Utility of the FebriDx point-of-care assay in supporting a triage algorithm for medical admissions with possible COVID-19: an observational cohort study

  1. Hamish Houston
  2. Gavin Deas
  3. Shivam Naik
  4. Kamal Shah
  5. Shiras Patel
  6. Maria Greca Dottori
  7. Michael Tay
  8. Sarah Ann Filson
  9. James Biggin-Lamming
  10. John Ross
  11. Natalie Vaughan
  12. Nidhi Vaid
  13. Guduru Gopal Rao
  14. Amit K Amin
  15. Ankur Gupta-Wright
  16. Laurence John

Reviewed by

Read the full article

Listed in

Log in to save this article

Abstract

To evaluate a triage algorithm used to identify and isolate patients with suspected COVID-19 among medical patients needing admission to hospital using simple clinical criteria and the FebriDx assay.

Design

Retrospective observational cohort.

Setting

Large acute National Health Service hospital in London, UK.

Participants

All medical admissions from the emergency department between 10 August 2020 and 4 November 2020 with a valid SARS-CoV-2 RT-PCR result.

Interventions

Medical admissions were triaged as likely, possible or unlikely COVID-19 based on clinical criteria. Patients triaged as possible COVID-19 underwent FebriDx lateral flow assay on capillary blood, and those positive for myxovirus resistance protein A (a host response protein) were managed as likely COVID-19.

Primary outcome measures

Diagnostic accuracy (sensitivity, specificity and predictive values) of the algorithm and the FebriDx assay using SARS-CoV-2 RT-PCR from nasopharyngeal swabs as the reference standard.

Results

4.0% (136) of 3443 medical admissions had RT-PCR confirmed COVID-19. Prevalence of COVID-19 was 46% (80/175) in those triaged as likely, 4.1% (50/1225) in possible and 0.3% (6/2033) in unlikely COVID-19. Using a SARS-CoV-2 RT-PCR reference standard, clinical triage had sensitivity of 96% (95% CI 91% to 98%) and specificity of 61.5% (95% CI 59.8% to 63.1%), while the triage algorithm including FebriDx had sensitivity of 93% (95% CI 87% to 96%) and specificity of 86.4% (95% CI 85.2% to 87.5%). While 2033 patients were deemed not to require isolation using clinical criteria alone, the addition of FebriDx to clinical triage allowed a further 826 patients to be released from isolation, reducing the need for isolation rooms by 9.5 per day, 95% CI 8.9 to 10.2. Ten patients missed by the algorithm had mild or asymptomatic COVID-19.

Conclusions

A triage algorithm including the FebriDx assay had good sensitivity and was useful to ‘rule-out’ COVID-19 among medical admissions to hospital.

Article activity feed

  1. Version published to 10.1136/bmjopen-2021-049179
  2. Version published to 10.1101/2021.01.05.21249154v2 on medRxiv
  3. ScreenIT

    SciScore for 10.1101/2021.01.05.21249154: (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
    Results are reported in compliance with STARD and STROBE guidelines (see supplementary materials).
    STARD
    suggested: None
    Statistical analyses were performed using Stata version 14.0 (StataCorp, LLC, College Station TX)
    StataCorp
    suggested: (Stata, RRID:SCR_012763)

    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:
    There are, however, several limitations. A single SARS-CoV-2 RT-PCR is an imperfect reference standard, and does not account for RT-PCR negative COVID-19 patients. We used multiple RT-PCR platforms, which will have different PCR targets and performance. 10% of patients in the possible group did not get tested with FebriDx for unclear reasons, potentially introducing bias. The prevalence of COVID-19 was 4% in this cohort, and it is unclear what impact a higher prevalence of COVID-19 or other respiratory pathogens such as influenza would have on these findings. The criteria for likely and possible COVID-19 groups changed subtly during the study period, although this is unlikely to significantly alter the outcomes. These data build on previous studies of FebriDx showing good sensitivity, and utility as a ‘rule-out’ test for COVID-19.17–20 We may have underestimated the sensitivity by not testing those patients deemed most likely to have COVID-19, although testing this group would have been unlikely to alter clinical decisions, even if FebriDx negative, given the high pre-test probability. The FebriDx test allowed patients with possible COVID-19 to be divided into two groups with similar characteristics and clinical features, but vastly different COVID-19 prevalence (0.5% in FebriDx negative, and 31.1% in FebriDx positive). However, about 10% of patients in this group were not eligible for FebriDx testing. Only ten patients with COVID-19 were incorrectly triaged by the algorithm,...

    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.
    • Thank you for including a protocol registration statement.

    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.

  4. Version published to 10.1101/2021.01.05.21249154v1 on medRxiv