Open Access
May 2022 Learning and Predicting from Dynamic Models for COVID-19 Patient Monitoring
Zitong Wang, Mary Grace Bowring, Antony Rosen, Brian Garibaldi, Scott Zeger, Akihiko Nishimura
Author Affiliations +
Statist. Sci. 37(2): 251-265 (May 2022). DOI: 10.1214/22-STS861


COVID-19 has challenged health systems to learn how to learn. This paper describes the context, methods and challenges for learning to improve COVID-19 care at one academic health center. Challenges to learning include: (1) choosing a right clinical target; (2) designing methods for accurate predictions by borrowing strength from prior patients’ experiences; (3) communicating the methodology to clinicians so they understand and trust it; (4) communicating the predictions to the patient at the moment of clinical decision; and (5) continuously evaluating and revising the methods so they adapt to changing patients and clinical demands.

To illustrate these challenges, this paper contrasts two statistical modeling approaches—prospective longitudinal models in common use and retrospective analogues complementary in the COVID-19 context—for predicting future biomarker trajectories and major clinical events. The methods are applied to and validated on a cohort of 1678 patients who were hospitalized with COVID-19 during the early months of the pandemic. We emphasize graphical tools to promote physician learning and inform clinical decision making.

Funding Statement

Dr. Zeger was partially supported by NIH Grants P30AR070254, 5UL1TR003098 and 1U54CA260492 and by the Scleroderma Research Foundation Grant IPN 21054327. Dr. Garibaldi is a member of the FDA Pulmonary-Asthma Drug Advisory Committee; is a consultant for Janssen Research and Development, LLC; has received speaker fees and served on an advisory panel for Gilead; has received speaker fees from Atea. Ms. Bowring is supported by NIH grant T32GM136577. Ms. Wang is partially supported by the Patrick C. Walsh Prostate Cancer Research Fund.
The COVID-19 PMAP Registry was funded by Johns Hopkins Medicine through the Precision Medicine Program. The studies were also supported through the generosity of the collective community of donors to the Johns Hopkins University School of Medicine for COVID-19 research.
This work was supported by funding from John Hopkins inHealth, the Johns Hopkins Precision Medicine initiative through JH-CROWN, and the coronavirus disease 2019 (COVID-19) Administrative Supplement for the US Department of Health and Human Services (HHS) Region 3 Treatment Center from the Office of the Assistant Secretary for Preparedness and Response.


The authors would like to thank Dr. Brian Caffo, the referees, Associate Editor and Editor for their constructive comments that improved the quality of this paper. We also thank the many persons who built the CROWN registry and the CADRE team that has administered its use.


Download Citation

Zitong Wang. Mary Grace Bowring. Antony Rosen. Brian Garibaldi. Scott Zeger. Akihiko Nishimura. "Learning and Predicting from Dynamic Models for COVID-19 Patient Monitoring." Statist. Sci. 37 (2) 251 - 265, May 2022.


Published: May 2022
First available in Project Euclid: 16 May 2022

MathSciNet: MR4422307
zbMATH: 07535202
Digital Object Identifier: 10.1214/22-STS861

Keywords: decision support , inverse regression , longitudinal data analysis , prediction , statistical graphics

Rights: Copyright © 2022 Institute of Mathematical Statistics

Vol.37 • No. 2 • May 2022
Back to Top