Learning Clinical Concepts for Predicting Risk of Progression to Severe COVID-19

TL;DR

This study develops and compares high-performance survival models for predicting severe COVID-19 progression using comprehensive features and clinically interpretable concepts, outperforming previous models.

Contribution

Introduces a pipeline that learns clinical concepts to enhance risk prediction models for severe COVID-19, balancing accuracy and interpretability.

Findings

Concept-based models outperform feature-based models (C-index 0.858 vs. 0.844).

Models outperform previous studies (C-index 0.844-0.872 vs. 0.598-0.810).

Concept learning improves out-of-sample prediction accuracy.

Abstract

With COVID-19 now pervasive, identification of high-risk individuals is crucial. Using data from a major healthcare provider in Southwestern Pennsylvania, we develop survival models predicting severe COVID-19 progression. In this endeavor, we face a tradeoff between more accurate models relying on many features and less accurate models relying on a few features aligned with clinician intuition. Complicating matters, many EHR features tend to be under-coded, degrading the accuracy of smaller models. In this study, we develop two sets of high-performance risk scores: (i) an unconstrained model built from all available features; and (ii) a pipeline that learns a small set of clinical concepts before training a risk predictor. Learned concepts boost performance over the corresponding features (C-index 0.858 vs. 0.844) and demonstrate improvements over (i) when evaluated out-of-sample (subsequent time periods). Our models outperform previous works (C-index 0.844-0.872 vs. 0.598-0.810).