This repository contains code necessary to produce findings associated with the manuscript "Surveillance for TB drug resistance using routine rapid diagnostic testing data: Methodological development and application in Brazil". A pre-print version of this manuscript can be found on medrxiv: https://www.medrxiv.org/content/10.1101/2024.07.22.24310845v1
This project uses routine Xpert MTB/RIF data to estimate levels and trends in Rifampin-resistant Tuberculosis (RR-TB) using a hierarchical generalized additive model in R. We apply this model to Brazil, using patient-level TB data from their National Disease Notification Registry (SINAN).
master_run_all.R will run all sub-directories to prepare analytic datasets, run models, and produce figures and tables. Each sub-directory contains a 01_run_all.R that will run components of that directory. Note that if running within sub-directory, you will need to manually set the version for which files to load and save (e.g. file_version_load, file_version_save). Otherwise, these can be set in master_run_all.R.
-
01_data_processing - 01_run_all.R cleans SINAN data and prepares analytics datasets for model, by sourcing relevant scripts in folder.
-
02_run_models - 05_run_all.R runs HGAM models and gets the associated fitted values, uncertainty intervals, and compiles modeled and observed estimates. Results are compiled at the national- and state-level by year and by quarter.
-
03_make_figures_and_tables - 01_run_all.R reproduces figures and table in manuscript. manuscript_calculations.R produces values for in-text references of results.
See dependencies.R for packages necessary to run code.
- Patient-level TB data from Brazil's national Notifiable Diseases Registry, Sistema de Informação de Agravos de Notificação (SINAN)
- Municipality centroid shapefile (IBGE)
- National and state-level population denominators from 2010 Census (IBGE)
- Region associated with each state.