Examining the Research Gap in Causal Inference Methods for Health Policy Evaluation in Nigeria

Rigorous causal inference is essential for evaluating whether health programmes actually improve outcomes rather than merely correlating with them, yet advanced causal inference methods remain underutilised in Nigerian health policy analytics, creating a gap between methodological advances and practical application. This study examined the research gap in causal inference for health policy evaluation in Nigeria through systematic review and illustrative analysis. A systematic scoping review identified 126 publications from 2019 to 2024 on health programme evaluations in Nigeria, of which 21 applied any form of causal inference method. Difference-in-differences was the most common method (11 studies), while instrumental variable analysis, regression discontinuity, and synthetic control methods were applied in one, three, and zero studies respectively. An illustrative analysis using difference-in-differences re-estimated the effect of the National Health Insurance Scheme expansion on antenatal care utilisation in participating versus non-participating LGAs, using NHMIS data from 2018 to 2022. The DiD estimate confirmed a significant positive programme effect (ATT = +14.2 percentage points, 95 percent CI: 9.8 to 18.6), larger than naive cross-sectional estimates had suggested. Parallel trends testing and placebo tests confirmed the DiD assumptions. The study identifies three methodological gaps: absent use of synthetic control for state-level policy evaluation, limited regression discontinuity applications despite programme eligibility thresholds, and inadequate data infrastructure for longitudinal causal studies. Recommendations include a NHP-sponsored causal inference training programme for government health data analysts.

Keywords: causal inference, health policy evaluation, Nigeria, difference-in-differences, NHIS