What stops Nigerian women from using maternity care?
Whether Nigerian women used maternity care depended not only on their own education and wealth but also on their community and the state they lived in.
Source
Determinants of use of maternal health services in Nigeria--looking beyond individual and household factors
Study at a glance
- Design
- Cross-sectional — Secondary analysis of the 2005 Nigerian National HIV/AIDS and Reproductive Health Survey with multilevel logistic models (individual, household, community and state fixed effects plus state random effects).
- N
- N=2148 · 2,148 women aged 15-49 who had a baby in the five years before the survey, drawn from 4,685 surveyed women in all 36 states; each outcome refers to the most recent birth.
- Population
- Women of reproductive age across Nigeria with a recent birth.
- Outcome
- Use of antenatal care, delivery assisted by a doctor, nurse or midwife, and postnatal care.
Structured fields used in claim comparison tables when every cited study has a complete layer.
What they did
The authors analysed a 2005 national household survey covering all 36 Nigerian states, focusing on 2,148 women who had given birth in the previous five years. For the most recent birth they looked at three outcomes: any antenatal care, delivery assisted by trained medical staff, and postnatal care. Multilevel models tested individual and household factors (education, age, ethnicity, family-planning attitudes, wealth), community factors (urban residence, local media reach, how common small-family norms were) and a state measure of residents per primary health care facility, while allowing for unmeasured differences between states.
What they found
Use was low: 60.3% had antenatal care, 43.4% had skilled help at delivery and 41.2% received postnatal care. Before adding predictors, 36.8% of the variation in antenatal care use lay between states. Education, household wealth and urban residence predicted all three services; women with post-secondary education had about five times the odds of antenatal care, and the richest women almost six times the odds of the poorest. Community media reach and small-family norms mattered for delivery care, more residents per primary care facility meant less antenatal and delivery care, and significant unexplained state differences remained after all adjustments.
The limits
What it doesn't show
The data are cross-sectional, so the authors cannot rule out confounding, reverse causation or selection; they note they did not use methods to correct for endogeneity. Service use was self-reported without checking clinic cards, and key access factors such as distance to the nearest facility were not available, so the state ratio of residents to clinics is only a rough proxy. The survey dates from 2005, and the paper identifies associations to target rather than testing any intervention.
Key terms
- Multilevel model
- A regression model for nested data (women within households, communities and states) that separates effects at each level and gives correct standard errors.
- Intraclass correlation (ICC)
- The share of total variation in an outcome that lies between groups (here, states) rather than between individuals within groups.
- Random effect
- A term that captures unmeasured differences between groups, such as states, allowing each to have its own baseline level of service use.
- Skilled birth attendant
- A doctor, nurse or midwife trained to manage normal deliveries and recognise complications.
- Community media saturation
- How widely media such as radio and TV reach people in a woman's local government area, used as a community-level exposure.
Flashcards
0 of 11 answers reviewed
Research intelligence for this paper
See its role on concept claims, tensions it is part of, placement history, and related discoveries.
Quiz yourself
About what share of women used antenatal care for their most recent birth?
Common questions
Why use a multilevel model instead of ordinary logistic regression?
Women in the same community or state share conditions, so their answers aren't independent. Ignoring that nesting can make standard errors too small and misattribute community effects to individuals; a multilevel model handles both problems.
Why was antenatal care more common than skilled delivery?
Antenatal visits can be planned, but labour starts unpredictably, and poor roads, transport and understaffed clinics make reaching skilled help at that moment harder. The stronger effect of clinic density on delivery care fits this explanation.
What does the leftover state-level variation mean for policy?
Even after accounting for education, wealth and community factors, states still differed, so unmeasured state-specific factors matter and programmes should be tailored to each state.
More on Maternal health