A precise score for the regular monitoring of COVID-19 patients condition validated within the first two waves of the pandemic

  1. Evgeny A. Bakin
  2. Oksana V. Stanevich
  3. Vassily A. Belash
  4. Anastasia A. Belash
  5. Galina A. Savateeveva
  6. Veronika A. Bokinova
  7. Natalia A. Arsentieva
  8. Ludmila F. Sayenko
  9. Evgeny A. Korobenkov
  10. Dmitry A. Lioznov
  11. Areg A. Totolian
  12. Yury S. Polushin
  13. Alexander N. Kulikov

Reviewed by

Read the full article See related articles

Listed in

Log in to save this article

Abstract

Purpose

The sudden outbreak of COVID-19 pandemic has shown that the medical community needs an accurate and interpretable aggregated score not only for an outcome prediction but also for a daily patient’s condition assessment. Due to a continuously changing pandemic landscape, robustness becomes a crucial additional requirement for the score.

Patients and methods

In this research, real-world data collected within the first two waves of the COVID-19 pandemic was used. The first wave data (1349 cases collected from 27.04.2020 to 03.08.2020) was used as a training set for the score development, while the second wave data (1453 cases collected from 01.11.2020 to 19.01.2021) was used as a validating set. For all the available patients’ features we tested their association with an outcome using robust linear regression. Statistically significant features were taken to the further analysis for each of which their partial sensitivity, specificity and promptness were estimated. The sensitivity and the specificity were further combined into a feature informativeness index.

Results

The developed score was derived as a weighted sum of the following 9 features showed the best trade-off between informativeness and promptness: APTT (> 42 sec, 4 points), CRP (> 146 mg/L, 3 points), D-dimer (> 2149 mkg/L, 4 points), Glucose (> 9 mmol/L, 4 points), Hemoglobin (< 115 g/L, 3 points), Lymphocytes (< 0,7*10^9/L, 3 points), Total protein (< 61 g/L, 6 points), Urea (> 11 mmol/L, 5 points) and WBC (> 13,5*10^9/L, 4 points). Thus, the proposed score ranges between 0 and 36 points. Internal and temporal validation showed that sensitivity and specificity over 90% may be achieved with an expected prediction range >7 days. Moreover, we demonstrated a high robustness of the score to the varying peculiarities of the pandemic. For the additional simplicity of application we split the full range of the score into five grades delimited with 9, 14, 19 and 24 points which determine expected death:discharge odds 1:100, 1:25, 1:5 and 1:1 correspondingly.

Conclusions

An extensive application of the score within the second wave of the COVID-19 pandemic showed its potential for the optimization of patients management as well as improvement of medical staff attentiveness during high workload stress. The transparent structure of the score, as well as tractable cut-off bounds, simplified its implementation into clinical practice.

Article activity feed

  1. ScreenIT

    SciScore for 10.1101/2021.02.09.21249859: (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

    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:
    Study limitations: The study has the following limitations. Firstly, it was a single-center study with a temporal validation. Secondly, a few features that showed the efficiency in previous studies (such as ferritin, LDH, procalcitonin and troponin), were excluded from the analysis due to insufficiency of data. Also, laboratory tests for various features were done asynchronously and with different updating periods. This indicates different degrees of the score components relevance at the time of its calculation. For this reason, we assume that the score distribution will depend on the rate of analyses sampling in each particular hospital. Finally, the score doesn’t contain any components reflecting a patient respiratory function. Thus, an objective estimation of a patient condition may be performed only in combination with oxygenation parameters.

    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 10.1101/2021.02.09.21249859v3 on medRxiv
  3. Version published to 10.1101/2021.02.09.21249859v2 on medRxiv