Forecasting hospital demand during COVID-19 pandemic outbreaks

TL;DR

This paper introduces a Bayesian SEIRD model to forecast hospital demand during COVID-19, incorporating detailed infection dynamics and intervention effects, aiding public health policy in Mexico.

Contribution

The paper presents a novel compartmental model with explicit hospital dynamics and Bayesian inference for timely COVID-19 hospital demand forecasting.

Findings

Model successfully forecasted hospital demand in over 70 metropolitan areas.

Informed public policy decisions during COVID-19 outbreaks.

Demonstrated the model's applicability across diverse regions.

Abstract

We present a compartmental SEIRD model aimed at forecasting hospital occupancy in metropolitan areas during the current COVID-19 outbreak. The model features asymptomatic and symptomatic infections with detailed hospital dynamics. We model explicitly branching probabilities and non exponential residence times in each latent and infected compartments. Using both hospital admittance confirmed cases and deaths we infer the contact rate and the initial conditions of the dynamical system, considering break points to model lockdown interventions. Our Bayesian approach allows us to produce timely probabilistic forecasts of hospital demand. The model has been used by the federal government of Mexico to assist public policy, and has been applied for the analysis of more than 70 metropolitan areas and the 32 states in the country.