Predictive Modeling using Point of Care Lung Ultrasound (P-LUS) for Emergency Triage of Patients with Acute Respiratory Symptoms Related to COVID-19
Abstract
The unprecedented number of patients with COVID-19 disease (infected or presumably infected by SARS-CoV-2) seeking urgent care for acute respiratory symptoms severely disrupted the allocation of hospital personnel and resources. Progressive respiratory failure may occur rapidly (<72h)during COVID-19. It is imperative to identify safe and effective diagnostic tools to appropriately allocate scarce resources early while minimizing further viral spread. Point-of-care lung ultrasound (P-LUS) imaging is available with the portable, cost-effective Butterfly iQ probes. We tested the hypothesis that P-LUS imaging and other predictors provide effective emergency triage and early identification of hospital resources required by COVID-19 patients presenting with acute respiratory symptoms. We developed a simplified scoring system (adjusted from prior published score rubric) and tested the score inter-user agreement. We then used this scoring matric to evaluate P-LUS images obtained for clinical purposes from patients presenting with possible COVID-19-related acute respiratory symptoms to the Emergency Department (ED) at three participating sites between March 2020 and April 2021. At least three blinded ultrasound-trained investigators evaluated the images available from eligible patients and provided the worst score (from normal, 0, to most abnormal, 3) for all the P-LUS exams available for each patient. A predictive model was then developed for the need of hospital admission (vs. discharge) from the ED, including clinically relevant vital signs and laboratory values, and P-LUS findings. Reviewers had uniform agreement on the low (scores 0-1) or high (2-3) risk classification of 66.7% cases. The presence of diffuse B-lines showed the highest level of agreement among reviewers (ICC 0.906).
Document Details
- Document Type
- Technical Report
- Publication Date
- Jun 26, 2022
- Accession Number
- AD1203387
Entities
People
- Ana Fernandez-bustamante
- Joe Maddry
- John Kendall
- Patrick C. Ng
Organizations
- 59th Medical Wing
- Denver Health Medical Center
- University of Colorado School of Medicine