WSzegda-UMich/BDSY-PHM-Malawi-Project

This repository contains all relevant R files used by the Malawi subgroup of the public health modeling group at the Big Data Summer Immersion at Yale. # Impact of 13-Valent Pneumococcal Conjugate Vaccine (PCV13) on Invasive Pneumococcal Disease (IPD) in Malawi Anna Abraham, Daniel Schiffman, and William Szegda ## Abstract Pneumococcal conjugate vaccines (PCVs) have significantly reduced the global burden of invasive pneumococcal disease (IPD). In this project, we model the overall and serotype-specific impacts of the 13-valent PCV (PCV13) in Malawi, an HIV-prevalent nation in sub-Saharan Africa. ### Overview - **Objective**: To assess the overall and serotype-specific effects of PCV13 on IPD in Malawi. - **Methods**: - **Time Series Analysis**: We examine aggregate IPD incidence data and incorporate HIV-related covariates (antiretroviral therapy coverage) to better isolate the vaccine’s impact. - **Bayesian Hierarchical Modeling**: Using data from the Global Pneumococcal Sequencing Project, we employ Bayesian hierarchical modeling to stabilize trends for vaccine-targeted serotypes with small case counts and look into serotype-specific trends. - **Market Basket Analysis**: We assess potential serotype-switching patterns post-vaccine introduction. ### Key Findings - **Overall trend**: - By controlling for antiretroviral therapy (ART) coverage in our time series model we better isolate the impact of PCV13 - **Serotype-Specific Impact**: - Serotypes 3 and 19F were **not well targeted** by PCV13 according to our models. - Serotypes 5, 6A, and 6B were **better targeted** by PCV13 according to our models. - **Serotype-Switching**: - We identified **three GPSCs** for which serotype switching may have occurred following the vaccine’s introduction. ### Conclusion Through a combination of modeling approaches, we provide insights into both the overall and serotype-specific impacts of PCV13 in Malawi. These findings could inform future vaccine policy and government expectations for vaccine rollouts in similar HIV-prevalent, sub-Saharan African settings and modeling which serotypes are not well-targeted and seroty …