Abstract
Under-five mortality remains a global health issue, especially in sub-Saharan Africa, where preventable conditions largely drive the high mortality rates. Understanding the heterogeneity in utilization of reproductive, maternal, newborn, and child health services is crucial for reducing under-five mortality. Here we show that among 9307 under-five mortality cases across 31 sub-Saharan African countries (2014–2024), maternal and child health service utilization falls into three distinct patterns—lowest, medium, and highest. Socioeconomic status strongly predicts subgroup membership: higher maternal education, employment, urban residence, and wealth are associated with lower odds of being in the lowest utilization group. Inequality indices further reveal disparities by education, wealth, residence, and employment. Our findings show a strong link between socioeconomic status and maternal and child health services utilization. To address under-five mortality in sub-Saharan Africa, targeted strategies are needed to improve access and uptake of essential health services among socioeconomically disadvantaged groups.
Similar content being viewed by others
Introduction
Under-five Mortality (U5M), defined as the death of a child between birth and their fifth birthday, is considered a critical global health issue. The Sustainable Development Goal target 3.2 aims to end preventable U5M, with all countries aiming to reduce the U5M rate to at least as low as 25 per 1000 live births by 20301. Substantial progress has been made towards meeting this target, with the global U5M rate dropping from 93 deaths per 1000 live births in 1990 to 37 in 2022. However, some regions are still left behind. In 2022, Sub-Saharan Africa (SSA) accounted for approximately 57% of the 4.9 million under-five deaths, yet only 30% of global live births took place in the region2. Tragically, most of these deaths were caused by preventable or treatable conditions (such as preterm birth, pneumonia, diarrhea, and malaria)1,3.
Reproductive, maternal, newborn, and child health (RMNCH) services (such as antenatal care, and institutional births) have received widespread attention in reducing preventable U5M. Additionally, protective environmental factors (such as safely managed drinking water services, sanitation services, and reliance on clean fuels), and cultural practices (such as marriage at a mature age) play a critical role in enhancing maternal and child health4. However, the coverage, and utilization of these services vary widely within and between SSA populations, due to socioeconomic, behavioral, and cultural heterogeneity5,6. This highlights the limitations of a uniform approach in improving service uptake and utilization and reducing U5M. A deeper understanding of populations’ nuanced needs could aid in efforts to tailor targeted services. This is why, the World Health Organization (WHO) highlights the need to leverage data-driven and numerical approaches to customize interventions to specific populations and countries based on quantitative evidence7.
Previous studies have employed various methods to assess utilization of maternal and child health services. Approaches include the use of summary measures such as the composite utilization index (CCI) and co-utilization indicators, as well as analyses of individual utilization indicators8,9. The CCI, calculated as a weighted average of utilization across a set of maternal and child health services, provides a single measure of overall utilization at the population level, allowing for comparisons across countries and tracking progress over time. Co-utilization indicators, on the other hand, measure the proportion of the target population receiving all or most of a set of essential services. Studies using these methods have revealed substantial socioeconomic inequalities in maternal and child health services utilization5,10,11. However, these methods do not reveal how different services are co-utilized, and they often miss key information on the relationship between different indicators12. Evidence suggests that utilization or non-utilization of these life-saving services rarely takes place in isolation13. Instead, different patterns of utilization tend to co-occur, and yet it is not always an obvious relationship. For example, studies have found that mothers who attend recommended four antenatal care visits do not always deliver in a health facility with a skilled birth attendance14,15. In addition, existing composite indices rarely take into account other environmental and cultural practices that are protective of maternal and child health. Yet this information is crucial for understanding the interplay of risk factors and utilization gaps experienced by specific subgroups, and for designing targeted interventions to address these disparities effectively.
Our study takes a novel approach by conducting multilevel latent class analysis (MLCA) using individual-level data of births that ended in U5M based on utilization/uptake of 16 key RMNCH services and protective environmental and cultural factors (collectively referred to as maternal and child health services hereafter). This study makes important contributions to the literature. First, it defines data-driven patterns of co-utilization of maternal and child health services. Second, it examines the association between these utilization patterns and SES. Third, it investigates the inequalities in utilization of these life-saving services along dimensions of wealth, education, place of residence, and employment. We employed MLCA to account for the hierarchical structure of our data. At the individual level, we categorized births into distinct service utilization subgroups, while at the country level, we identified clusters of nations with similar service utilization profiles. This dual-level approach captures both within-country heterogeneity and between-country similarities in service use patterns. By exploring SES predictors of subgroup membership, we tested the hypothesis that SES is associated with health service utilization patterns. Additionally, we tested the hypothesis that there exist disparities in service utilization along SES dimensions by obtaining relative concentration indices (RCI) and slope indices of inequality (SII) related to membership in the optimal utilization subgroup.
Results
Descriptive results
We analyzed Demographic and Health Survey (DHS) (2014–2024) data for 31 SSA countries, focusing on 9307 births that ended in U5M. Descriptive statistics in Table 1 show that slightly more than half of the children were male (54.5%), with a median age at death of 1 month (IQR: 0–8). The median mothers’ age was 29 years (IQR: 23–36 years). The majority (81.8%) of the mothers were married or living with a partner, 36.4% were unemployed, and 42.9% did not attend any school. Approximately a quarter (26.7%) of the mothers were in the lowest wealth quintile, and 71.2% lived in homes located in rural areas. Table 2 presents indicators of utilization of maternal and child health services.
Patterns of utilization of maternal and child health services
We employed MLCA analysis to identify subgroups of utilization patterns of maternal and child health services. Nine consecutive low-level latent class models were estimated (starting with one to nine classes), with individual births as the low-level units. Fit statistics for each model are presented in Fig. 1a. The Bayesian information criterion (BIC) and Akaike information criterion (AIC) decreased notably from one to three classes and then plateaued, indicating diminishing returns on model complexity beyond three classes. Similarly, the integrated complete likelihood (ICL-BIC) showed a decline up to three classes before starting to increase, suggesting that additional classes did not improve the model fit. More importantly, the substantive interpretation of the three-class solution is theoretically meaningful, useful, and parsimonious. As such, we chose the three-class solution as the best model for the low-level latent classes. Using this three-class low-level latent class solution, we fitted nine multilevel nonparametric random effects models to accommodate the nested data structure, with countries serving as high-level units. We explored one to nine high-level latent classes. There was a substantial decline in BIC, AIC, and ICL-BIC values from one to three high-level classes before leveling off (see Fig. 1b). We therefore selected the high-level class solution with three classes as the most appropriate. The three-class low-level and high-level latent class solution was selected as the optimal balance between model fit and parsimony.
Each low-level unit was assigned to a lower-level class with the highest posterior probability following marginalization over the higher-level classes. To characterize the low-level classes, we calculated the standardized mean values of all utilization indicators within each low-level class and obtained a profile of the typical characteristics associated with each low-level class. The radar plot in Fig. 2 shows the low-level latent class profiles. We named classes as highest utilization, medium utilization, and lowest utilization, based on these profiles. The highest utilization subgroup included approximately 61.1% of the overall cohort. Compared to the overall cohort and other subgroups, it was characterized by the highest prevalence of majority (62.5%) of maternal and child health services (that is, institutional births; births attended by a skilled provider; mothers who attended antenatal care visits (4+); mothers with a satisfied need for family planning methods; practiced recommended birth spacing (33 months+); married at a mature age; took part in decision making; lived in homes which rely primarily on clean cooking fuel; use improved sanitation facilities; and have access to protected drinking-water sources). The other two subgroups were characterized by a lower prevalence of maternal and child health services. Both had a much lower proportion of institutional births. The lowest utilization subgroup included approximately 19.0% of the overall sample. Compared to the overall cohort and the other two subgroups, it was characterized by the lowest proportion of all maternal and child health services, except for breastfeeding. The third subgroup had medium utilization and included 19.9% of the overall sample. It had a very low proportion of institutional childbirths and yet had the highest proportion of postnatal care, postpartum care, mothers who received iron supplementation during pregnancy; and received neonatal tetanus protection shots (2+) during pregnancy. On the rest of the indicators, this subgroup had approximately average prevalence of maternal and child health services utilization.
The estimated model resulted in three high-level latent classes for the countries, labeled as HL1, HL2, and HL3. We found that 47.94% of the countries belong to the HL1, 26.26% to HL2, and 25.80% to HL3. Considering first the conditional probabilities for the three low-level classes given these country-level classes (see Table 3), we can see that HL1 has the highest conditional probability for the “high utilization” low-level class, and HL2 similarly has a high conditional probability for the “high utilization” low-level class. HL3 has the highest probabilities for both the “medium utilization” and “low utilization” low-level classes. Figure 3 shows the assignment of countries to classes, with the assignment done based on the highest probabilities given the utilization patterns in the countries. The plot also shows each country’s U5M rates obtained from the corresponding DHS survey data. Fifteen of the thirty-one countries were assigned to HL1, and eight to both HL2 and HL3. Countries classified as HL1 appear to have predominantly lower U5M rates below 60 deaths per 1000 live births, and countries in the HL3 group tend to have higher U5M rates. However, we do not observe a clear relationship between class membership and the U5M rates.
The height of the bars and center of error bars indicate each country’s estimated probability that a child will die before the age of five years, reported as deaths per 1000 live births, obtained from the DHSs. The error bars represent 95% confidence intervals of these under-five mortality rates. Source data are provided as a Source Data file.
Associations between utilization patterns and SES indicators
When examining predictors of the lower utilization subgroups vs the highest utilization subgroup (Table 4), younger maternal age and high parity were associated with a significant increase in odds of membership to the lowest and medium utilization low-level subgroups. Being separated/divorced and having a partner were associated with higher odds of membership in the lowest utilization low-level subgroup. In terms of SES, mothers having attained some level of education, being in the middle, high, or highest wealth quintile, and living in an urban area were associated with decreased odds of being in the lowest and medium utilization subgroups. Additionally, mothers being employed was associated with decreased odds of being in the lowest utilization subgroup.
Inequality measures
We examined utilization inequality using the SII and RCI across key SES indicators (see Table 5). Our findings revealed disparities, particularly pronounced in education, wealth, and place of residence, where higher positive values of SII and RCI indicate that better health outcomes are disproportionately concentrated among more affluent, educated, and urban populations. Employment-related disparities, although less marked, also highlight utilization inequalities.
Discussion
Using MLCA, we revealed patterns of co-utilization of maternal and child health services, in 31 SSA countries. This method signaled three distinct utilization subgroups, that is, highest, medium, and lowest utilization, highlighting ways in which services are co-utilized. Notably, mothers in the medium utilization subgroup attended antenatal and postnatal care yet had home deliveries, while mothers in the highest utilization subgroup primarily had institutional deliveries but did not attend postnatal and postpartum care. These findings indicate unique gaps in the continuum of care. At the country level, we uncovered three high-level latent classes (HL1, HL2, and HL3), revealing country similarities and differences in service utilization. HL3 countries had the highest share of their populations belonging to the lowest and medium utilization subgroups.
Multinomial regression analysis revealed associations between maternal and child health service utilization and SES. Mothers with higher education had lower odds of belonging to the lowest utilization subgroup (OR: 0.00, 95% confidence interval (CI): 0.00–0.00). Mothers in higher wealth quintiles were less likely to be in the lowest utilization group (0.23, 0.12–0.43). Urban residence was associated with decreased odds of belonging in the lowest utilization subgroup (0.37, 0.27–0.51), and mothers’ employment showed a protective effect against belonging to the lowest utilization subgroup (0.60, 0.47–0.78). Furthermore, SII and RCI inequality measures indicated disparities in service utilization along dimensions of education (SII: 0.679, RCI: 0.197), wealth (SII: 0.525, RCI: 0.155), place of residence (SII: 0.599, RCI: 0.120), and employment (SII: 0.128, RCI: 0.028). Our findings corroborate a well-established body of literature indicating a SES gradient in maternal and child healthcare utilization and access5,6,16,17. However, co-utilization patterns of maternal and child health services along the continuum of care remain underexplored, and our study adds insights to address this gap.
The highest utilization subgroup had the highest prevalence of environmental protective factors, namely, access to safe drinking water, sanitation facilities, and clean cooking fuel—described in a 2021 study as “services delivered in household environment”5,18. These advantages are attributable to higher SES, as households with a high SES are more likely to afford constructing sanitation facilities19, access piped water grids due to living in urban areas, and to afford clean cooking fuel20,21. This subgroup also had the highest utilization of health facility-based services that is antenatal care, and institutional childbirths, and moderate utilization of postnatal and postpartum care services. This may reflect better accessibility through proximity to health facilities, or the ability to afford transportation to health facilities22, as well as knowledge of the benefits of skilled healthcare during pregnancy, childbirth, and following childbirth23. Other prevalent attributes were marriage at a mature age and the mother’s decision-making participation. These align with this subgroup’s association, with higher education, and older mothers’ age, potentially explaining the delayed marriage24, and the higher autonomy and agency of women in this subgroup.
The medium utilization subgroup was characterized by a mixed pattern of health facility-based service use—marked by moderate antenatal care attendance, very low rates of institutional births, and the highest prevalence of postnatal and postpartum care. In a qualitative study conducted among rural women in Tanzania25, most mothers reported attending antenatal care services and understanding its importance, yet more than half had home deliveries due to accessibility barriers. Additionally, in that study, the majority of mothers who had home deliveries, reported seeking postnatal care to ensure the good health of their newborns. Our findings in this subgroup provide empirical support for practices reported in this qualitative study25.
Conversely, the lowest utilization subgroup exhibited the lowest prevalence on all indicators except breastfeeding. This higher prevalence of breastfeeding is consistent with earlier studies showing that breastfeeding rates are often higher among mothers who are unemployed26, as they can practice extended breastfeeding more easily compared to working mothers. Moreover, unemployment and poverty restrict affordable access to breast milk substitutes. Therefore, while financial constraints hinder uptake of health facility-based services in this subgroup, a mother’s unemployment appears to facilitate adherence to optimal breastfeeding practices, possibly out of necessity and lack of alternatives.
At the country level, we observed notable heterogeneity in the prevalence of the low-level utilization subgroups across countries. Countries classified as HL1, representing nearly half of our sample, demonstrate high rates of “high utilization” (79.08%), suggesting these nations have likely developed more accessible maternal and child health services. Meanwhile, HL3 countries exhibit a more fragmented utilization landscape, with substantially higher proportions of “medium” and “low” utilization individuals. To contextualize our findings, we compare them with a recent multi-country study on wealth-based inequalities in the continuum of maternal health service utilization across 16 SSA countries27. This study found that completion of the maternal continuum of care was below 50% in nine of the 16 countries they analyzed, four of which (Angola, Ethiopia, Guinea, and Nigeria) have been classified as HL3 countries in our study. This further supports our findings. Additionally, our latent classes, reveal co-utilization patterns of health services, within these countries, which provides insights into gaps and opportunities for intervention delivery.
By demonstrating clear subgroup differences, the policy implications for our findings are a call for targeted interventions to address specific subgroup needs. As such, identification of mothers, belonging to a given subgroup, is crucial, to enable suitable recommendations and follow-ups. This can be done using demographics and SES characteristics. For the low utilization subgroup, comprehensive individual and community-level strategies, aimed at improving SES, and access to maternal and child health services are necessary. At the individual level, poverty alleviation programs such as conditional cash transfers have been shown to increase antenatal care attendance and institutional deliveries by reducing financial barriers28. At the community level, community health worker programs have been shown to effectively increase antenatal care, and reduce infant mortality rates, by providing home-based care and referrals29,30. Similarly, women empowerment programs have been shown to enhance education, financial independence, and decision-making power to improve maternal and child healthcare utilization31,32. For the medium utilization subgroup, interventions should focus on addressing the specific barriers that prevent women from accessing institutional delivery services, while leveraging the relatively higher uptake of antenatal and postnatal care services. For example, community-based antenatal counseling. A study in Burkina Faso found that behavior change communication during antenatal care significantly improved facility-based births33. Additionally, a study in Nigeria showed that engaging traditional birth attendants has been successful in increasing institutional deliveries34. These evidence-based interventions can effectively increase institutional delivery rates in this subgroup, improving outcomes. It is worth noting that the patterns derived from our findings are still preliminary. Confirmatory studies are imperative to assess subgroup differences in relevant outcomes, that is, maternal and child morbidity and mortality.
Our findings should be interpreted considering some limitations. First, the data are self-reported, which may introduce response biases, and the cross-sectional nature of the surveys limits our ability to establish causality. Second, our use of asset indices to classify family wealth quintiles includes variables like housing quality and access to water and sanitation, inherently linking wealth to environmental protective measures. Despite this, a comprehensive UNICEF analysis shows persistent disparities in the distribution of these indicators, even when removed from the asset index35. Furthermore, the indicators we used did not include certain critical variables on the continuum of maternal and child health services, that is, immunizations, minimum dietary diversity, and health-seeking practices, as these indicators are not available at the individual-level among births that ended in U5M.
Despite its limitations, the findings of this study provide insight into understanding heterogeneity in utilization patterns of maternal and child health services. By identifying utilization subgroups at both individual and country levels and the associated SES inequalities, interventions can be better tailored to the needs of specific populations. The substantial disparities revealed highlight the need for equity-focused approaches that prioritize socioeconomically disadvantaged populations to improve maternal and child health outcomes across SSA.
Methods
Data source
In this multi-national study, we analyzed publicly available data from the DHS program36. The DHS are nationally representative cross-sectional surveys, mostly done in low-and middle-income countries. The surveys contain information on household characteristics, HIV, reproductive health, women’s and children’s health, nutrition, and mortality37. DHS are conducted by national central statistics agencies or research institutes. Procedures and questionnaires utilized are reviewed and approved by ICF and country-specific Institutional Review Boards. The institutions that approved, provided funding for, or implemented the surveys were responsible for ethical clearance, which guaranteed consent (for children, consent was given by their caregiver), and confidentiality of the respondents’ information. The surveys use similar multi-stage cluster sampling methods to select women of reproductive age (15–49 years) and children younger than five years for inclusion38. We used the most recent survey (conducted no earlier than 2014 until May 2024) from 31 SSA countries (see details in Supplementary Note 1). We utilized the birth recode file, which contains full birth history for women interviewed, including information on pregnancy, postnatal care, and health for children born within the last five years. We focused exclusively on births that ended in U5M, and on the last birth for each woman, as the majority of available RMNCH utilization indicators pertain to the last birth. The final dataset contained 9307 births meeting these criteria. No statistical method was used to predetermine the current study’s sample size. In this pooled dataset, we rescaled the sampling weights (provided by DHS), such that each country’s total weight is proportional to the country’s population size during the year of survey. We obtained population estimates from the United Nations World Population Prospects39.
Utilization indicators
We selected 16 key internationally recognized core health indicators, including ten RMNCH intervention indicators, and six protective environmental and cultural factors4. The RMNCH indicators describe essential services along the continuum of care for women, neonates, and children, including family planning demand satisfaction; antenatal care; antenatal iron supplementation; neonatal tetanus protection; institutional childbirth; births attended by skilled health personnel; postpartum care; postnatal care; birth spacing; and breastfeeding. The environmental factors include primary reliance on clean fuels, improved sanitation facilities, and protected drinking water sources. Tobacco use, marriage at a mature age, and decision-making participation by mothers were the three cultural factors. We included these 16 variables based on their relevance in the scientific literature pertaining to U5M4,40. We did not include crucial indicators on immunizations and case management for common illnesses (diarrhea and pneumonia) among children because this data was not collected for births that ended in U5M. Details about the variables are provided in the Supplementary Material (Supplementary Note 2).
Predictors of utilization subgroups
Child’s sex, mother’s age, marital status, and parity were considered as demographic predictors of the utilization subgroups. The family’s wealth quintile, mother’s employment status, education level, and place of residence (i.e., rural or urban) were included as SES indicators.
Data preprocessing
The dataset contained missing values for 14 variables, with relatively low missingness percentages ranging from 0% to 7.84%. These included mother’s employment status (0.16%), health facility delivery (0.17%), skilled delivery provider (0.20%), breastfeeding (0.59%), postnatal check (0.73%), iron pills during pregnancy (0.89%), family planning demand satisfaction (1.34%), neonatal tetanus protection (1.62%), antenatal care visits (4+) (1.71%), protected drinking water source (2.25%), clean cooking fuel (2.25%), improved sanitation facility (2.29%), postpartum check (5.78), and marriage at mature age (7.84%). Mode imputation was performed for each variable. We conducted a sensitivity analysis to assess the impact of using an imputed dataset compared to performing a complete-case analysis. We found no significant differences in the revealed latent classes (see Supplementary Note 3).
MLCA
The overall analytical strategy of this study involved two stages. First, we defined latent variables representing patterns of maternal and child health service utilization across multiple countries. Second, we examined the association between utilization patterns and SES, and subsequently quantified inequality gaps in service utilization across population subgroups.
To identify data-driven patterns of maternal and child health service utilization, we used latent class analysis (LCA). LCA is a “person-centered” approach that groups individuals into discrete classes or groups based on a set of responses to a set of observable variables41. The LCA model assumes that observations are independent of each other. However, this assumption is often violated when the data have a multilevel structure, namely when lower-level units are nested in higher-level ones42, as is our case (individuals are nested within primary sampling units nested within countries). For this reason, we employed the MLCA method42 to account for the hierarchical data structure. To maintain analytical tractability while still accounting for the hierarchical structure in the data, we focused exclusively on country-level clustering in our multilevel framework. We fitted a multilevel latent class model estimated with a two-step estimator using the multiLCA43 function in R. The model was specified with individual births as the lower-level units and countries as the higher-level units. The model was initialized using the k-means algorithm to establish low-level latent classes, followed by the application of the expectation maximization (EM) algorithm to optimize the model’s fit to the data by maximizing the log likelihood formally defined as:
and parameterized by \({{{\mathbf{\Phi }}}}\), the class-specific item-response probabilities for the low-level latent classes; \({{{\mathbf{\Pi }}}},\) the conditional low-level class-membership probabilities for individuals given their country’s high-level class; and \({{{\mathbf{\omega }}}}\), the distribution of countries in high-level latent classes. \(P({{{{\bf{Y}}}}}_{j})\) represents the probability of observing a specific combination of responses within each country, and is formally defined as:
In this context, \({Y}_{{ij}}={({Y}_{{ij}1},\ldots,{Y}_{{ijH}})}^{{\prime} }\) is the vector of observed responses, \({Y}_{{ijh}}\) is the response of an individual \(i=(1,\ldots,{n}_{j})\) in a country \(j=\left(1,\ldots,J\right)\) on the \({h}_{{th}}\) utilization indicator variable, with \(h=1,\ldots,H\). A multilevel latent class model specifies the probability \(P({{{{\bf{Y}}}}}_{j})\) of observing a particular response configuration for each country \(j.\) The probability expression \({{{{\rm{\omega }}}}}_{m}\) represents the probability that country \(j\) belongs to class \(m\), \({\pi }_{t|m}\) denotes the conditional probability that an individual belongs to latent class \(t\) given that their country is in latent class \(m\), and \(P({Y}_{{ijh}}|{X}_{{ij}}=t)\) denotes the probability of the response \({Y}_{{ijh}}\) given the latent class \(t\). Details of this model have been published elsewhere42.
To determine the optimal number of high-level and low-level latent classes, we systematically explored various model configurations by estimating latent class models for one to nine classes to ascertain the optimal structure for latent classes. We first examined low-level latent classes using the 16 indicators of utilization of maternal and child health services for each individual. The selection criteria were based on evaluating the AIC44, BIC45, and ICL-BIC46 for each model. AIC is defined as:
where: \(L\) is the maximized likelihood of the model given the data; \(k\) is the number of parameters in the model; and \(n\) is the number of data points. BIC, on the other hand, is defined as:
The ICL-BIC is an extension of BIC for latent class and mixture models, by adding an entropy term. It is defined as:
Entropy measures how well the model separates data into distinct latent classes. Therefore, ICL-BIC accounts for both model fit and the quality of the latent class solution, preferring models that produce clear, well-separated groups. This approach enabled us to assess the trade-off between model complexity and fit, guiding our selection of the most parsimonious and theoretically sound model. Upon establishing the optimal low-level latent class model, we extended our analysis to include high-level latent classes, focusing on the nested nature of the data with countries as the high-level indicators. We similarly evaluated models ranging from one to nine high-level latent classes based on AIC, BIC, and ICL-BIC, aiming to identify the configuration that best balanced detailed representation of the data with model simplicity.
Countries were assigned to high-level classes based on the maximum posterior probabilities derived from the MLCA. Each country was categorized into one of three groups by identifying the class with the highest probability from the model outputs. Bar plots were used to represent the assigned high-level class assignment for each country. Similarly, for individuals, class assignments were made by selecting the low-level class with the highest posterior probability for each individual, after marginalization over the high-level classes. We used radar plots to visually describe the generated low-level latent classes. The plots were generated by normalizing the mean of each categorical indicator for a specific class to the mean of the overall cohort.
Multinomial analysis
Following MLCA, a multivariable multinomial regression model was fitted, with low-level class membership as the dependent variable, and demographic and SES indicators as the independent variables. The multinomial logistic regression model for the three latent classes was formally defined as:
where \({{{{\bf{Y}}}}}_{i}\) is the low-level latent class for individual, \(i\), \({X}_{{ip}}\) are predictor variables, and \(\beta\)’s represent regression coefficients. To properly account for the complex survey design of the DHS datasets, we implemented the analysis using the svydesign framework, incorporating sampling weights, cluster identifiers (primary sampling units), and stratification variables as specified in the DHS sampling methodology. This approach ensures that our estimates correctly reflect the hierarchical sampling structure and provides appropriate standard errors accounting for the design effects. The highest utilization subgroup was used as the reference subgroup for class membership. Odds ratios (OR) with 95% CI of belonging to a given class were reported, and variables with p-value < 0.05 in the multivariable analysis were considered significant predictors of utilization subgroups. All statistical tests were two-tailed to account for both positive and negative associations.
Measures of inequality
To quantify utilization inequalities across wealth, education, place of residence, and employment dimensions, we employed two key inequality measures: the SII and the RCI. SII is an absolute measure of inequality that shows the difference in estimated indicator values between the most-advantaged and most-disadvantaged subgroups, while accounting for all other subgroups—through appropriate regression modeling47. RCI, on the other hand, is a relative measure showing the gradient across population subgroups on a relative scale47. It indicates the extent to which an indicator is concentrated among disadvantaged or advantaged subgroups. Subgroups are weighted by their population share in both measures. For both measures, a value of zero indicates no inequality, positive values indicate a concentration of the indicator among the advantaged, and negative values indicate a concentration of the indicator among the disadvantaged. Since the highest utilization subgroup represents the most desirable health-related characteristics, we applied these inequality measures to evaluate its distribution across the population across various SES groups. We implemented this analysis using the healthequal library in R. To properly account for the complex survey design of the DHS datasets, we specified the sampling weights, cluster identifier (primary sampling units), and stratification variable as specified in the DHS sampling methodology.
All modeling and statistical analysis were performed using R (version 4·3·1) with packages multilevLCA (2.0.1), survey (4.4.2), svrepmisc (0.2.2), and healthequal (1.0.1). Data preprocessing and visualizations were generated using Python (version 3.11.5) with packages pandas (2.2.3) and matplotlib (3.9.2).
Reporting summary
Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.
Data availability
The datasets analyzed in this study are publicly available and can be accessed through the DHS Program website (https://dhsprogram.com/). Source data are provided with this paper.
Code availability
The code used to generate results in this study is available and can be accessed through this link (https://github.com/CNajjuuko/RMNCH_utilization_patterns)48.
References
End Preventable Deaths of Newborns and Children Under 5 Years of Age (WHO, 2024) https://www.who.int/data/gho/data/themes/topics/sdg-target-3_2-newborn-and-child-mortality.
Azevedo, J. P., Banerjee, A., Wilmoth, J., Fu, H. & You, D. Hard truths about under-5 mortality: call for urgent global action. Lancet 404, 506–508 (2024).
Global child deaths reach historic low in 2022—UN report. https://www.who.int/news/item/13-03-2024-global-child-deaths-reach-historic-low-in-2022---un-report.
2018 Global reference list of 100 core health indicators (plus health-related SDGs) https://www.who.int/publications-detail-redirect/2018-global-reference-list-of-100-core-health-indicators-(-plus-health-related-sdgs).
Leventhal, D. G. P. et al. Delivery channels and socioeconomic inequalities in coverage of reproductive, maternal, newborn, and child health interventions: analysis of 36 cross-sectional surveys in low-income and middle-income countries. Lancet Glob. Health 9, e1101–e1109 (2021).
Alam, N., Hajizadeh, M., Dumont, A. & Fournier, P. Inequalities in maternal health care utilization in sub-Saharan African Countries: a multiyear and multi-country analysis. PLoS One 10, e0120922 (2015).
World Health Organization. Nurturing Care for Early Childhood Development: A Framework for Helping Children Survive and Thrive to Transform Health and Human Potential (World Health Organization, 2018).
Barros, A. J. et al. Equity in maternal, newborn, and child health interventions in Countdown to 2015: a retrospective review of survey data from 54 countries. Lancet 379, 1225–1233 (2012).
Barros, A. J. D. & Victora, C. G. Measuring coverage in MNCH: determining and interpreting inequalities in coverage of maternal, newborn, and child health interventions. PLoS Med. 10, e1001390 (2013).
Boerma, T. et al. Countdown to 2030: tracking progress towards universal coverage for reproductive, maternal, newborn, and child health. Lancet 391, 1538–1548 (2018).
Wehrmeister, F. C., Barros, A. J. D., Hosseinpoor, A. R., Boerma, T. & Victora, C. G. Measuring universal health coverage in reproductive, maternal, newborn and child health: an update of the composite coverage index. PLoS One 15, e0232350 (2020).
Verma, A., Angelini, O. & Di Matteo, T. A new set of cluster driven composite development indicators. EPJ Data Sci. 9, 1–21 (2020).
Singh, K., Story, W. T. & Moran, A. C. Assessing the continuum of care pathway for maternal health in South Asia and Sub-Saharan Africa. Matern. Child Health J. 20, 281–289 (2016).
Anastasi, E. et al. Losing women along the path to safe motherhood: why is there such a gap between women’s use of antenatal care and skilled birth attendance? A mixed methods study in northern Uganda. BMC Pregnancy Childbirth 15, 287 (2015).
Wang, W. & Hong, R. Levels and determinants of continuum of care for maternal and newborn health in Cambodia-evidence from a population-based survey. BMC Pregnancy Childbirth 15, 62 (2015).
Langlois et al. Inequities in postnatal care in low- and middle-income countries: a systematic review and meta-analysis. Bull. World Health Organ. 93, 259–270G (2015).
Atake, E.-H. Socio-economic inequality in maternal health care utilization in Sub-Saharan Africa: eEvidence from Togo. Int. J. Health Plann. Manage. 36, 288–301 (2021).
Balde, M. D. Coverage of RMNCH interventions: more analysis is needed. Lancet Glob. Health 9, e1037–e1038 (2021).
Simiyu, S., Swilling, M., Rheingans, R. & Cairncross, S. Estimating the cost and payment for sanitation in the informal settlements of Kisumu, Kenya: a cross sectional study. Int. J. Environ. Res. Public. Health 14, 49 (2017).
Muller, C. & Yan, H. Household fuel use in developing countries: review of theory and evidence. Energy Econ. 70, 429–439 (2018).
EkouevI, K. & Tuntivate, V. Household Energy Access for Cooking and Heating: Lessons Learned and the Way Forward (The World Bank, 2012); https://doi.org/10.1596/978-0-8213-9604-9
Gage, A. J. & Guirlène Calixte, M. Effects of the physical accessibility of maternal health services on their use in rural Haiti. Popul. Stud. 60, 271–288 (2006).
Brandstetter, S. et al. Information on, knowledge and utilisation of support services during pregnancy and after childbirth: cross-sectional analyses of predictors using data from the KUNO-Kids health study. BMJ Open 10, e037745 (2020).
Marphatia, A. A. et al. How much education is needed to delay women’s age at marriage and first pregnancy? Front. Public Health 7, 199341 (2020).
Mrisho, M. et al. The use of antenatal and postnatal care: perspectives and experiences of women and health care providers in rural southern Tanzania. BMC Pregnancy Childbirth 9, 10 (2009).
Habte, M. H. et al. The effect of unemployment and post-natal care on the exclusive breast-feeding practice of women in Ethiopia: a systematic review and meta-analysis. Reprod. Health 19, 94 (2022).
Asefa, A. et al. Wealth-based inequality in the continuum of maternal health service utilisation in 16 sub-Saharan African countries. Int. J. Equity Health 22, 203 (2023).
Glassman, A. et al. Impact of conditional cash transfers on maternal and newborn health. J. Health Popul. Nutr. 31, 48–66 (2013).
Kayentao, K. et al. Effect of community health worker home visits on antenatal care and institutional delivery: an analysis of secondary outcomes from a cluster randomised trial in Mali. BMJ Glob. Health 8, e011071 (2023).
Nadella, P., Subramanian, S. V. & Roman-Urrestarazu, A. The impact of community health workers on antenatal and infant health in India: a cross-sectional study. SSM Popul. Health 15, 100872 (2021).
Prost, A. et al. Women’s groups practising participatory learning and action to improve maternal and newborn health in low-resource settings: a systematic review and meta-analysis. Lancet Lond. Engl. 381, 1736–1746 (2013).
Fottrell, E. et al. The effect of increased coverage of participatory women’s groups on neonatal mortality in Bangladesh: a cluster randomized trial. JAMA Pediatr 167, 816–825 (2013).
Soubeiga, D., Sia, D. & Gauvin, L. Increasing institutional deliveries among antenatal clients: effect of birth preparedness counselling. Health Policy Plan 29, 1061–1070 (2014).
Oyeneyin, L., Osunmakinwa, O. & Olagbuji, Y. Incorporating traditional birth attendants into the mainstream maternal health system in Nigeria—an evaluation of the Ondo State Agbebiye program. Afr. J. Reprod. Health Rev. Afr. Santé Reprod. 25, 82–88 (2021).
Martel, P. Review of options for reporting water, sanitation and hygiene coverage by wealth quintile. In MICS Methodological Papers, No. 4. Data and Analytics Section, Division of Data, Research, and Policy (UNICEF, New York, 2016).
The DHS Program—Data. https://dhsprogram.com/data/ (2024).
Corsi, D. J., Neuman, M., Finlay, J. E. & Subramanian, S. Demographic and health surveys: a profile. Int. J. Epidemiol. 41, 1602–1613 (2012).
Hancioglu, A. & Arnold, F. Measuring coverage in MNCH: tracking progress in health for women and children using DHS and MICS household surveys. PLOS Med. 10, e1001391 (2013).
United Nations. World Population Prospects (UN); https://population.un.org/wpp/
Organization, W. H. Monitoring Maternal, Newborn, and Child Health: Understanding Key Progress Indicators (WHO, 2019).
Hagenaars, J. A. & McCutcheon, A. L. Applied Latent Class Analysis (Cambridge University Press, 2002).
Mari, R. D., Bakk, Z., Oser, J. & Kuha, J. A two-step estimator for multilevel latent class analysis with covariates. Psychometrika 88, 1144–1170 (2023).
Lyrvall, J., Di Mari, R., Bakk, Z., Oser, J. & Kuha, J. Multilevel Latent Class Analysis: State-of-the-Art Methodologies and Their Implementation in the R Package multilevLCA. Multivar. Behav. Res. 60, 731–747 (2025).
Cavanaugh, J. E. & Neath, A. A. The Akaike information criterion: background, derivation, properties, application, interpretation, and refinements. WIREs Comput. Stat. 11, e1460 (2019).
Penny, W. D., Mattout, J. & Trujillo-Barreto, N. CHAPTER 35—Bayesian model selection and averaging. In Statistical Parametric Mapping (eds. Friston, K., Ashburner, J., Kiebel, S., Nichols, T. & Penny, W.) 454–467 (Academic Press, London, 2007); https://doi.org/10.1016/B978-012372560-8/50035-8
Biernacki, C. Assessing a mixture model for clustering with the integrated completed likelihood. In IEEE Transactions on Pattern Analysis and Machine Intelligence (ACM, 2000); https://dl.acm.org/doi/10.1109/34.865189
Schlotheuber, A. & Hosseinpoor, A. R. Summary measures of health inequality: a review of existing measures and their application. Int. J. Environ. Res. Public. Health 19, 3697 (2022).
Najjuuko, C. Patterns of maternal and child health services utilization and associated socioeconomic disparities in sub-Saharan Africa. Zenodo https://doi.org/10.5281/zenodo.15490575 (2025).
Acknowledgements
This study was supported in part by funding from the NIH Researcher Resilience Training grant R25MH118935-01 to Dr. Fred M. Ssewamala, as well as support from the AI Health Institute and the Fullgraf Foundation to Dr. Chenyang Lu. The sponsors had no role in the study design, data collection, data interpretation, manuscript preparation, or the decision to submit the work for publication.
Author information
Authors and Affiliations
Contributions
C.N. did the data curation, formal analysis, visualization, and wrote the original draft of the manuscript. F.M.S. conceptualized the study, guided on results, and reviewed and significantly improved the manuscript. C.L. led the study design and methodology, supervised the analyses, and reviewed and significantly improved the manuscript. Z.X. critically reviewed the analyses and contributed to the writing of the manuscript. S.K. critically reviewed the analyses and edited the manuscript. All authors had final responsibility for the decision to submit for publication.
Corresponding author
Ethics declarations
Competing interests
The authors declare no competing interests.
Peer review
Peer review information
Nature Communications thanks Md Mizanur Rahman, Lance Waller and the other anonymous reviewer(s) for their contribution to the peer review of this work. A peer review file is available.
Additional information
Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Supplementary information
Source data
Rights and permissions
Open Access This article is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License, which permits any non-commercial use, sharing, 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 you modified the licensed material. You do not have permission under this licence to share adapted material derived from this article or parts of it. 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-nc-nd/4.0/.
About this article
Cite this article
Najjuuko, C., Xu, Z., Kizito, S. et al. Patterns of maternal and child health services utilization and associated socioeconomic disparities in sub-Saharan Africa. Nat Commun 16, 7840 (2025). https://doi.org/10.1038/s41467-025-61350-8
Received:
Accepted:
Published:
DOI: https://doi.org/10.1038/s41467-025-61350-8