An statistical analysis of COVID-19 intensive care unit bed occupancy data

TL;DR

This paper introduces a robust statistical framework using mixed models to estimate and predict COVID-19 ICU bed occupancy, addressing data scarcity and heterogeneity challenges during the pandemic.

Contribution

The study develops an innovative Small Area Estimation methodology with mixed models for accurate ICU bed occupancy prediction.

Findings

Effective estimation of ICU bed occupancy in COVID-19 context

Application to Castilla y León data from Nov 2020 to Mar 2022

Addresses data scarcity and heterogeneity issues

Abstract

The COVID-19 pandemic has had far-reaching consequences, highlighting the urgency for explanatory and predictive tools to track infection rates and burden of care over time and space. However, the scarcity and inhomogeneity of data is a challenge. In this research we develop a robust framework for estimating and predicting the occupied beds of Intensive Care Units by presenting an innovative Small Area Estimation methodology based on the definition of mixed models with random regression coefficients. We applied it to estimate and predict the daily occupancy of Intensive Care Unit beds by COVID-19 in health areas of Castilla y Le\'on, from November 2020 to March 2022.