chinembiri24500/NCD_HIV_SharedComponent_Model

Bayesian spatio-temporal modelling of HIV and NCD mortality in Sub-Saharan Africa using INLA and BYM2 models # SADC HIV-NCD Bayesian spatial analysis This repository contains the R code used for the revised manuscript: **Uncovering Spatial Co-dependence and Potential Syndemic Overlap: A Bayesian Stacked-Likelihood Model of HIV Prevalence and NCD Mortality in Southern Africa** The analysis estimates a primary stacked-likelihood Bayesian areal model and a shared-component sensitivity model for HIV prevalence and major NCD mortality in 13 Southern African Development Community countries, 2000-2019. ## Repository structure ```text . ├── sadc_hiv_ncd_inla_analysis.R ├── README.md ├── data/ │ ├── sadc2025.csv │ ├── HIV_df.csv │ ├── sadc.shp │ ├── sadc.dbf │ ├── sadc.shx │ └── sadc.prj └── outputs/ # created by the script ├── figures/ ├── tables/ └── models/ ``` ## Required input files The script expects the following files in a `data/` folder: 1. `sadc2025.csv`: NCD panel data. Required columns are: - `country` - `year` - `observed_deaths` - `expected_deaths` - `ncdstype` - `gdpercpta` - `urbrte` - `hcexp` - `avrgprecp` - `avrgtemp` 2. `HIV_df.csv`: HIV prevalence data. Required columns are: - `country` - `year` - `hiv_prev` 3. `sadc.shp` and associated shapefile sidecar files. The shapefile must contain a `country` column. ## R version and packages The script requires R >= 4.1 because it uses the base R pipe operator (`|>`). Required R packages: - `INLA` - `sf` - `spdep` - `dplyr` - `ggplot2` - `readr` - `stringr` - `tmap` - `scales` - `tidyr` - `purrr` - `tibble` - `data.table` - `grid` INLA is not installed from CRAN. It can be installed using: ```r install.packages( "INLA", repos = c(getOption("repos"), INLA = "inla.r-inla-download.org"), dep = TRUE ) ``` ## How to run Place the input files in the `data/` folder, set the working directory to the repository root, and run: ```r source("sadc_hiv_ncd_inla_analysis.R") ``` The script creates the `outputs/` folder automatically. ## Main analytical steps The script …