Estimating Discontinuous Time-Varying Risk Factors and Treatment Benefits for COVID-19 with Interpretable ML

TL;DR

This paper introduces a method using tree-based generalized additive models to estimate discontinuous, time-varying risk factors and treatment benefits in COVID-19, revealing changing biomarker importance over the pandemic.

Contribution

It presents a novel approach for modeling discontinuous, time-varying effects in observational data, specifically applied to COVID-19 risk factors and treatment outcomes.

Findings

Thrombosis biomarkers increasingly predicted mortality over time.

Inflammation biomarkers' association with thrombosis weakened.

Methodology can estimate unknown, discontinuous effects in other contexts.

Abstract

Treatment protocols, disease understanding, and viral characteristics changed over the course of the COVID-19 pandemic; as a result, the risks associated with patient comorbidities and biomarkers also changed. We add to the conversation regarding inflammation, hemostasis and vascular function in COVID-19 by performing a time-varying observational analysis of over 4000 patients hospitalized for COVID-19 in a New York City hospital system from March 2020 to August 2021. To perform this analysis, we apply tree-based generalized additive models with temporal interactions which recover discontinuous risk changes caused by discrete protocols changes. We find that the biomarkers of thrombosis increasingly predicted mortality from March 2020 to August 2021, while the association between biomarkers of inflammation and thrombosis weakened. Beyond COVID-19, this presents a straightforward methodology to estimate unknown and discontinuous time-varying effects.