1042 Improving the accuracy of frontline clinicians in detecting SARS-COV-2 on chest X-rays using a bespoke virtual training platform. Issue 3 (21st February 2022)

• Record Type: Journal Article
• Authors: Bahra, Jasdeep; Ather, Sarim; Wilson, Sarah; Keating, Liza; Gulati, Divyansh; Banerji, Abhishek; Gleeson, Fergus; Novak, Alex
• Is Part Of: Emergency medicine journal. Volume 39:Issue 3(2022)
• Page Start: 252
• Page End: 252
• Publication Date: 2022-02-21
• Subjects: Emergency medicine -- Periodicals; 616.02505
• DOI: 10.1136/emermed-2022-RCEM.18
• Languages: English
• ISSNs: 1472-0205

Abstract

Abstract : Aims/Objectives/Background: The non-specific symptoms of COVID-19 and the lack of a highly-sensitive point-of-care test make it difficult to reliably detect and diagnose in acute care settings. The early identification of COVID-19 using chest X-rays (CXR) in the Emergency Department (ED) is a crucial skill for frontline clinicians. We wanted to measure the accuracy of ED clinicians in detecting COVID-19 CXR changes and assess for improvement using an adaptive online learning module. Methods/Design: ED clinicians working across five hospitals in the Thames Valley Emergency medicine Research Network (TaVERN) were recruited over six months. Participants' reporting performance was assessed by interpreting 30 anonymised CXR via the Report and Image Quality Control (RAIQC) online platform, using an image bank which contained both COVID-19 and non-COVID-19 pathological findings. Participants subsequently completed an online training module, and repeated the assessment using different image sets. Diagnostic accuracy and speed of CXR reporting was assessed both before and after training, with results compared against radiologists. The ground truth for each case was established by consensus of three thoracic radiologists. RT-PCR results were reviewed for each case to ensure that all the COVID-19 cases were positive and all COVID-19 cases were negative. Results/Conclusions: ED clinicians working in emergency departments across five hospitals in the Thames Valley Emergency Medicine Research Network (TaVERN) were recruited over a six month period. 112 clinicians completed the initial assessment. 56 clinicians completed all three training components. The initial mean accuracy for clinicians in identifying COVID-19 on chest X-rays was 43%. The mean accuracy was 57% amongst clinicians who completed all three online training components. These clinician showed improved reporting speed with mean time reduction to CXR interpretation from 69 to 50 seconds. ED clinicians do not perform well at detecting COVID-19 CXR related changes on CXR, but accuracy and speed can be improved by online training.