Individual and community-level factors associated with non-institutional delivery of women of childbearing-age in Nigeria

Maternal mortality is a persistent public health challenge, particularly in sub-Saharan Africa, which accounts for about two-thirds of global maternal deaths. Health facility delivery is a key strategy to reduce maternal mortality, yet many women in sub-Saharan Africa, including Nigeria, still face barriers to accessing modern maternal care. Nigeria’s maternal mortality rate is high (906 per 100,000 live births), with a large share of births occurring at home and without skilled attendants. While prior Nigerian studies have examined determinants of health facility delivery, few have incorporated community contextual factors when investigating non-institutional delivery. This study addresses that gap by exploring individual, household, and community-level determinants of non-institutional delivery among Nigerian women of reproductive age, with the aim of informing strategies to reduce maternal mortality and meet SDG 3.1 targets by 2030.

Prior studies link non-institutional delivery to socioeconomic and geographic disparities, including low education, poverty, and rural residence. Research from various countries (e.g., Nepal, Tanzania, Cambodia, Ghana) shows substantial rates of home delivery, often driven by financial and regional constraints that limit access to standard health services. Other studies indicate that uneducated, poor, and rural women may lack awareness of the risks associated with home delivery. Media exposure has been associated with greater awareness of pregnancy-related complications and increased utilization of facility delivery. Despite this evidence, Nigerian research has rarely integrated community-level contextual factors such as community poverty, media exposure, cultural norms, and women’s autonomy when analyzing non-institutional delivery, a gap this study seeks to fill.

Data source: 2018 Nigeria National Demographic and Health Survey (NDHS), publicly available. Study population: women aged 15–49 who had given birth in the five years preceding the survey; weighted sample size N = 12,567. Study design: cross-sectional. Outcome variable: non-institutional delivery, coded 1 for deliveries outside health institutions and 0 for deliveries in health institutions (government or private facilities). Individual-level variables: maternal age (15–19, 20–24, 25–29, 30–34, 35–39, 40–44, 45–49), religion (Catholic, other Christians, Islam, African traditional/others), maternal education (none, primary, secondary, post-secondary), partner’s education (none, primary, secondary, post-secondary), health insurance (yes/no), employment status (working/not working), timing of ANC initiation (early vs late booking), media exposure (any of TV/radio/newspaper; yes/no), woman’s autonomy in healthcare decision-making (autonomous if respondent alone or respondent with partner; otherwise non-autonomous). Household-level variables: sex of household head (male/female), household size (1–4, 5–7, ≥8), wealth index quintiles (poorest, poorer, average, wealthier, wealthiest) derived using principal components analysis of household assets. Community-level variables (constructed from enumeration areas): residence (urban/rural), cultural norms justifying wife-beating (≥50% vs <50% justification), share of women with no education (≥50% vs <50%), community-level poverty (high vs low), community-level media exposure (high vs low), community-level women’s autonomy (high vs low). Statistical analysis: Descriptive statistics reported prevalence of non-institutional delivery; chi-square tests examined associations with sociodemographic variables (significant variables at 10% retained). Multilevel binary logistic regression modeled fixed and random effects with three levels: individual (level 1), household (level 2), and community (level 3). Five models were estimated: Model 0 (empty model), Model 1 (individual factors), Model 2 (household factors), Model 3 (community factors), and Model 4 (full model including individual, household, and community factors). Results reported as adjusted odds ratios (AOR) with 95% confidence intervals; model fit assessed with ICC, AIC, BIC, and log-likelihood. Analyses conducted in Stata.

- Prevalence: 58.6% of childbearing women delivered outside health institutions (home delivery). Group-specific prevalence included 64.3% among Muslim women; 85.8% among mothers with no formal education; and 50% among mothers whose partners had no formal education. Home delivery was higher among women without health insurance, those not working, those initiating ANC late, women in polygamous unions, women without media exposure or autonomy, those in the poorest wealth quintiles, households with ≥8 members, and rural residents. Communities with high acceptance of wife-beating, low media use, high poverty, and low women’s autonomy showed higher home delivery prevalence. - Multilevel model fit: The full model (controlling for individual, household, and community factors) had the best fit (lowest AIC = 987; ICC around 0.59), indicating substantial between-cluster variance. - Significant predictors (fixed effects): • Maternal education: Compared to no education, primary (AOR≈0.34), secondary (AOR≈0.56), and post-secondary (AOR≈0.45) education were associated with lower odds of non-institutional delivery. • Partner’s education: Compared to no education, primary (AOR≈0.12), secondary (AOR≈0.34), and post-secondary (AOR≈0.21) were associated with lower odds of non-institutional delivery. • Media exposure: No media exposure increased odds of non-institutional delivery (AOR≈2.89) relative to exposure. • Wealth: Wealthiest quintile associated with lower odds vs poorest (AOR≈0.48). • Residence: Rural residence increased odds vs urban (AOR≈2.75). • Community poverty: High community-level poverty increased odds vs low (AOR≈2.21). • Community media use: High media use decreased odds vs low (AOR≈0.98; reported significant). • Community women’s autonomy: High autonomy decreased odds vs low (AOR≈0.35).

The study’s findings indicate that both individual and contextual factors significantly influence the likelihood of non-institutional delivery in Nigeria. Educational attainment (of women and their partners), higher wealth, urban residence, and media exposure reduce the propensity for home delivery, likely through improved knowledge of risks, greater autonomy, and better access to health services. Conversely, poverty—both at the household and community levels—rural residence, and low media exposure are linked to higher odds of home delivery, reflecting socioeconomic and infrastructural barriers and prevailing community norms that can normalize home births. These results align with prior evidence from Nigeria and other low- and middle-income settings, reinforcing the importance of addressing socioeconomic disparities, enhancing access to information, and improving the distribution and quality of maternal health services, particularly in rural and impoverished communities. The influence of community-level media exposure and women’s autonomy underscores the role of social norms and information ecosystems in shaping delivery choices, suggesting that community-based interventions can meaningfully shift behavior toward facility-based deliveries.

Non-institutional delivery among Nigerian women of reproductive age is prevalent and associated with low socioeconomic status, limited education, low media exposure, rural residence, high community-level poverty, low community media use, and low women’s autonomy. To improve facility-based delivery rates and advance toward SDG 3.1 targets, the study recommends expanding educational opportunities for women, strengthening advocacy and community education (e.g., focus groups, peer education, mentor-mentee programs), increasing media access and health information dissemination, reducing financial barriers to facility delivery, and expanding and equitably distributing health facilities—especially in rural areas. Future research should incorporate qualitative methods to deepen understanding of cultural and contextual drivers of home delivery decisions.