Infectious Disease
IV+Survival with IPDMA data in outbreaks Heather* Heather Hufstedler
Quasi-experimental methods are increasingly utilized in infectious disease research to address confounding and establish causal relationships. While instrumental variable analysis has been combined with survival models in prior work, its application within the context of individual participant data meta-analyses (IPDMAs) remains unexplored.
In this study, we propose a novel methodological framework that integrates IV analysis with survival outcomes in an IPDMA setting. We develop and compare frequentist and Bayesian implementations, exploring their strengths and limitations. A secondary aim is to evaluate the practicality of this approach in low-resource settings, where data collection challenges can further complicate analysis. This application demonstrates the method’s potential in real-world scenarios characterized by incomplete or heterogeneous data, in addition to its inherent challenges. This work is part of ongoing research.