ABSTRACT
Background: Point of care lung ultrasound (LUS) has become an integral part in the clinical care and evaluation of patients presenting with respiratory complaints in the setting of COVID-19 infection. Since the start of the COVID-19 pandemic, LUS has been used not only to help identify possible COVID-19 infection, but also to help prognosticate and risk stratify patients with known, or highly suspicious for, COVID-19 infection 24. Study Objective: To determine if point-of-care LUS can be used to risk stratify patients presenting under suspicion of COVID-19 infection. Methods: 118 patients were scanned using 8-point LUS score method looking at 4 lung fields on each side in order to evaluate the diagnostic and prognostic value of LUS in COVID-19 patients. Scores were assigned to each field based on presence of B-lines, pleural abnormalities, and subpleural consolidations. All lung ultrasounds were performed in the emergency department on persons under investigation (PUI) for COVID-19 respiratory infections. Result: There is a clear trend of increasing mean total LUS score with increasing severity of illness. The increasing severity was defined in ascending order as patients discharged, admitted to floor, admitted to ICU, and death in hospital. The mean total LUS score for each was: discharged (5.18 ±1.47 [95% CI 3.71-6.65]), admitted to floor (9.82 ± 1.57 [95% CI 8.25-11.4]), admitted to ICU (10.83 ± 1.99 [95% CI 8.84-12.8] ), and death in hospital (13.14 ± 4.64 [95% CI 8.5-17.8]). One of the deaths was a patient with a means total LUS score of 3 who was placed on comfort care and then terminally extubated in the setting of metastatic lung disease. If this patient is removed, the mean LUS score associated with death in hospital is 14.83 ± 3.83 [95% CI 11-18.7]. Overall, patient’s that tested positive for COVID-19 had a higher mean LUS score (8.71 ± 1.3 [95% CI 7.41-10) than those that tested negative (7.24 ± 1.86 [95% CI 5.38-9.1). A SpO2 greater than or equal to 90% was associated with a lower average LUS score (7.76 ± 1.24 [95% CI 6.52-9), than an SpO2 less than 90% (12.24 ± 2.24 [95% CI 10-14.5). Patient’s requiring high flow nasal cannula, non-invasive positive pressure ventilation, or intubation had a mean LUS score of 12.75 ± 2.05 [95% CI 10.7-14.8], while those who only required nasal cannula or no supplemental oxygen had mean LUS score of 8.76 ± 1.5 [95% CI 7.26-10.3]. Conclusion: Our results show that by using an 8 zone lung ultrasound protocol not only are we able to identify those patients more likely to test positive for COVID, but also to risk stratify those patients under suspicion of a COVID infection.
ABSTRACT
Since the COVID-19 pandemic surges around the world and officially entered a dangerous new phase, one of the important concerns is when to take aggressive public health measures to slow the spread of COVID-19 and to know impact of the use of protection tools. Many studies have dealt with the prediction of the evolution of cases affected by the C'OVID-19 virus. Given the unreliability of the data collected about the number of new cases and the uncertainties in values, the results found cannot be accurate and present a bias. In this paper, we will present a study using artificial intelligence algorithms more precisely machine and deep learning algorithms to predict the evolution of cases reached by COVID-19 in the future given the application of confinement and the use of protection tools. To improve the accuracy of the results and to take into account the uncertain aspect of the data we will apply the theory of belief functions. Among objectives of this theory is the fusion of different sources of information, given by artificial intelligence algorithms in our case, in order to obtain a global knowledge in the form of a more precise and reinforced belief function. Results shows that applying the home isolation and the use of protection tools with the rate over of 80% can reduce considerably the number of cases. © 2020 International Information and Engineering Technology Association. All rights reserved.