Modeling COVID-19 care capacity in a major health system

  1. Margret Erlendsdottir
  2. Soheil Eshghi
  3. Forrest Crawford

Reviewed by

Read the full article

Listed in

Log in to save this article

Abstract

Hospital resources, especially critical care beds and ventilators, have been strained by additional demand throughout the COVID-19 pandemic. Rationing of scarce critical care resources may occur when available resource limits are exceeded. However, the dynamic nature of the COVID-19 pandemic and variability in projections of the future burden of COVID-19 infection pose challenges for optimizing resource allocation to critical care units in hospitals. Connecticut experienced a spike in the number of COVID-19 cases between March and June 2020. Uncertainty about future incidence made it difficult to predict the magnitude and duration of the increased COVID-19 burden on the healthcare system. In this paper, we describe a model of COVID-19 hospital capacity and occupancy that generates estimates of the resources necessary to accommodate COVID-19 patients under infection scenarios of varying severity. We present the model structure and dynamics, procedure for parameter estimation, and publicly available web application where we implemented the tool. We then describe calibration using data from over 3,000 COVID-19 patients seen at the Yale-New Haven Health System between March and July 2020. We conclude with recommendations for modeling tools to inform decision-making using incomplete information during future crises.

Article activity feed

  1. ScreenIT

    SciScore for 10.1101/2021.11.18.21266407: (What is this?)

    Please note, not all rigor criteria are appropriate for all manuscripts.

    Table 1: Rigor

    NIH rigor criteria are not applicable to paper type.

    Table 2: Resources

    Software and Algorithms
    SentencesResources
    This study received approval from the Institutional Review Board of Yale University’s Human Research Protection Program (IRB ID: 2000028666).
    Human Research Protection Program
    suggested: None

    Results from OddPub: We did not detect open data. We also did not detect open code. Researchers are encouraged to share open data when possible (see Nature blog).

    Results from LimitationRecognizer: We detected the following sentences addressing limitations in the study:
    However, this method may have several weaknesses. The model slightly underpredicted floor occupancy. This could be due to inaccurate assumptions used to estimate the model parameters. The model parameters were estimated for 3 age groups. The model fit could be improved in future versions by creating additional subgroups according to age and other risk factors for hospitalization and severe disease in COVID-19 patients. The model also assumes that rates of transition between departments remains constant over time. However, several factors, including changing hospital protocols for triage and treatment, may have resulted in fluctuations in the rates of transition over time. Such non-stationary behavior is challenging to capture and would not be captured by the model or parameter estimation procedure. In addition, the model is deterministic, and the estimates of variance in occupancy are based on uncertainty in parameter estimation rather than inherent stochasticity in the model. Improved estimates of variance might be achieved by making the model fully stochastic. The urgency of the crisis caused by surging COVID-19 patients contributed substantially to the challenge of developing a useful model. We would like to conclude by providing a few recommendations for creating a model in a crisis. First, early collaboration with end-users of the product was essential. After creation of the model structure and early implementation of the web application, we met several times with admini...

    Results from TrialIdentifier: No clinical trial numbers were referenced.

    Results from Barzooka: We did not find any issues relating to the usage of bar graphs.

    Results from JetFighter: We did not find any issues relating to colormaps.

    Results from rtransparent:
    • Thank you for including a conflict of interest statement. Authors are encouraged to include this statement when submitting to a journal.
    • Thank you for including a funding statement. Authors are encouraged to include this statement when submitting to a journal.
    • No protocol registration statement was detected.

    Results from scite Reference Check: We found no unreliable references.

    About SciScore

    SciScore is an automated tool that is designed to assist expert reviewers by finding and presenting formulaic information scattered throughout a paper in a standard, easy to digest format. SciScore checks for the presence and correctness of RRIDs (research resource identifiers), and for rigor criteria such as sex and investigator blinding. For details on the theoretical underpinning of rigor criteria and the tools shown here, including references cited, please follow this link.

  2. Version published to 10.1101/2021.11.18.21266407v1 on medRxiv