Identification on Admission of COVID-19 Patients at Risk of Subsequent Rapid Clinical Deterioration

  1. J. Beals
  2. J. Barnes
  3. D. Durand
  4. J. Rimar
  5. T. Donohue
  6. M. Hoq
  7. K. Belk
  8. A. Amin
  9. M.J. Rothman

Reviewed by

Read the full article

Listed in

Log in to save this article

Abstract

Introduction

Recent localized surges in COVID-19 cases have resulted in the hospitals serving those areas being overwhelmed. In such cases, the ability to rapidly and objectively determine a patient’s acuity and predict near-term care needs is a major challenge. At issue is the clinician’s ability to correctly identify patients at risk for subsequent rapid clinical deterioration. Data-driven tools that can support such determinations in real-time may be a valuable adjunct to clinician judgement during COVID-19 surges.

Objective

To assess the effectiveness of the Rothman Index (RI) predictive model in distinguishing the risk of subsequent deterioration or elevated care needs among hospitalized COVID-19 patients at the time of hospital admission.

Methods

We evaluated the initial RI score on admission to predict COVID-19 patient risk for 216 COVID-19 patients discharged from March 21 st to June 7 th , 2020 at Sinai LifeBridge Hospital and 1,453 COVID-19 patients discharged from any of Yale New Haven Health System’s Yale New Haven, Bridgeport, and Greenwich hospitals from April 1 st to April 28 th , 2020. In-hospital mortality as a function of age and RI on admission for COVID-19 and non-COVID-19 patients were compared. AUC values using each COVID-19 patient’s initial RI on admission to predict in-hospital mortality, mechanical ventilation, and ICU utilization were computed, as were precision and recall for mortality prediction at specific RI thresholds.

Results

The RI computed at the time of admission provides a high degree of objective discrimination to differentiate the COVID-19 population into high and low risk populations at the outset of hospitalization. The high risk segment based on initial RI constitutes 20-30% of the COVID-19 positive population with mortality rates from 40-50%. The low risk segment based on initial RI constitutes 40%-55% of the population with mortality rates ranging from 1%-8%. Of note is that COVID-19 patients who present with heightened but generally unremarkable acuity can be identified early as having considerably elevated risk for subsequent physiological deterioration.

Conclusion

COVID-19 patients exhibit elevated mortality rates compared to non-COVID-19 medical service patients and may be subject to rapid deterioration following hospital admission. A lack of predictive indicators for identifying patients at high risk of subsequent deterioration or death can pose a challenge to clinicians. The RI has excellent performance characteristics when stratifying risk among COVID-19 patients at the time of admission. The RI can assist clinicians in real-time with a high degree of objective discrimination by segmenting the COVID-19 population into high and low risk populations. This supports rapid and optimal patient bed assignment and resource allocation.

Article activity feed

  1. ScreenIT

    SciScore for 10.1101/2020.08.13.20171751: (What is this?)

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

    Table 1: Rigor

    Institutional Review Board StatementIRB: This research was approved by both the Sinai LifeBridge and Yale New Haven Health Bridgeport Institutional Review Boards.
    Randomizationnot detected.
    Blindingnot detected.
    Power Analysisnot detected.
    Sex as a biological variablenot detected.

    Table 2: Resources

    No key resources detected.

    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:
    Limitations of the study: This work was a retrospective analysis of COVID-19 patients, and as such presupposes knowledge of whether or not a patient was accurately diagnosed with COVID-19. This work does not account for possible coding error nor uncertainty arising from the possibility of false positive or false negative diagnostic lab tests. However, with the dramatic improvement in testing capacity it is reasonable to expect that diagnoses are correct in most instances. Additionally, the care of COVID-19 patients is evolving rapidly, and treatments such as dexamethasone, remdesivir and convalescent plasma, proning, and anti-coagulation therapies are now being used in a manner that was not routine during the period of our study data32-38 while the benefits of early, aggressive intubation are increasingly debated.39,40 Without chart review, it is not possible to know what patient goals of care may have been and the extent to which they influenced decisions related to treatment or to receiving care in the ICU. We also note that the data reflecting non-COVID-19 cases largely overlaps with a time period when most U.S. hospitals were taking extraordinary measures to manage population volumes and hence the medical admissions during this time will not be perfectly representative of typical non-COVID-19 medical patient populations. Conclusion: COVID-19 patients have high mortality rates compared to other medical service patients. They are subject to rapid deterioration which may fol...

    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.

    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 10.1101/2020.08.13.20171751v1 on medRxiv