Study setting

Nigeria is in West Africa, between 4° and 14° N latitude and 3° and 14° E longitude. It borders Benin, Niger, Chad, and Cameroon and is Africa’s most populous country, with over 230 million people, 54.9% of whom live in urban areas [46]. The country has a median age of 18.1 years, indicating a substantially young population [46]. Nigeria is highly diverse, with over 374 ethnic groups and languages. The three main groups, Hausa-Fulani, Yoruba, and Igbo, and several others, make the country one of the most multilingual nations in the world [47]. Administratively, Nigeria is divided into 36 states and the Federal Capital Territory (FCT), organised into six geopolitical zones (‘regions’): North-East (six states), North Central (six states and the FCT), North-West (seven states), South-South (six states), South-West (six states), and South-East (five states) [Fig 1]. Nigeria also has 774 local government areas (LGAs), with wards or enumeration areas serving as the smallest administrative units. Approximately 63% of Nigerians experience multidimensional poverty, reflecting deep socioeconomic inequalities [48].

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1. Map of Nigeria showing the geo-political zones.

Note: Map of Nigeria showing states and geopolitical zones. The map was generated by the authors using R with data from the rnaturalearth and rnaturalearthdata packages (public domain). A freely accessible script for generating the map is available at https://github.com/actright1/Nigeria-Geopolitical-Map.

https://doi.org/10.1371/journal.pone.0337438.g001

Nigeria’s healthcare system comprises both public and private sectors, operating in parallel and with limited integration. The public sector follows a three-tier structure: primary, secondary and tertiary. Primary healthcare is managed at the local government level through primary health centres, secondary care is provided by state-owned general hospitals, and tertiary care is delivered through teaching hospitals and federal medical centres offering specialised services. Despite this structure, the private sector, including formal and informal providers, plays a major role in service delivery and accounts for a considerable proportion of healthcare utilisation in Nigeria [49].

Health facility distribution in Nigeria is highly uneven across states and regions. Of the 34,423 registered facilities, 88% are primary, 12% secondary, and only 0.25% tertiary centres, with about two-thirds publicly owned and one-third private [50]. On average, there are 16.6 primary, 2.2 secondary, and 0.05 tertiary facilities per 100,000 population [50]; however, this figure masks sharp regional disparities. For example, the North Central has the highest density of primary facilities (23.3 per 100,000), while the South-South has the lowest (11.8 per 100,000). Secondary facilities are concentrated in the South-East (5.4 per 100,000) but remain scarce in the North-West (0.5 per 100,000). Tertiary facilities are clustered in the South-West and South-South, with a minimal presence in the North-East and North-West. Moreover, private facilities are more prevalent in southern states, whereas northern states such as Jigawa, Sokoto, and Zamfara rely more on public facilities [50].

Facility distribution also reflects a clear North–South and urban–rural divide: southern and urban areas host more private, secondary, and tertiary facilities, while northern and rural areas depend heavily on public primary centres, often under-resourced [50]. These disparities are corroborated by the Lancet Nigeria Commission, which found that facility density is consistently higher in the south than in the north, and that several northern states also scored lowest on the health facility quality index, underscoring gaps in availability, infrastructure and staffing quality [51]. The uneven distribution of healthcare facilities in Nigeria may contribute to poorer access and lower quality of care in northern and rural regions. Despite the Nigerian government’s efforts to improve maternal and neonatal health in line with global health goals, through initiatives such as the MSS, CCTs, and BHCPF [15–18,20–22], mentioned earlier, poor maternal and neonatal health outcomes continue to persist. Home birth is a potential contributor, and national data may obscure within-population disparities [34,51,52]; hence, the approach in this study.

Data source and sampling design

We utilised data from the cross-sectional Nigerian Demographic and Health Survey (NDHS) 2018. The NDHS is a nationally representative survey conducted periodically since its inception in 1990 [13]. The survey employs a rigorously tested and validated methodology. The 2018 version was implemented by the Nigerian National Population Commission in collaboration with international partners and technical oversight from ICF International. The survey provides information on key demographic and health indicators, including fertility, family planning, maternal and child health, and health service utilisation. The NDHS 2018 used a stratified two-stage cluster sampling method to select 42,000 households from 1,400 clusters [13]. Data collection for the survey was conducted between 14 August and 29 December 2018 using standardised, validated questionnaires administered by trained interviewers [13]. The completed survey covered 40,427 households within 1,389 clusters. A total of 41,821 mothers aged 15–49 participated (16,984 from urban and 24,837 from rural areas). The overall response rate was high at 99.3%, with urban and rural response rates of 99.2% and 99.4%, respectively.

In this study, we analysed a weighted representative sample of 21,512 mothers—8,467 from urban areas and 13,055 from rural areas—who had a live birth in the five years before the survey [13]. These mothers provided complete data on the place of childbirth for their most recent live birth. We used the Children Recode (KR) dataset, restricted to the most recent live birth per mother. This dataset includes detailed records on pregnancy, childbirth, postnatal care, and child health, together with maternal information, fully anonymised to protect personally identifiable data. Documentation on the NDHS 2018 methodology is publicly available [13]. Ethical approval for the NDHS 2018 survey was granted by the Nigerian National Health Research Ethics Committee, and informed consent was obtained from participants aged 18 and older, while parental consent was required for those under 18. The dataset can be accessed online at https://dhsprogram.com/data/available-datasets.cfm, upon approval from the DHS Program [13]. As our study involved secondary analysis of fully de-identified publicly available data, no additional ethics approval was required.

Study factors

Explanatory variables.

We adopted Andersen’s Behavioural Model as the conceptual framework and utilised it for selecting explanatory variables, a well-established approach in healthcare services utilisation research [11,34,35,53]. The model categorises factors influencing healthcare utilisation into four key domains: environmental factors (external conditions, including broader contextual influences), predisposing factors (demographic and social characteristics shaping healthcare-seeking behaviour), enabling factors (resources that facilitate or hinder access to care), and need factors (individual perceptions and evaluations of health needs) [41]. Using this framework, we examined how various factors are associated with having a birth at home rather than in a healthcare facility. The selection of explanatory variables was also guided by a review of previous research and the availability of information in our dataset [10,11,34,35,53]. These variables were grouped into four broad categories (Fig 2), categorised or re-categorised as presented in Table 1.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 2. Theoretical framework for home birth in Nigeria.

https://doi.org/10.1371/journal.pone.0337438.g002

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Sample characteristics and home birth in Nigeria by rural and urban residences, NDHS 2018.

https://doi.org/10.1371/journal.pone.0337438.t001

1. (i). External environmental factors included respondents’ place of residence (urban or rural) and regions (northern geopolitical zones: North-West, North-Central and North-East; and southern geopolitical zones: South-South, South-West, and South-East) classified separately (Figs 1 and 2). We followed the same classification of these variables as in the original DHS 2018 data and report [13].
2. (ii). Predisposing factors were classified into socio-demographic characteristics and health knowledge indicators. Socio-demographic characteristics included maternal age, categorised into seven groups: 15–19, 20–24, 25–29, 30–34, 35–39, 40–44, and 45–49 years [13,34,52]. Household wealth, a composite measure of socioeconomic status derived from principal component analysis of household assets, was reclassified from the original five NDHS 2018 categories into three groups for this study: poor (poorest and poorer), middle (middle), and rich (richer and richest) [10,13,35]. Other socio-demographic factors include religion, education level, working status of the mother and her husband/partner, decision-making autonomy (say on own health), ethnicity, birth order, preceding birth interval, and birth type [13]. Health knowledge indicators include frequency of exposure to newspapers, television, radio, and the internet, all of which were categorised as ‘not at all’, ‘less than once a week’, and ‘at least once a week’. The variables were categorised as described in Table 1.
3. (iii). Enabling factors encompass key aspects influencing healthcare access, such as health insurance, antenatal care contacts, financial capacity to pay for services, distance to facility, and difficulties in obtaining permission to seek care. The categories of these variables are shown in Table 1.
4. (iv). Need factors focused on pregnancy intention or desire for pregnancy, representing the mother’s intention at the time of conception.

Further details on the classification and operationalisation of these variables are in Table 1.

Statistical analysis

We assessed both unadjusted (univariable) and adjusted (multivariable) relationships between home birth and explanatory variables using a two-step approach. First, we performed analyses using aggregated data for the overall Nigerian population; second, we disaggregated our data and conducted separate analyses for urban and rural settings. In each setting (overall, rural and urban areas), we estimated the proportion of home births, along with the corresponding 95% confidence intervals (CIs), using frequency tabulation. Unadjusted associations between home birth and each explanatory factor were evaluated using Chi-square tests, and only variables with statistically significant associations (p < 0.05) were advanced to the multivariable analyses. Analyses were restricted to observations with complete data; cases with ‘missing’ or ‘don’t know’ responses were excluded.

For the multivariable analyses, we built logistic regression models for all four variable categories and employed a backward elimination strategy to systematically remove non-significant variables, retaining only those that were statistically significant at 5% level. Before fitting the models, we assessed multicollinearity using Variance Inflation Factors (VIFs) and tolerance values to assess the independence of predictor variables. The VIFs values <10 and tolerance values >0.1 were considered acceptable thresholds [54]. Variable selection for model building was guided by both theoretical relevance (based on the literature) and statistical evidence (significance in our univariable analysis). All statistical analyses and data management were conducted in IBM SPSS Statistics for Windows (Version 21.0). For both univariable and multivariable analyses, we used the Complex Samples module of SPSS to account for sampling weights, stratification, and clustering in the NDHS 2018 design. This approach ensures that our estimates accurately reflect the survey’s complex sampling structure, as well as that standard errors and significance tests reflect the true population variability [55].

To ensure the robustness of the logistic regression model, we applied a model evaluation strategy appropriate for complex survey data. Model fit was assessed using design-adjusted Wald F-statistics and pseudo R-squared values (Cox & Snell, Nagelkerke, McFadden) [56]. Although pseudo R² values do not represent explained variance in the same way as in linear regression, they provide a useful comparative metric for assessing model adequacy. The Hosmer–Lemeshow test was not used, as it assumes simple random sampling [57] and is not supported in SPSS Complex Samples Logistic Regression. We further evaluated the final parsimonious model by examining potential confounders and factors previously reported to be associated with home birth.

Sample characteristics for the overall Nigerian data

The overall weighted sample consisted of 21,512 mothers aged 15–49 years. Regionally, the North-West had the highest representation at 35.5%, while the South-South had the lowest at 8.8% [Table 1]. Nearly half of the mothers (45.2%) had no formal education, while 8.8% had higher education. In terms of age, the cumulative percentage reaches just over 50% in the 25–29 age group. Using the grouped data formula, the estimated median age was 29.9 years. Teenagers comprised 5.6% of the sample, while most mothers were between 20 and 34 years (66.8%), highlighting a predominantly young-middle-aged population. Additionally, 68.1% of the mothers were working, 35.5% were in rich households, while 44.2% were in poor households.

Sample characteristics for rural and urban residences

Rural-urban data disaggregation revealed notable differences (Table 1). Of the total sample, 60.6% (13,045) resided in rural areas and 39.4% (8,467) in urban areas. In rural areas, a higher proportion of mothers had no formal education (60.8%) compared to urban mothers (21.0%). Working status also differed, with 64.4% of rural mothers working compared with 73.8% of urban mothers. However, urban residents typically had higher educational attainment (17.5% with higher education) than their rural counterparts (3.21%), suggesting variations in the type and quality of work. In urban areas, 67.5% of mothers were in the rich category, and only 12.8% were in the poor category. In contrast, just 14.7% of rural mothers were rich, and 64.6% were in the poor category.

Prevalence of home birth in Nigeria by rural and urban residence

Overall, 58.1% (95% CI: 56.5, 59.7) of mothers in Nigeria gave birth at home, while 41.9% (95% CI: 40.3, 43.5, p < 0.001) delivered at a health facility (Fig 3). Home birth prevalence was twice as high in rural (72.4% [95% CI: 70.7, 74.0]) as in urban areas (36.1% [95% CI: 33.6, 38.7], p < 0.001). At the national level, the North-West had the highest home birth prevalence (83.6% [95% CI: 81.5, 85.6]), while the South-East (18.4%; 95% CI: 15.4, 21.9) and South-West (18.4%; 95% CI: 16.0, 21.1, p < 0.001) had the lowest.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 3. Home and health facility birth prevalence in Nigeria by national, rural and urban contexts.

https://doi.org/10.1371/journal.pone.0337438.g003

Considering rural–urban residence-specific differences, the South-East had the lowest home birth prevalence in rural areas (16.2%; 95% CI: 11.0, 23.3), while the South-West had the lowest in urban areas (16.7%; 95% CI: 14.1, 19.7) (Fig 4). Conversely, the North-West recorded the highest prevalence, with 89.4% (95% CI: 87.6, 91.0) in rural areas and 66.6% (95% CI: 60.5, 72.2, p < 0.001) in urban areas. Across all regions, except the South-East, rural areas consistently reported a higher home birth prevalence than urban areas. Notably, in the South-East region, home birth was slightly higher (but not statistically significant) in urban settings (19.2%; 95% CI: 15.7, 23.4) than in rural areas (16.2%; 95% CI: 11.0, 23.3) with overlapping CIs. Overall, the southern regions demonstrated lower home birth prevalence than the northern regions (Fig 4). In all settings, the Hausa-Fulani ethnic group had the highest home birth prevalence (82.7% in the overall, 88.5% in rural and 66.1% in urban areas) compared to other ethnic groups. Also, mothers who had no education, identified as Muslim, had fewer than eight antenatal care contacts, were teenagers, or belonged to the low wealth index category had higher home birth prevalence.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 4. Home birth in Nigeria’s geopolitical zones by national, rural and urban residences.

https://doi.org/10.1371/journal.pone.0337438.g004

Multicollinearity and model fit results

We assessed multicollinearity and model fit in examining predictors of home birth. All VIFs were below five, and all tolerance values exceeded 0.1, indicating no significant multicollinearity among the independent variables [54]. These results suggest that the predictors in the final model are sufficiently distinct and that the estimated coefficients are stable and interpretable. All models also demonstrated statistically significant overall fit (p < 0.001), suggesting that the predictors collectively contributed meaningfully to the likelihood of home birth in each setting. For instance, in the national model, the overall test was significant (F(36, 1278) = 67.511, p < 0.001), indicating strong model performance. Similarly, both the rural (F (30, 746) = 52.10, p < 0.001) and urban (F (35, 504) = 23.68, p < 0.001) models showed significance. The smaller Wald F-statistic values in the rural (52.10) and urban (23.68) models may partly reflect reduced sample size and increased estimation variability in those subpopulations. Regardless, all models remained statistically significant. Pseudo R-squared values supported the Wald F-statistics. For the national model, the Cox & Snell R² was 0.395, the Nagelkerke R² was 0.533, and the McFadden R² was 0.372. The urban and rural models also demonstrated acceptable model performance. The pseudo R-squared values for the residences fall within similar ranges (though slightly lower for rural and urban compared to the national). Notably, McFadden R² values of 0.2–0.4 are typically regarded as indicative of good fit in logistic regression [58]. Together, these results suggest the final models achieved a good balance of explanatory power and parsimony, providing confidence in interpreting the associations.

Factors associated with home birth in the overall Nigerian population

In the multivariable analyses, region of residence (an environmental factor) was significantly associated with home birth. Specifically, mothers residing in the North-East region (AOR = 1.87; 95%CI: 1.34, 2.61), North-West (AOR = 3.30; 95%CI: 2.36, 4.64), and South-South (AOR = 4.54; 95%CI: 3.21, 6.43) had significantly higher odds of home birth compared to the South-West (Table 2). Nine sociodemographic characteristics (predisposing factors) were significantly associated with home birth in the overall Nigerian population (Table 2). For instance, mothers with no education (AOR = 3.18; 95% CI: 2.35, 4.30), and those with primary education (AOR = 2.41; 95% CI: 1.80, 3.23), or secondary education (AOR = 1.87; 95% CI: 1.43, 2.45) had greater odds of home birth in comparison with higher education. A similar pattern of results was observed for the husband/partner’s education level (Table 2).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 2. Factors associated with home birth in Nigeria across the overall, rural and urban areas.

https://doi.org/10.1371/journal.pone.0337438.t002

Additionally, mothers under 35 had higher odds of home birth compared to older age groups. Notably, in a stepwise pattern, mothers from poor households (AOR = 2.19; 95% CI: 1.85–2.59) and those in the middle wealth category (AOR = 1.35; 95% CI: 1.17–1.54) had greater odds of a home birth compared to those from rich households. In addition, giving birth to a singleton, rather than multiples, was associated with increased odds of home birth. Higher birth order, compared to first births, and identifying as Muslim, compared to Christian, were also associated with elevated odds. Belonging to the Hausa-Fulani ethnic group increased the odds of home birth, while Igbo ethnicity was associated with substantially lower odds (Table 2).

Among the health knowledge indicators (predisposing factors), mothers who reported not watching television at all had higher odds of home birth (AOR = 1.42; 95% CI: 1.22, 1.64) compared to those who watched at least once weekly. Similarly, those who did not listen to the radio at all had increased odds of home birth (AOR = 1.36; 95% CI: 1.18, 1.56). Not using the internet at all also increased the odds of home birth (AOR = 1.51; 95% CI: 1.12–2.02).

Lastly, among the enabling factors, mothers who reported challenges in accessing health services recorded higher odds of home birth (Table 2). Specifically, those who found obtaining permission to visit a facility problematic (AOR = 1.22; 95% CI: 1.03, 1.44) and those who described distance to a facility as a ‘big problem’ (AOR = 1.29; 95% CI: 1.14, 1.46) had increased odds of home birth. Finally, ANC contact was inversely related to home birth: mothers who had fewer than seven ANC contacts had markedly higher odds compared to those with eight or more contacts.

Similarities in factors associated with home birth between rural and urban settings

In both rural and urban settings, mothers with no education had significantly higher odds of home birth compared to those with higher education. The effect was slightly stronger in rural areas than in urban areas, but remained comparable (Table 2). Similarly, lower educational attainment at the primary and secondary levels was associated with increased odds of home birth in both settings. In both contexts, mothers from poor households also had greater odds of home birth compared to those from rich households (Table 2). A lack of partner’s education was associated with higher odds of home birth in both settings, with a slightly stronger effect in rural areas (AOR = 2.77; 95% CI: 2.18–3.53) than in urban areas (AOR = 1.57; 95% CI: 1.11–2.21). Mothers who recorded fewer than seven antenatal contacts had higher odds of home birth in both settings, with a stronger association in rural areas (AOR = 3.03; 95% CI: 2.43–3.78) than in urban areas (AOR = 1.91; 95% CI: 1.57–2.32). Additionally, higher birth order, Hausa-Fulani ethnicity, singleton births, and limited media exposure (TV, radio, and internet) were common factors associated with home birth across both rural and urban areas (Table 2).

Differences in factors associated with home birth between rural and urban settings

Notable differences emerged when comparing rural and urban residences. First, the odds of home birth were significantly higher in four rural regions: North-East, North-Central, North-West, and South-South, compared to the South-West, and by extension, the South-East. That is, in rural Nigeria, all the northern regions (North-Central, North-West and North-East), and only one southern region (South-South) had higher odds of home birth. Conversely, in urban areas, higher odds were observed in only two regions: one from the northern area (North-West) and one from the southern area (South-South). Regional disparities were more pronounced in rural areas. For example, the North-West region was associated with higher odds of home birth in both settings, but the effect was over twice as strong in rural (AOR = 5.83; 95% CI: 3.43–9.90) than urban areas (AOR = 2.09; 95% CI: 1.38–3.17). Similarly, in the South-South region, rural mothers had nearly three-fold higher odds (AOR = 7.82; 95% CI: 4.55–13.44) compared to urban counterparts (AOR = 2.71; 95% CI: 1.74–4.20).

Secondly, certain enabling factors were uniquely associated with rural areas (not significant in urban areas). Mothers who had difficulty obtaining permission to visit a health facility had higher odds of home birth (AOR = 1.45; 95% CI: 1.15–1.82), as did those reporting distance as a major barrier (AOR = 1.37; 95% CI: 1.18–1.60). Home birth odds were also higher among mothers not involved in healthcare decision-making. Specifically, those deciding alone or jointly with partners had lower odds of home birth, a pattern observed only in rural areas (Table 2). Thirdly, in urban areas, maternal age, religion, and financial access were uniquely associated with home birth. Younger mothers, particularly those aged 15–19 (AOR = 2.19; 95% CI: 1.25–3.82) and 20–24 (AOR = 1.79; 95% CI: 1.21–2.65), had higher odds compared to older groups. Religion was significant only in urban settings; Muslim mothers had higher odds than Christians (AOR = 1.55; 95% CI: 1.19–2.02). Financial constraints also mattered in urban areas, where mothers in the middle wealth index (AOR = 1.46; 95% CI: 1.22–1.75) and those who reported difficulty affording care (AOR = 1.25; 95% CI: 1.07–1.47) had higher odds of home birth.

Policy and programmatic implications

Our study highlights the need for targeted, equity-oriented policies to address home births in Nigeria, framed through Andersen’s model and a social justice perspective. This approach recognises that addressing predisposing, enabling, and environmental barriers requires interventions that are both feasible and responsive to systemic inequities. First, our findings implicate maternal and husband education as critical socio-demographic predisposing factors. Hence, it is crucial to prioritise policies that promote educational opportunities, particularly in rural, all northern and South-South regions, where home birth prevalence and odds were highest. Expanding access to quality education for girls is a critical first step. Improving maternal health literacy is also essential, enabling women to understand birth complications and make informed healthcare decisions. From a social justice perspective, empowering girls and women through formal education and health literacy addresses long-standing systemic inequities and strengthens their capacity to make autonomous health choices [89]. Community-driven initiatives, such as women’s support groups, mobile health messaging, and the engagement of community health volunteers, can complement formal education by providing culturally appropriate health education and encouraging shifts toward skilled birth attendance.

Second, expanding access to communication infrastructure nationally should be a policy priority to ensure that health messages reach all populations. Our findings highlight the value of increased access to radio, television, and the internet, which serve as low-cost predisposing indicators for health information sharing across all settings in Nigeria. Policies should actively leverage these media to disseminate culturally appropriate maternal health messages, particularly targeting marginalised and underserved populations. Complementary approaches, including leveraging local community networks and engaging local champions, have proven effective in reaching marginalised populations [90,91]. For instance, lessons from the Abiye program (in Ondo state, Nigeria) demonstrate that community involvement and advocacy improved the uptake and sustainability of facility-based childbirth, helping to reduce rural–urban disparities [90,91]. Scaling such approaches nationally could strengthen health communication and reinforce behaviour change toward safer birth practices.

Third, addressing Andersen’s environmental factors, particularly regional and rural–urban disparities, requires targeted structural action. In the North, interventions should combine infrastructure development, improved facility access, and enhanced security. In the South-South, policies should prioritise hard-to-reach areas such as riverine locations and address environmental degradation. Medium- to long-term priorities include investing in transportation networks, road infrastructure, and well-equipped healthcare centres across underserved regions. Short-term strategies, such as mobile health services and functional ambulances, can reach remote communities and enable timely transfers. Strengthening the health workforce and ensuring equitable distribution of skilled birth attendants remain essential. Yet improving access alone is insufficient; the quality of care, especially emergency obstetric services, must also improve [14,65]. The MSS offers a practical model, showing the effectiveness of deploying skilled personnel to underserved areas [15,19]. A revitalised MSS, integrated within UHC reforms, supported by sustainable financing, fair remuneration, and clear career pathways, would be feasible and equity-enhancing.

Fourth, targeted financial interventions are essential to address economic barriers that may contribute to home births in both rural and urban settings, including among middle-income households. Policies should expand and sustain maternal health subsidies to improve affordability and promote equitable access, for instance, through an expanded National Health Insurance Scheme (NHIS). Lessons from previous initiatives, such as the Abiye program and the CCT scheme [15–18,20,90,91], highlight the potential of demand-side financing to increase facility-based childbirth when adequately funded and monitored. Community-based financing models, including women’s savings groups, micro-insurance, and voucher programs, can complement national strategies. Effective implementation requires strong governance, transparency, and integration within broader health financing reforms. The experience of the MSS, which suffered from limited collaboration and inconsistent funding, underscores the need for predictable financing and sustained political commitment to ensure program success [15,19].

Fifth, promoting maternal autonomy is vital, especially in rural areas. Engaging male partners and community networks in maternal health education can create supportive environments [85]. Programs that challenge restrictive gender norms and encourage shared decision-making are crucial for improving maternal health outcomes. Potential strategies could include participatory workshops, peer-support networks, and male engagement initiatives. These intervention efforts can help advance social justice by addressing power imbalances, enhancing maternal agency, and ensuring that marginalised women can make informed choices.

Finally, our findings call for regionally tailored, culturally sensitive strategies addressing Nigeria’s distinct inequalities. In rural areas, policies should prioritise physical and informational access and women’s empowerment [85]. Targeted investments in regions such as the North, where the Hausa-Fulani ethnic group is concentrated and home birth is high, are essential. In urban areas, focus should be on financial accessibility and culturally sensitive care, particularly for Muslim women who may face barriers related to privacy and gender preferences [86,87]. Faith-based approaches could help reduce home births among Muslims and other groups [88]. Across Nigeria, poor maternal outcomes are driven by systemic gaps undermining UHC. The Abiye Programme’s success highlights that political will, sustainable financing, and community ownership are key [90,91]. Framing interventions through a social justice lens ensures prioritisation of the most marginalised: rural mothers with geographic or autonomy constraints, poor mothers, urban middle-wealth mothers, Muslim mothers in urban areas, younger urban mothers, and mothers with limited education, ANC, or health information access. Future research should evaluate the feasibility, scalability, and sustainability to ensure place-based innovations equitably reduce home births and maternal mortality.

Strengths and limitations

This study demonstrates key strengths, including analysis at national, urban, and rural levels, which clarifies disparities in home birth practice in Nigeria. The findings offer evidence-based insights for policies to reduce home births in Nigeria and are generalizable to the larger population due to the nationally representative sample. However, there are limitations to consider. Firstly, data collection relied on mothers’ recall ability; while it would be minimal given the substantial nature of the event, there appears to be a possibility of recall bias. Secondly, due to data limitations, the study did not include potential culture-related factors that could influence home births in Nigeria. However, we captured wide-ranging factors based on the NDHS dataset, providing comprehensive insights. The dataset used is over five years old and may not accurately reflect the current situation. Nonetheless, it remains the most recent edition and our findings serve as valuable evidence for further research on home births in Nigeria and other low- to middle-income countries. Finally, the data utilised were cross-sectional; hence, causality cannot be established, limiting the scope of our conclusions.