Hierarchical logistic regression analysis of skilled birth attendance in Northern Nigeria using Andersen’s Behavioural Model

Abstract Background In low and middle-income countries, maternal and newborn mortality continue to be major challenges for public health. In Nigeria, the Northwest region remains to record some of the highest maternal and child mortality rates despite improvements in antenatal care coverage. But the utilization of expert birth services remains suboptimal, with many deliveries still occurring outside health facilities. This study examined the determinants of skilled birth service utilization among women of reproductive age in Northern Nigeria. Methods A cross-sectional analytical study was conducted among 1,004 women aged 15–49 years who had delivered within the previous five years across six northern Nigerian states (Jigawa, Bauchi, Niger, Katsina, Kaduna, and Kano). Participants were selected using purposive and quota sampling techniques from both rural and urban communities. Data was collected using a pretested interviewer-administered questionnaire and analyzed using descriptive statistics and hierarchical logistic regression. The analysis was guided by Bivariate associations were examined using Pearson’s chi-square tests and Andersen’s Behavioral Model of Health Service Use, which categorizes determinants into predisposing, enabling, and need factors. Two outcome variables were examined: skilled birth attendance (SBA), defined as delivery assisted by a trained health professional regardless of location, and place of delivery, defined as delivery at a health facility versus home. Separate hierarchical logistic regression models were specified for each outcome. Results Overall, 72.8% of respondents reported receiving skilled birth assistance, while 27.2% did not. At the bivariate level, phone ownership, ANC satisfaction, and previous place of delivery were significantly associated with healthcare delivery assistance. After multivariable adjustment, partner’s education and previous place of delivery remained independently associated with healthcare delivery assistance, whereas the bivariate associations observed for phone ownership and ANC satisfaction were attenuated and were no longer statistically significant. For place of delivery, education, year of last delivery, partner’s employment, settlement type, mode of transportation, and phone ownership were significantly associated at the bivariate level. After adjustment, year of last delivery and phone ownership remained statistically associated with place of delivery, whereas education, partner’s employment, settlement type, and mode of transportation were no longer statistically significant overall. Category-specific associations for year of last delivery were observed for 2021, 2022, and 2024. Conclusion Policies aimed at improving maternal health outcomes should prioritize strengthening primary healthcare infrastructure, expanding financial protection mechanisms, improving transportation and referral systems, and improving the standard of prenatal care services. Addressing these structural and socioeconomic barriers is essential to increase facility-based deliveries and reducing maternal and neonatal mortality.