01 / Continuum dropout
Coverage falls at every stage.
Interpretation: measuring individual service contacts alone can overstate continuity across the full maternal-care pathway.
← Back to selected work
Health data analytics · maternal health · Nigeria
Maternal Continuum of Care in Nigeria
An end-to-end analysis of how women move from antenatal care through skilled delivery and early postnatal care, using the 2024 Nigeria Demographic and Health Survey.
Decision problem
Individual indicators can hide losses across the full care pathway.
Maternal care is not a single contact with the health system. A woman may begin antenatal care but fail to receive enough contacts, deliver without skilled attendance, or miss timely postnatal care.
The project asks how much of the pathway is completed nationally, where the largest drop-offs occur, how unequal completion is across socioeconomic and geographic groups, and which characteristics remain associated with completion after adjustment.
Primary outcome: at least four antenatal care contacts, skilled birth attendance, and early postnatal care for both mother and newborn within two days of birth.
Executive view
Only one in four completed the primary continuum.
73.7%Any antenatal care
53.6%Four or more ANC contacts
45.1%Skilled birth attendance
24.5%Complete primary continuum
Analytical architecture
From restricted microdata to a governed reporting layer.
The workflow separates respondent-level analysis from public reporting outputs so the project remains reproducible without distributing restricted DHS microdata.
01R: import, audit & cohort derivation
02Survey-weighted national, equity & geographic analysis
03Regression & sensitivity analysis
04PostgreSQL staging, analytics & reporting views
05Power BI semantic model & reporting
Findings
The national average hides substantial dropout and inequality.
02 / Equity
Completion gaps are wide across social and geographic groups.
63.2ppEducation gap
61.2ppWealth gap
53.3ppGeopolitical-zone gap
30.0ppResidence gap
Interpretation: national coverage improvement alone would not guarantee equitable continuity of care.
03 / Geography
State-level completion ranges from 72.7% to 3.3%.
In the public reporting output, complete-continuum coverage ranged from 72.7% in Lagos to 3.3% in Kebbi.
Interpretation: the geographic spread is large enough that programme design and monitoring should not rely on national averages alone.
04 / Adjusted associations
Education and wealth remain strongly associated with completion.
In the adjusted survey-weighted model, higher education had an adjusted odds ratio of 6.32 versus no education, while the richest wealth quintile had an adjusted odds ratio of 6.31 versus the poorest quintile.
Important: these are associations from cross-sectional data and should not be interpreted as causal effects.
05 / Sensitivity
The story holds under stricter definitions.
Replacing ANC4 with ANC8 reduced continuum prevalence to 11.0%. Replacing skilled birth with facility delivery produced a prevalence of 24.6%.
Interpretation: the headline conclusion is not dependent on a single operational definition.
Quality & governance
Reproducibility is part of the analysis, not an afterthought.
The final R reconciliation reported 0 missing required analytical files, 0 invalid national estimates or CI-ordering problems, 37 state/FCT units represented, 296/296 expected state-indicator estimates generated and 0 critical historical QA failures.
Respondent-level NDHS microdata are not distributed. Public reporting uses validated aggregate outputs, with small-cell suppression and caution rules applied before version control.
Inspect the work
The full analytical workflow is documented.
The repository contains the R pipeline, PostgreSQL architecture, QA, safe aggregate outputs, PBIP source, methods documentation and reproducibility notes.