From predictions to prescriptions: A data-driven response to COVID-19

Abstract

The COVID-19 pandemic has created unprecedented challenges worldwide. Strained healthcare providers make difficult decisions on patient triage, treatment and care management on a daily basis. Policy makers have imposed social distancing measures to slow the disease, at a steep economic price. We design analytical tools to support these decisions and combat the pandemic. Specifically, we propose a comprehensive data-driven approach to understand the clinical characteristics of COVID-19, predict its mortality, forecast its evolution, and ultimately alleviate its impact. By leveraging cohort-level clinical data, patient-level hospital data, and census-level epidemiological data, we develop an integrated four-step approach, combining descriptive, predictive and prescriptive analytics. First, we aggregate hundreds of clinical studies into the most comprehensive database on COVID-19 to paint a new macroscopic picture of the disease. Second, we build personalized calculators to predict the risk of infection and mortality as a function of demographics, symptoms, comorbidities, and lab values. Third, we develop a novel epidemiological model to project the pandemic’s spread and inform social distancing policies. Fourth, we propose an optimization model to re-allocate ventilators and alleviate shortages. Our results have been used at the clinical level by several hospitals to triage patients, guide care management, plan ICU capacity, and re-distribute ventilators. At the policy level, they are currently supporting safe back-to-work policies at a major institution and vaccine trial location planning at Janssen Pharmaceuticals, and have been integrated into the US Center for Disease Control’s pandemic forecast. Electronic supplementary material The online version of this article (10.1007/s10729-020-09542-0) contains supplementary material, which is available to authorized users. Supplementary Information The online version contains supplementary material available at (10.1007/s10729-020-09542-0)

Where to locate COVID‐19 mass vaccination facilities?

“…Ultimately, DELPHI involves 16 parameters that define the transition rates between the 11 states. We calibrate seven of them from a database on clinical outcomes (Bertsimas et al, 2020 ). Using nonlinear optimization, we estimate the other nine parameters from historical data on the number of cases and deaths in each region.…”

Task complexity moderates group synergy

“…The connection to operations research is useful because unlike other “toy” problems, CSOPs map in a relatively intuitive way to a range of practical resource allocation problems and have been used to model many problems that are of practical interest. Examples of CSOPs include staffing software projects where there are several potential developer-to-activity assignments to evaluate ( 19 ); forming learning groups based on some criteria related to the collaboration goals ( 20 ); railway timetabling ( 21 ); and allocating vaccines, ventilators, and medical supplies during the COVID-19 pandemic ( 22 ). Furthermore, while CSOPs capture important features of real-world group problem-solving exercises, they do not require participants to have specialized skills.…”

COVID-19: Data-Driven optimal allocation of ventilator supply under uncertainty and risk

“…(2020) Optimize the allocation of ventilators COVID-19 Tanner et al. (2008) Optimize the vaccination policy General Multi-Stage Stochastic Programming Yin & Büyüktahtakın (2021a) Optimize resources and centers allocation under uncertainty and equity Ebola Yin & Büyüktahtakın (2021b) Optimize vaccine allocation and treatment logistics under a mean-CVaR objective Ebola This paper Optimize ventilator allocation under asymptomatic uncertainty and risk COVID-19 MIP and/or Machine Learning Bertsimas et al. (2020) Optimize ventilator allocation in the US COVID-19 Bushaj et al.…”

“…Operations Research (OR) methods have been widely used to determine optimal resource allocation strategies to control an epidemic or pandemic. Several studies have used multi-period OR models to optimize the allocation and redistribution of ventilators (see, e.g., Mehrotra, Rahimian, Barah, Luo, & Schantz (2020) , Bertsimas et al. (2020) , and Blanco, Gázquez, & Leal (2020) ).…”