On April 27th, Daniel Dempsey presented a novel method for the analysis of ANCA vasculitis flare using Bayesian distributed lag regression. Details for Daniel’s talk are below.
Title
Bayesian Distributed Lag Regression and Model Selection Method for the Analysis of ANCA Vasculitis Flares.
Abstract
ANCA vasculitis is an autoimmune disease characterised by relapses, or flares, that can have a severe detrimental impact on a patient’s health. Flares can be prevented by suppressing the immune system but this obviously exposes the patient to infection, resulting in a dangerous balancing act. This is made more difficult by the fact that researchers are still unclear on what causes flare events. While much research has been conducted on possible genetic explanations, there has been less focus on the environment and this is what we aim to address.
Our statistical method uses an aggregation of distributed lags, similar to the MIDAS model in econometric literature, to model the accumulation of environmental exposure over time in a parsimonious manner. This is incorporated into a quantile regression framework to handle the large degree of response imbalance. A covariate indicator is also included to quantify our posterior belief that the corresponding covariate should be included in the model. We collect samples from the posterior using Reversible-Jump MCMC.
The method is validated via simulation study, and then applied to real data comprising of clinical information for flare events and satellite data that tracks weather and pollution indices for the region of residence of each patient. Despite our focus on vasculitis, we believe this model is applicable to many similar research problems.