Sedem-77/health-financing-mortality-africa

We use generalized additive mixed models (GAMMs) to allow for potential non-linear relationships between crude mortality rates and public health expenditure, out-of-pocket payments, political stability, and age dependency across 51 African countries from 2002 to 2022 (N = 1,071 country-year observations). # Health Financing and Mortality in Africa, 2002-2022 This repository contains the data, statistical analysis code, and a static dashboard associated with the manuscript: > **Beyond Spending: Nonlinear Impacts of Health Financing and Out-of-Pocket > Costs on Mortality in Africa** > > Denis Folitse, Christopher Odoom, Alexander Boateng, and Luis Alberiko > Gil-Alaña The manuscript has been submitted to **Health Economics (Springer Nature)**. The study examines how domestic government health expenditure, out-of-pocket health spending, political stability, and age dependency are associated with crude mortality rates. It uses a country-year panel covering 51 African countries over 2002-2022 and applies descriptive analysis, linear regression, and generalized additive models (GAMs) to evaluate potentially nonlinear relationships. > **Submission status:** The manuscript is under journal review. Results in > this repository should be treated as pre-publication findings and may change > during peer review. ## Research Objectives The analysis is designed to investigate: 1. How mortality and health-financing indicators changed across African countries and regions between 2002 and 2022. 2. Whether government health expenditure per capita, out-of-pocket spending, and political stability are associated with crude mortality rates. 3. Whether these associations are nonlinear or vary by African region. 4. Whether the relationship between government health expenditure and mortality depends on the level of out-of-pocket spending. 5. Whether the observed relationships remain after accounting for age dependency, country-level heterogeneity, time trends, lagged exposures, and the COVID-19 period. The study is observational. Associations estimated by the models should not be interpreted as causal effects. ## Study Design - **Unit of analysis:** country-year - **Geographic coverage:** 51 African countries grouped into Central, East, Northern, Southern, and West regions - **St …