Validation of home oxygen saturations as a marker of clinical deterioration in patients with suspected COVID-19

  1. Matthew Inada-Kim
  2. Francis P. Chmiel
  3. Michael J. Boniface
  4. Helen Pocock
  5. John J. M. Black
  6. Charles D. Deakin

Reviewed by

Read the full article

Listed in

Log in to save this article

Abstract

Background

The early identification of deterioration in suspected COVID-19 patients managed at home enables a more timely clinical intervention, which is likely to translate into improved outcomes. We undertook an analysis of COVID-19 patients conveyed by ambulance to hospital to investigate how oxygen saturation and measurements of other vital signs correlate to patient outcomes, to ascertain if clinical deterioration can be predicted with simple community physiological monitoring.

Methods

A retrospective analysis of routinely collected clinical data relating to patients conveyed to hospital by ambulance was undertaken. We used descriptive statistics and predictive analytics to investigate how vital signs, measured at home by ambulance staff from the South Central Ambulance Service, correlate to patient outcomes. Information on patient comorbidities was obtained by linking the recorded vital sign measurements to the patient’s electronic health record at the Hampshire Hospitals NHS Foundation Trust. ROC analysis was performed using cross-validation to evaluate, in a retrospective fashion, the efficacy of different variables in predicting patient outcomes.

Results

We identified 1,080 adults with a COVID-19 diagnosis who were conveyed by ambulance to either Basingstoke & North Hampshire Hospital or the Royal Hampshire County Hospital (Winchester) between March 1 st and July 31 st and whose diagnosis was clinically confirmed at hospital discharge. Vital signs measured by ambulance staff at first point of contact in the community correlated with patient short-term mortality or ICU admission. Oxygen saturations were the most predictive of mortality or ICU admission (AUROC 0.772 (95 % CI: 0.712-0.833)), followed by the NEWS2 score (AUROC 0.715 (95 % CI: 0.670-0.760), patient age (AUROC 0.690 (95 % CI: 0.642-0.737)), and respiration rate (AUROC 0.662 (95 % CI: 0.599-0.729)). Combining age with the NEWS2 score (AUROC 0.771 (95 % CI: 0.718-0.824)) or the measured oxygen saturation (AUROC 0.820 (95 % CI: 0.785-0.854)) increased the predictive ability but did not reach significance.

Conclusions

Initial oxygen saturation measurements (on air) for confirmed COVID-19 patients conveyed by ambulance correlated with short-term (30-day) patient mortality or ICU admission, AUROC: 0.772 (95% CI: 0.712-0.833). We found that even small deflections in oxygen saturations of 1-2% below 96% confer an increased mortality risk in those with confirmed COVID at their initial community assessments.

Article activity feed

  1. ScreenIT

    SciScore for 10.1101/2020.11.06.20225938: (What is this?)

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

    Table 1: Rigor

    Institutional Review Board Statementnot detected.
    Randomizationnot detected.
    Blindingnot detected.
    Power Analysisnot detected.
    Sex as a biological variablenot detected.

    Table 2: Resources

    Software and Algorithms
    SentencesResources
    Data analysis: Analysis was performed in Python, primarily making use of the statsmodels library.
    Python
    suggested: (IPython, RRID:SCR_001658)

    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, even oxygen saturations at the lower end of the normal range are associated with a risk of deterioration (sensitivity of 94% saturations = 0.713) and it therefore appears that oxygen saturation alone has significant limitations when it is within a normal range. Our data (Table 3) suggests that the addition of age to oxygen saturation measurements may further increase sensitivity to identify risk of mortality at 30 days, but a larger study is required to ascertain whether this is the case. We also examined co-morbidities known to be risk factors for adverse outcomes with COVID-19 infections.[12] Age was strongly associated with mortality at 30-days, which rose rapidly in those over 60 years of age. COPD patients had almost double the observed 30-day mortality rate compared to those without COPD, but this did not reach statistical significance, likely because of sample size of our study. Contrary to most other studies, diabetes was not a risk factor for adverse outcome in our study, the reason for this is not clear. In total, 21.7% of our sample was diabetic which is similar to that reported in a large UK study of hospitalised COVID-19 patients.[12] Being under cancer care was also not a significant factor for adverse outcome. Although oxygen saturations as a risk factor for COVID-19 patients on presentation to the Emergency Department are widely reported,[13, 14] [15] the ability of oxygen saturations measured at home to indicate disease severity and the need for hosp...

    Results from TrialIdentifier: No clinical trial numbers were referenced.

    Results from Barzooka: We found bar graphs of continuous data. We recommend replacing bar graphs with more informative graphics, as many different datasets can lead to the same bar graph. The actual data may suggest different conclusions from the summary statistics. For more information, please see Weissgerber et al (2015).

    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. ScreenIT

    SciScore for 10.1101/2020.11.06.20225938: (What is this?)

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

    Table 1: Rigor

    Institutional Review Board StatementGovernance and ethics approval Regulatory and ethical approval for the study were provided by the Health Research Authority (REC reference 20/HRA/5445) and by the University of Southampton Ethics Committee (REF ERGO/61242).Randomizationnot detected.Blindingnot detected.Power Analysisnot detected.Sex as a biological variablenot detected.

    Table 2: Resources

    Software and Algorithms
    SentencesResources
    Data analysis Analysis was performed in Python, primarily making use of the statsmodels library.
    Python
    suggested: (IPython, RRID:SCR_001658)

    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, even oxygen saturations at the lower end of the normal range are associated with a risk of deterioration (sensitivity of 94% saturations = 0.713) and it therefore appears that oxygen saturation alone has significant limitations when it is within a normal range. Our data (Table 3) suggests that the addition of age to oxygen saturation measurements may further increase sensitivity to identify risk of mortality at 30 days, but a larger study is required to ascertain whether this is the case. We also examined co-morbidities known to be risk factors for adverse outcomes with COVID-19 infections.[12] Age was strongly associated with mortality at 30-days, which rose rapidly in those over 60 years of age. COPD patients had almost double the observed 30-day mortality rate compared to those without COPD, but this did not reach statistical significance, likely because of sample size of our study. Contrary to most other studies, diabetes was not a risk factor for adverse outcome in our study, the reason for this is not clear. In total, 21.7% of our sample was diabetic which is similar to that reported in a large UK study of hospitalised COVID-19 patients.[12] Being under cancer care was also not a significant factor for adverse outcome. Although oxygen saturations as a risk factor for COVID-19 patients on presentation to the Emergency Department are widely reported,[13, 14] [15] the ability of oxygen saturations measured at home to indicate disease severity and the need for hosp...

    Results from TrialIdentifier: No clinical trial numbers were referenced.

    Results from Barzooka: We found bar graphs of continuous data. We recommend replacing bar graphs with more informative graphics, as many different datasets can lead to the same bar graph. The actual data may suggest different conclusions from the summary statistics. For more information, please see Weissgerber et al (2015).

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

    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.

  3. Version published to 10.1101/2020.11.06.20225938v1 on medRxiv