Your browser doesn't support javascript.
Show: 20 | 50 | 100
Results 1 - 4 de 4
Annals of Emergency Medicine ; 76(4):S109-S110, 2020.
Article in English | Web of Science | ID: covidwho-921379
Annals of Emergency Medicine ; 76(4):S109-S110, 2020.
Article in English | EMBASE | ID: covidwho-898439


Study Objectives: The surge and long tail of patients in acute respiratory distress during the coronavirus-19 (CoVID19) pandemic has inspired new innovations in diagnosing, treating and dispositioning patients during high census conditions with constrained resources. During the first wave of the pandemic, we deployed an artificial intelligence (AI) algorithm for assisted interpretation of chest x-ray for use by radiologists and emergency department (ED) physicians. We report first experiences of physician interaction with this novel AI algorithm designed to enhance physician abilities to identify ground glass and consolidation on chest radiographs. Methods: Design: We created a fully-automated pipeline into the clinical environment to provide AI augmentation of chest x-rays, utilizing a previously developed deep learning-based AI algorithm. Trained with 22,000 annotations by radiologists, the algorithm overlays X-rays with color-coded maps that indicate pneumonia probability. This was provided alongside standard chest x-ray images for physicians to use in real-time at the point of care with existing imaging software. For this prospective observational study, we developed a 3-point survey to characterize experiences with the tool regarding ease of use and impact on clinical decision-making. Setting: Surveys were conducted during a one-month period surrounding the projected CoVID-19 surge locally (April 8-May 9) at two academic hospitals in Southern California. A federal declaration of emergency occurred March 13, 2020 and the tool was urgently deployed on March 25. Types of Participants: Emergency medicine resident and attending physicians surveyed in real time by telephone. Results: Of the 5,125 total visits and 1,960 chest radiographs obtained in the ED during the study period, 1,855 were analyzed by the algorithm. Among these, emergency physicians were surveyed for their experiences on 202. Real-time computation and delivery of the tool took four minutes on average. Overall, 86% either strongly agreed or somewhat agreed that the intervention was easy to use in the existing workflow. 20% of all respondents reported that the algorithm impacted their clinical decision making. In general, resident physicians found the AI implementation easier to use than attendings (Mann Whitney U, p=0.005). Descriptive statistics regarding further impact are summarized below (table 1). Conclusion: This AI technology was rapidly deployed in a large academic health system in the first wave of a global pandemic. Surveyed ED physicians found this implementation easy to use within existing workflows. Twenty percent of physicians reported that the tool changed clinical decision making, and approximately one third of those found that it impacted diagnostic testing decisions and treatment plans. Several physicians reported ordering COVID-19 PCR testing as a direct result of the AI, resulting in positive tests and subsequent quarantining of patients who otherwise might not have been appropriately diagnosed. To our knowledge, this is the first published study evaluating the impact of medical imaging AI on clinical decision making in the ED setting and may prove to be a powerful tool during the pandemic response. [Formula presented]