, is the coefficient (unexplained) component: the part attributable to regional differences in the effects of those characteristics on vaccination uptake. For each predictor category, a positive contribution is associated with a wider northern–southern difference in vaccination, and a negative contribution is associated with a narrower difference. Within the endowment component, the contribution reflects the part of the difference associated with the regional distribution of that characteristic; within the coefficient component, it reflects the part associated with regional differences in the estimated association between that characteristic and vaccination. The sign and magnitude of each contribution are therefore interpreted as statistical associations rather than as causal effects. Ethical Consideration This study used publicly available secondary dataset with all identifiers deleted. NDHS received ethical approval from National Health Research Ethics Committee of Nigeria (NHREC) and obtained informed consent from the respondents in accordance with standard ethical procedure. The Health Research Committee assigned number: NHREC/01/01/2007. In addition, ICF institutional Review Board gave ethical approval with ICF Project Number: 132989.0.000.NG.DHS.01. However, the corresponding author of this study was granted permission to use the data, which was gathered in accordance with ethical guidelines, therefore no further participants’ consent was required. Information about DHS guidelines can be found here: http://goo.gl/ny8T6X. This study involved secondary analysis of anonymised, publicly available data from the Nigeria Demographic and Health Survey. The primary survey protocol received ethical approval from the National Health Research Ethics Committee of Nigeria (NHREC) and the ICF Institutional Review Board, and informed consent was obtained from survey participants. In accordance with the exemption provisions of the National Code of Health Research Ethics of Nigeria (2007) concerning research involving existing publicly available data, no additional ethical approval was required for this secondary analysis. This practice is consistent with a published scientific report.21 Results Table 1 shows the distribution of women’s characteristics by region. The northern region contributed a larger number of women than the southern region. Compared with the southern region, the northern region had higher proportions of women aged 20–24 years and of women with no formal education, of Muslim faith, without health insurance coverage, currently married or in union, in the high socioeconomic-disadvantage category, in male-headed and poorest households, in households of seven or more members, native to their communities, resident in rural areas, and in communities with high illiteracy, high poverty, and low media use. The corresponding distribution for the southern region is shown in Table 1. Table 1Distribution of the Respondents’ Characteristics (N= 36,555)VariableNorthern Region (n= 21,597)Southern Region (n= 14,958)Age in 5-year groups15-194511 (20.9)2866 (19.2)20-243994 (18.5)2269 (15.2)25-293788 (17.5)2168 (14.5)30-343138 (14.5)2124 (14.2)35-392610 (12.1)2139 (14.3)40-442004 (9.3)1922 (12.9)45-491552 (7.2)1470 (9.8)Highest educational levelNo education10,316 (47.8)520 (3.5)Primary2411 (11.2)1687 (11.3)Secondary6729 (31.2)9295 (62.1)Higher2141 (9.9)3456 (23.1)ReligionChristianity5170 (23.9)13,243 (88.5)Islam16,314 (75.5)1633 (10.9)Traditiona/no religion113 (0.5)82 (0.6)Covered by health insuranceNo20,983 (97.2)14,323 (95.8)Yes614 (2.8)635 (4.2)Marital statusSingle5267 (24.4)5459 (36.5)Currently married15,445 (71.5)8484 (56.7)Formerly married885 (4.1)1015 (6.8)Employment statusNot working10,468 (48.5)4298 (28.7)Working11,129 (51.5)10,660 (71.3)Wife beating justifiedLow15,599 (72.2)13,225 (88.4)High5998 (27.8)1733 (11.6)Socioeconomically disadvantaged statusLow2838 (13.1)9370 (62.6)Medium7089 (32.8)5093 (34.1)High11,670 (54.0)495 (3.3)Exposed to mediaNo10,818 (50.1)1909 (12.8)Yes10,779 (49.9)13,049 (87.2)Sex of household headMale19,005 (88.0)9842 (65.8)Female2592 (12.0)5116 (34.2)Household wealthPoorest5890 (27.3)365 (2.4)Poorer4963 (23.0)1000 (6.7)Middle4550 (21.1)2729 (18.2)Richer3714 (17.2)4716 (31.5)Richest2480 (11.5)6148 (41.1)Household size1-21557 (7.2)2212 (14.8)3-44386 (20.3)4877 (32.6)5-65249 (24.3)4633 (31.0)7+10,405 (48.2)3236 (21.6)Community-level women empowermentLow12,310 (57.0)2790 (18.7)High9287 (43.0)12,168 (81.3)Community-level willingness for COVID-19 vaccine uptakeLow9550 (44.2)8873 (59.3)High12,047 (55.8)6085 (40.7)Community-level nativityLow5348 (24.8)8251 (55.2)High16,249 (75.2)6707 (44.8)Community-level urbanicityHigh7515 (34.8)10,422 (69.7)Low14,082 (65.2)4536 (30.3)Community-level illiteracyLow8234 (38.1)13,566 (90.7)High13,363 (61.9)1392 (9.3)Community-level media useLow13,223 (61.2)1786 (11.9)High8374 (38.8)13,172 (88.1)Community-level povertyLow9139 (42.3)13,567 (90.7)High12,597 (57.7)1391 (9.3) Distribution of the Respondents’ Characteristics (N= 36,555) Figure 1 shows the percentage of COVID-19 vaccination by northern and southern regions. COVID-19 vaccination was reported by 29.0% of southern women and 24.1% of northern women. The regional difference was statistically significant (p<0.05). Percentage of COVID-19 vaccine uptake by northern–southern differences. Weighted percentage of women aged 15–49 years reporting at least one dose of a COVID-19 vaccine, by region of residence, 2024 Nigeria DHS (n = 36,555). Northern region = North Central, North East and North West zones; southern region = South East, South South and South West zones. The horizontal axis is scaled as a proportion (0.1 = 10%). Uptake was 24.1% in the north and 29.0% in the south, a difference of 4.9 percentage points; p value from a design-based Pearson chi-square test. Estimates are weighted and account for the complex survey design. Table 2 presents the estimated percentage of COVID-19 vaccination by northern and southern regions across levels of women’s characteristics. Within most categories, northern women reported a lower percentage of COVID-19 vaccination than southern women. Higher percentages of vaccination were reported among women aged 45–49 years, those with higher education, those covered by health insurance, those who were employed, those classified as empowered, those least socioeconomically disadvantaged, those exposed to media, those from the richest households, those from communities with high willingness for COVID-19 vaccine uptake, and those resident in urban areas. The regional difference was statistically significant for several characteristics (p<0.05). In Table 2, the percentage-point difference (∆%) was computed as the northern percentage minus the southern percentage; negative values therefore indicate higher uptake among southern women. The overall regional gap reported in the abstract and decomposed in Table 3 is expressed in the opposite direction, as the southern–northern difference (29.0% − 24.1% = 4.9 percentage points). Table 2Northern-Southern Percentage-Point Difference (∆%) in COVID-19 VaccinationVariableNorthern (%)Southern (%)Percentage Point Difference (∆% Northern–Southern)P-valueAge in 5-year groups<0.001*15-1917.915.82.120-2422.427.3−4.925-2923.228.0−4.830-3427.531.7−4.235-3927.833.7−5.940-4429.435.9−6.545-4929.938.6−8.7Highest educational level<0.001*No education17.425.2−7.8Primary26.828.0−1.2Secondary28.526.32.2Higher40.137.13.0Religion<0.001*Christianity29.126.82.3Islam22.746.8−24.1Traditional/no religion10.622.0−11.4Covered by health insurance0.021*No23.628.3−4.7Yes45.845.20.6Marital status<0.001*Single22.822.00.8Currently married24.232.9−8.7Formerly married32.033.4−1.4Employment status0.364Not working18.020.5−2.5Working30.032.4−2.4Wife beating justified0.029*Low25.329.1−3.8High21.327.6−6.3Socioeconomically disadvantaged status0.112Low31.631.60.0Medium29.225.53.7High19.315.43.9Exposed to media0.001*No18.922.6−3.7Yes29.529.9−0.4Sex of household head0.002*Male23.628.6−5.0Female28.329.7−1.4Household wealth<0.001*Poorest16.719.5−2.8Poorer22.023.7−1.7Middle28.727.61.1Richer29.630.8−1.2Richest29.829.60.2Household size<0.001*1-224.230.9−6.73-424.230.7−6.55-625.129.4−4.37+23.724.4−0.7Community-level women empowerment0.009*Low21.425.6−4.2High27.929.7−1.8Community-level willingness for COVID-19 vaccine uptake0.196Low10.717.1−6.4High34.946.3−11.4Community-level nativity<0.001*Low24.427.2−2.8High24.131.2−7.1Community-level urbanicity0.269High25.531.0−5.5Low22.424.4−2.0Community-level illiteracy<0.001*Low30.529.41.1High20.325.3−5.0Community-level media use0.876Low21.122.0−0.9High29.129.9−0.8Community-level poverty0.035*Low29.629.60.0High20.222.9−2.7Notes: * significant at P <0.05. ∆% was computed as the percentage of vaccinated women in the northern region minus that in the southern region; negative values indicate higher uptake in the southern region. The overall regional gap reported in the abstract, main text, and Table 3 is expressed in the opposite direction, as the southern–northern difference (29.0% − 24.1% = 4.9 percentage points). Northern-Southern Percentage-Point Difference (∆%) in COVID-19 Vaccination Notes: * significant at P <0.05. ∆% was computed as the percentage of vaccinated women in the northern region minus that in the southern region; negative values indicate higher uptake in the southern region. The overall regional gap reported in the abstract, main text, and Table 3 is expressed in the opposite direction, as the southern–northern difference (29.0% − 24.1% = 4.9 percentage points). Table 3Factors Contributing to Northern-Southern Inequality in COVID-19 VaccinationVariableEndowments (Explained) (E)Coefficients (Unexplained) (C)β (SE)Percentage Contribution (Pct)β (SE)Percentage Contribution (Pct)Overall explained disparity (%)−0.0165 (0.0145)−34.50.0644 (0.0152)134.5Age in 5-year groups15-19RefRef20-24−0.0047 (0.0017)*−9.730.0125 (0.0033)*26.0625-29−0.0045 (0.0017)*−9.410.0105 (0.0035)*21.9730-34−0.0006 (0.0002)*−1.260.0054 (0.0031)11.2935-390.0046 (0.0017)*9.620.0068 (0.0027)*14.1640-440.0084 (0.0030)*17.550.0053 (0.0021)*11.0145-490.0074 (0.0027)*15.510.0067 (0.0018)*14.05Highest educational levelNo educationRefRefPrimary0.0001 (0.0000)0.12−0.0018 (0.0026)−3.83Secondary0.0312 (0.0116)65.13−0.0034 (0.0069)−7.17Higher0.0241 (0.0102)*50.35−0.0028 (0.0025)−5.84ReligionChristianityRefRefIslam−0.0737 (0.0242)*−154.070.0681 (0.0105)*142.43Traditiona/no religion−0.0001 (0.0000)−0.020.0006 (0.0004)1.16Covered by health insuranceNoRefRefYes0.0016 (0.0006)*3.28−0.0002 (0.0006)−0.47Marital statusSingleRefRefCurrently married−0.0009 (0.0022)−1.970.0002 (0.0109)0.49Formerly married−0.0001 (0.0006)−0.06−0.0009 (0.0010)−1.90Employment statusNot workingRefRefWorking0.0101 (0.0043)*21.15−0.0179 (0.0056)*−37.39Wife beating justifiedLowRefRefHigh−0.0045 (0.0029)−9.290.0055 (0.0037)11.41Socioeconomically disadvantaged statusLowRefRefMedium0.0004 (0.0003)0.900.0109 (0.0056)22.57High0.0070 (0.0229)14.61−0.0074 (0.0187)−15.56Exposed to mediaNoRefRefYes0.0069 (0.0067)14.40−0.0168 (0.0070)*−35.06Sex of household headMaleRefRefFemale0.0039 (0.0029)8.100.0006 (0.0015)1.34Household wealthPoorestRefRefPoorer−0.0010 (0.0063)−1.99−0.0028 (0.0068)−5.81Middle−0.0008 (0.0012)−1.74−0.0040 (0.0063)−8.35Richer−0.0087 (0.0068)18.160.0037 (0.0054)7.67Richest0.0060 (0.0123)12.600.0028 (0.0038)5.86Household size1-2RefRef3-4−0.0010 (0.0018)−2.08−0.0015 (0.0034)−3.065-60.0001 (0.0010)0.020.0008 (0.0041)1.587+0.0032 (0.0046)6.75−0.0037 (0.0086)−7.71Community-level women empowermentLowRefRefHigh0.0116 (0.0067)24.26−0.0056 (0.0052)−11.77Community-level willingness for COVID-19 vaccine uptakeLowRefRefHigh−0.0493 (0.0171)−102.99−0.0076 (0.0052)−15.81Community-level nativityLowRefRefHigh−0.0114 (0.0047)*−23.850.0109 (0.0073)22.68Community-level urbanicityHighRefRefLow0.0163 (0.0085)34.08−0.0391 (0.0110)*−81.72Community-level illiteracyLowRefRefHigh−0.0320 (0.0162)*−66.930.0162 (0.0115)33.84Community-level media useLowRefRefHigh0.0065 (0.0096)13.560.0038 (0.0066)7.88Community-level povertyLowRefRefHigh0.0099 (0.0128)20.70−0.0123 (0.0126)−25.61Notes:* significant at P <0.05. The decomposed difference is the southern–northern gap in the predicted probability of COVID-19 vaccination (southern minus northern); positive contributions widen the gap in favour of the southern region and negative contributions narrow it. β., Decomposition coefficients; Pct., Percentage contribution of each variable category to COVID-19 vaccine uptake.Abbreviations: Ref, Reference; SE, standard error; E, endowment (explained/compositional) component; C, coefficient (unexplained) component. Factors Contributing to Northern-Southern Inequality in COVID-19 Vaccination Notes:* significant at P <0.05. The decomposed difference is the southern–northern gap in the predicted probability of COVID-19 vaccination (southern minus northern); positive contributions widen the gap in favour of the southern region and negative contributions narrow it. β., Decomposition coefficients; Pct., Percentage contribution of each variable category to COVID-19 vaccine uptake. Abbreviations: Ref, Reference; SE, standard error; E, endowment (explained/compositional) component; C, coefficient (unexplained) component. Table 3 presents the aggregate decomposition of the northern–southern difference in COVID-19 vaccination uptake into the endowment (explained, E) and coefficient (unexplained, C) components, together with the percentage contribution of each predictor category to the two components. The endowment component accounted for −34.5% of the observed difference, and the coefficient component accounted for 134.5%. Relative to women aged 15–19 years, the endowment contributions for women aged 20–24 and 25–29 years were negative and statistically significant (−9.73% and −9.41%, respectively), whereas the corresponding coefficient contributions were positive and significant (26.06% and 21.97%, respectively). For women aged 35–39, 40–44, and 45–49 years, both the endowment and coefficient contributions were positive and statistically significant (Table 3). The endowment contributions of secondary and higher education were 65.13% and 50.35%, respectively, with the higher-education contribution reaching statistical significance. The corresponding coefficient contributions for educational attainment were negative and did not reach statistical significance (Table 3). Relative to Christianity, Islamic affiliation showed an endowment contribution of −154.07% and a coefficient contribution of 142.43%, both statistically significant; the contributions for traditional or no religion were small and not significant (Table 3). Relative to women who were not working, working women showed an endowment contribution of 21.15% and a coefficient contribution of −37.39%, both statistically significant (Table 3). Several community-level characteristics had large endowment contributions. The endowment contribution of high community-level willingness for COVID-19 vaccine uptake was −102.99%; that of high community-level illiteracy was −66.93%; and that of high community-level nativity was −23.85% and statistically significant (Table 3). For community-level urbanicity, women in low-urbanicity communities showed a coefficient contribution of −81.72%, which was statistically significant (Table 3). For media exposure, the coefficient contribution of media exposure was −35.06% and statistically significant. The contributions of household wealth, household size, marital status, sex of household head, health insurance coverage, and community-level poverty were generally smaller, and most did not reach statistical significance (Table 3). At the aggregate level, the endowment component accounted for −34.5% and the coefficient component for 134.5% of the northern–southern difference in COVID-19 vaccination uptake (Table 3). This finding indicates that the observed northern disadvantage was not explained by differences in the distribution of the measured characteristics between the two regions. The negative compositional contribution suggests that, if northern women had the same observed characteristics as southern women, the regional vaccination gap would not necessarily be reduced and could instead widen in favour of the south. The substantially larger coefficient contribution indicates that regional differences in the associations between these characteristics and vaccination uptake, together with factors not captured by the measured covariates, accounted for most of the observed disparity. Discussion This study examined northern–southern inequalities in COVID-19 vaccination uptake among Nigerian women of reproductive age, using nationally representative data from the 2024 Nigeria Demographic and Health Survey. This study was conducted to decipher the findings from our previous paper that reported differences in COVID-19 vaccination by region.22 We sought to examine the factors that narrow or widen the gap in COVID-19 vaccination uptake by northern–southern dichotomy. Uptake was higher in the south than in the north, and decomposition analysis attributed most of this gap to the coefficient (unexplained) component rather than to differences in population composition. The principal finding of this study is that the northern–southern inequality in COVID-19 vaccination uptake among Nigerian women of reproductive age was predominantly coefficient-driven rather than compositional. Although vaccination uptake was higher in the south than in the north (29.0% versus 24.1%), the decomposition analysis showed that the coefficient component accounted for 134.5% of the observed gap, whereas the compositional component contributed −34.5%. This pattern is substantively important because it indicates that differences in the distribution of measured demographic, socioeconomic, and community characteristics did not explain the northern disadvantage. Instead, the regional gap was largely associated with differences in how these characteristics translated into vaccination uptake across regions, alongside factors not captured by the model. In other words, the issue is not primarily who northern women are in terms of measured demographic and socioeconomic characteristics, but how the regional environment in which they live shapes the way those characteristics translate into vaccination decisions. This focus also distinguishes the present study from our previous mixed-effects analysis of COVID-19 vaccination coverage22 and from earlier decomposition analyses of other maternal and reproductive health outcomes: rather than identifying determinants of uptake, it formally partitions the northern–southern gap into compositional and coefficient components and quantifies the contribution of each measured characteristic to that gap. The negative (−34.5%.) compositional component represents the portion of the regional difference attributable to differences in the distribution of observed characteristics between northern and southern women. Its value indicates that these differences, considered collectively, acted in the opposite direction to the observed northern disadvantage. Consequently, the counterfactual scenario in which northern women possessed the same observed characteristics as southern women would not necessarily reduce the vaccination gap; rather, the gap could become larger in favour of the south. This finding challenges an interpretation that attributes the regional disparity primarily to northern women having less favourable demographic or socioeconomic characteristics. The coefficient contribution exceeding 100% arises because the coefficient component and compositional component operate in opposite directions: the large positive coefficient contribution is partially offset by the negative compositional contribution, producing the observed net regional difference. The 134.5% coefficient contribution therefore indicates the dominant role of differences in estimated associations and residual contextual influences within the decomposition, rather than a causal proportion attributable to one specific mechanism. Approximately 29% of women in the south had received a COVID-19 vaccine, compared with 24.1% in the north. This pattern is consistent with earlier reports of geographical disparities in vaccine uptake in Nigeria.23,24 Similar inequalities exist across sub-Saharan Africa, where uptake is higher in urbanized and socioeconomically advantaged settings.25 The southern advantage is consistent with the region’s stronger health-system capacity and higher educational attainment, which facilitate both access to and acceptance of vaccination services. Older women, particularly those aged 40–49 years, reported higher uptake, consistent with greater risk perception and health-seeking behaviour.26–28 In the decomposition, the age coefficient contributed more to the gap than did age composition. Educational attainment was among the strongest determinants and a major contributor to the explained component of the gap. Women with secondary or higher education were more likely to be vaccinated, and differences in educational distribution explained a sizeable share of the disparity. Education improves health literacy and the capacity to appraise vaccine information and resist misinformation.29 Lower educational attainment among northern women may therefore contribute to lower uptake. This accords with evidence from Nigeria, Ghana, Kenya, and other low- and middle-income countries in which education consistently predicts uptake.30–32 Religion was also influential: in the decomposition, Islamic affiliation made a strong unexplained contribution, suggesting that the way religious identity shapes vaccination behaviour differs between regions. Religious leaders are known to shape public acceptance of health interventions in Nigeria and other African countries.33,34 During the polio and COVID-19 campaigns, misinformation spread through religious and community networks, particularly in the north. Similar dynamics were reported during the Ebola outbreak, where community trust and local beliefs strongly influenced adherence.35 These findings indicate that religious engagement, rather than religious identity itself, is the appropriate focus for intervention. Health insurance, employment, and media exposure were each associated with higher uptake. Insurance improves healthcare use and preventive behaviour among women in low- and middle-income countries.36 Working women reported higher uptake than those not working. Similar associations have been reported elsewhere in Africa.31 In the decomposition, employment contributed positively to the endowment component but negatively to the coefficient component, indicating that its association with vaccination differed between regions. Media exposure was associated with higher uptake. Mass media and digital communication are well-documented influences on vaccine acceptance.37,38 Media can also spread misinformation rapidly. Household wealth was positively associated with uptake. Similar socioeconomic gradients have been documented globally during the pandemic.39 These social determinants continued to shape access even though vaccines were provided free of charge. At the community level, vaccine acceptance was among the strongest contributors to regional inequality, with women in more accepting communities reporting higher uptake. This supports social ecological theories, in which individual decisions are shaped by community norms, and vaccine attitudes cluster within communities and networks. In Niger State, integrating routine immunization with COVID-19 vaccination improved coverage but also raised equity concerns linked to community characteristics.40 Community illiteracy and nativity also contributed; communities with high illiteracy reported lower uptake, reflecting reduced health literacy and greater vulnerability to misinformation. This is consistent with research showing that community educational and social contexts influence preventive behaviour independently of individual characteristics.41,42 Urban women reported higher uptake than rural women, likely reflecting closer proximity to health facilities and greater exposure to health promotion. Similar urban-rural disparities have been reported across Africa.43 Its large unexplained contribution suggests that differences in health infrastructure and service delivery between north and south remain important barriers to equitable access. Security-related constraints may also contribute to the large coefficient component observed in this study. In parts of northern Nigeria, insecurity and insurgency can disrupt routine health services, restrict population mobility, limit access to health facilities, and complicate the delivery of vaccination campaigns. Such mechanisms were not directly measured in the NDHS and may therefore be captured partly within the unexplained component. This possibility further supports the need to interpret the coefficient component as reflecting a combination of regional differences in observed associations and unmeasured contextual conditions rather than as a single causal mechanism. Notably, the unexplained component accounted for a greater share of the disparity than the explained component, so regional inequalities cannot be attributed solely to demographic and socioeconomic differences. Behavioural, sociocultural, institutional, and structural factors, such as institutional trust, perceived vaccine safety, misinformation, and service quality, appear to dominate. Similar findings during Ebola and COVID-19 show that contextual influences often shape health behaviour more than individual characteristics alone.44,45 The pandemic showed that equitable vaccine availability does not guarantee equitable uptake. Preparedness strategies should prioritize region-specific approaches that build institutional trust, improve female education, strengthen health communication and religious engagement, and invest in rural health infrastructure. These measures will be essential to reduce vaccination inequalities and strengthen resilience against future emergencies. Experience from COVID-19 and Ebola shows that effective emergency responses depend on addressing social, cultural, and structural determinants of health, not only biomedical interventions. Addressing these structural and contextual drivers, rather than vaccine supply alone, is likely to yield the greatest reduction in regional inequities. These priorities align closely with existing national frameworks. Nigeria’s COVID-19 vaccination programme, led by the National Primary Health Care Development Agency (NPHCDA), demonstrated the value of sustained leadership, accountability and transparency mechanisms, engagement of traditional and religious leaders, and community-based mobilization,6,7 and experience from Niger State indicates that integrating COVID-19 vaccination into routine immunization and primary health care services can improve coverage.40 Embedding the region-specific strategies proposed in this study within these NPHCDA-led structures would strengthen their practical relevance and sustainability. The findings suggest several practical region-specific strategies. In northern communities where religious context and limited media exposure contribute to the coefficient component, vaccination programmes could strengthen partnerships with trusted religious leaders, incorporate culturally appropriate vaccine messages into mosque- and community-based communication platforms, and integrate trained religious and community leaders into local vaccination mobilization networks. Mobile vaccination teams could also be deployed through community structures to reach women in rural and underserved areas, particularly where distance, insecurity, or limited health-service availability constrains access. These approaches should complement, rather than replace, broader efforts to improve health-system responsiveness, vaccine confidence, and access to reliable information. Because the primary objective was to quantify the contribution of observed covariates to the disparity in COVID-19 vaccination uptake (binary outcome) rather than to partition residual heterogeneity across hierarchical levels, we used a multilevel logistic regression to account for the clustered structure of the data, followed by a nonlinear decomposition of predicted probabilities using average marginal effects. Covariate-specific contributions were subsequently aggregated to obtain the explained component of the observed disparity. This approach preserves the interpretability of conventional multivariate decomposition while accounting for the hierarchical structure of the data. Strengths and Limitations This study has several strengths. It draws on the 2024 Nigeria Demographic and Health Survey, a large, nationally representative dataset collected using standardized and validated instruments, which supports the generalizability of the findings to women of reproductive age in Nigeria. The use of multivariate decomposition is a further strength, as it partitions the northern–southern gap into explained (endowment) and unexplained (coefficient) components and quantifies the contribution of individual, household, and community-level factors to the observed inequality, while the inclusion of community-level measures allowed contextual influences to be examined alongside individual characteristics. Several limitations should nevertheless be considered. The cross-sectional design precludes causal inference, as the temporal relationship between the determinants and vaccination uptake cannot be established. Vaccination status was self-reported and was not verified against vaccination cards or health records; the 2024 NDHS did not include card-based verification or a card-verified subsample for COVID-19 vaccination, so a verification-based sensitivity analysis was not possible. If the accuracy of self-report differed systematically between northern and southern respondents, differential recall or social desirability bias could have influenced the estimated regional difference and decomposition components. Because the direction and magnitude of such differential misclassification cannot be established from the available data, the decomposition estimates should be interpreted with appropriate caution. In addition, several community-level indicators were derived by aggregating individual responses within enumeration areas. Although this approach provides useful proxies for the social environment, it may introduce ecological or compositional bias because aggregated characteristics do not necessarily represent the experiences or beliefs of every woman within a community. The findings involving community-level variables should therefore be interpreted as contextual associations rather than individual-level effects. Because the analysis was restricted to women of reproductive age, the findings may not be generalizable to men or to other age groups. Finally, the large unexplained component is likely to reflect unmeasured factors, such as vaccine hesitancy, trust in institutions, and supply-side constraints, that were not captured in the survey, and the decomposition estimates may be sensitive to the choice of reference categories and model specification. Conclusion COVID-19 vaccination uptake among women of reproductive age was significantly lower in northern (24.1%) than southern (29.0%) Nigeria, and the decomposition showed that this gap was predominantly coefficient-driven (134.5%) rather than compositional (−34.5%). The northern disadvantage therefore reflects differences in how women’s characteristics translate into vaccination uptake across regions, together with unmeasured contextual conditions, rather than differences in the measured composition of the two populations; equalizing measured characteristics alone would be unlikely to close the gap and could even widen it in favour of the south. Three actionable priorities follow from these findings. First, vaccination programmes in northern Nigeria should work through trusted religious and community leaders, integrating them into local mobilization and mobile immunization networks. Second, health communication should be tailored to communities with low literacy and low media exposure, using local languages and community-based channels. Third, service delivery should be strengthened where rurality and insecurity constrain access, including through mobile outreach integrated into routine primary health care under existing NPHCDA-led structures. Vaccine supply alone will not deliver equitable uptake; addressing the regional context in which vaccination decisions are made is the clearest route to closing the gap. Acknowledgment The authors appreciate the Demographic and Health Survey (DHS) for the approval and access to the original data. Funding Statement This research received no grant from any funding agency in the public, commercial or not-for-profit sectors. Data Sharing Statement Data for this study were sourced and available here: https://dhsprogram.com/data/dataset/Nigeria_Standard-DHS_2024.cfm?flag=1. Ethical Statement The Nigerian DHS is conducted according to the local Nigerian research ethics requirements. Data for this analysis were accessed via the publicly available DHS datasets, with access granted through the DHS programme. As this was a secondary data analysis, further research ethics approval was not required; however, in accordance with DHS regulations, all data extracted from the NDHS for the purpose of this study were handled as confidential and survey respondents remained unidentified. This study conforms to the principles of the Declaration of Helsinki. Ethics Approval and Consent to Participate This study is a secondary data analysis of the DHS which is publicly available, and approval was granted for its use. Written consent was obtained from mothers/caregivers and data were recorded anonymously at the time of data collection during the DHS. Author Contributions All authors made a significant contribution to the work reported, whether that is in the conception, study design, execution, acquisition of data, analysis and interpretation, or in all these areas; took part in drafting, revising or critically reviewing the article; gave final approval of the version to be published; have agreed on the journal to which the article has been submitted; and agree to be accountable for all aspects of the work. Disclosure The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest. References Associated Data Data Availability Statement Data for this study were sourced and available here: https://dhsprogram.com/data/dataset/Nigeria_Standard-DHS_2024.cfm?flag=1.