ABSTRACT
‘Federated Learning’ (FL) is a method to train Artificial Intelligence (AI) models with data from multiple sources while maintaining anonymity of the data thus removing many barriers to data sharing. During the SARS-COV-2 pandemic, 20 institutes collaborated on a healthcare FL study to predict future oxygen requirements of infected patients using inputs of vital signs, laboratory data, and chest x-rays, constituting the “EXAM” (EMR CXR AI Model) model. EXAM achieved an average Area Under the Curve (AUC) of over 0.92, an average improvement of 16%, and a 38% increase in generalisability over local models. The FL paradigm was successfully applied to facilitate a rapid data science collaboration without data exchange, resulting in a model that generalised across heterogeneous, unharmonized datasets. This provided the broader healthcare community with a validated model to respond to COVID-19 challenges, as well as set the stage for broader use of FL in healthcare.
Subject(s)
COVID-19 , InfectionsABSTRACT
OBJECTIVES: The symptoms of Coronavirus disease 2019 (COVID-19) vary among patients. The aim of this study was to investigate the clinical manifestation and disease duration in young versus elderly patients. METHODS: We retrospectively analyzed 187 patients (87 elderly and 100 young patients) with confirmed COVID-19. The clinical characteristics and chest computed tomography (CT) extent as defined by a score were compared between the two groups. RESULTS: The numbers of asymptomatic cases and severe cases were significantly higher in the elderly group (elderly group vs. young group; asymptomatic cases, 31 [35.6%] vs. 10 [10%], p < 0.0001; severe cases, 25 [28.7%] vs. 8 [8.0%], p = 0.0002). The proportion of asymptomatic patients and severe patients increased across the 10-year age groups. There was no significant difference in the total CT score and number of abnormal cases. A significant positive correlation between the disease duration and patient age was observed in asymptomatic patients (ρ = 0.4570, 95% CI 0.1198-0.6491, p = 0.0034). CONCLUSIONS: Although the extent of lung involvement did not have a significant difference between the young and elderly patients, elderly patients were more likely to have severe clinical manifestations. Elderly patients were also more likely to be asymptomatic and a source of COVID-19 viral shedding.
Subject(s)
Asymptomatic Infections/epidemiology , Coronavirus Infections/epidemiology , Pneumonia, Viral/epidemiology , Virus Shedding , Adult , Age Factors , Aged , Betacoronavirus , COVID-19 , Coronavirus Infections/diagnostic imaging , Coronavirus Infections/pathology , Female , Humans , Male , Middle Aged , Pandemics , Pneumonia, Viral/diagnostic imaging , Pneumonia, Viral/pathology , Retrospective Studies , SARS-CoV-2 , Severity of Illness Index , Tomography, X-Ray ComputedABSTRACT
Background: The ongoing outbreak of the coronavirus disease 2019 (COVID-19) is a global threat. Identification of markers for symptom onset and disease progression is a pressing issue. We compared the clinical features on admission among patients who were diagnosed with asymptomatic, mild, and severe COVID-19 at the end of observation. Methods: This retrospective, single-center study included 104 patients with laboratory-confirmed COVID-19 from the mass infection on the Diamond Princess cruise ship from February 11 to February 25, 2020. Clinical records, laboratory data, and radiological findings were analyzed. Clinical outcomes were followed up until February 26, 2020. Clinical features on admission were compared among those with different disease severity at the end of observation. Univariate analysis identified factors associated with symptom onset and disease progression. Findings: The median age was 68 years, and 54 patients were male. Briefly, 43, 41, and 20 patients on admission and 33, 43, and 28 patients at the end of observation had asymptomatic, mild, and severe COVID-19, respectively. Serum lactate hydrogenase levels were significantly higher in 10 patients who were asymptomatic on admission but developed symptomatic COVID-19 compared with 33 patients who remained asymptomatic throughout the observation period. Older age, consolidation on chest computed tomography, and lymphopenia on admission were more frequent in patients with severe COVID-19 than those with mild COVID-19 at the end of observation. Interpretation: Lactate dehydrogenase level is a potential predictor of symptom onset in COVID-19. Older age, consolidation on chest CT images, and lymphopenia might be risk factors for disease progression of COVID-19 and contribute to the clinical management. Funding: Not applicable.