Symptom-Based Predictive Model of COVID-19 Disease in Children

• Published in: Viruses Vol. 14; no. 1; p. 63
• Main Authors: Antoñanzas, Jesús M, Perramon, Aida, López, Cayetana, Boneta, Mireia, Aguilera, Cristina, Capdevila, Ramon, Gatell, Anna, Serrano, Pepe, Poblet, Miriam, Canadell, Dolors, Vilà, Mònica, Catasús, Georgina, Valldepérez, Cinta, Català, Martí, Soler-Palacín, Pere, Prats, Clara, Soriano-Arandes, Antoni, The Copedi-Cat Research Group
• Format: Journal Article
• Language: English
• Published: Switzerland MDPI AG 30.12.2021; MDPI
• Subjects: Adolescent; Artificial intelligence; Bacterial infections; Child; Child, Preschool; Children; Children & youth; Coronaviruses; COVID-19; COVID-19 - diagnosis; COVID-19 - epidemiology; COVID-19 Testing - methods; Decision making; deep learning; Diagnosis; Diagnostic tests; Disease; Epidemiology; Female; Fever; Humans; Infant; Infant, Newborn; Infections; Learning algorithms; Machine Learning; Male; Medical research; Models, Statistical; Olfaction; paediatrics; Pandemics; Pediatrics; Population; Prediction models; Predictive Value of Tests; Primary care; Questionnaires; SARS-CoV-2; SARS-CoV-2 - isolation & purification; Severe acute respiratory syndrome coronavirus 2; Statistical analysis; Systematic review; Teenagers; Viral infections; Viruses; Spain; COVID-19; paediatrics; deep learning; epidemiology; SARS-CoV-2; microbiology; machine learning
• ISSN: 1999-4915; 1999-4915
• DOI: 10.3390/v14010063

Testing for severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) infection is neither always accessible nor easy to perform in children. We aimed to propose a machine learning model to assess the need for a SARS-CoV-2 test in children (<16 years old), depending on their clinical symptoms. Epidemiological and clinical data were obtained from the REDCap registry. Overall, 4434 SARS-CoV-2 tests were performed in symptomatic children between 1 November 2020 and 31 March 2021, 784 were positive (17.68%). We pre-processed the data to be suitable for a machine learning (ML) algorithm, balancing the positive-negative rate and preparing subsets of data by age. We trained several models and chose those with the best performance for each subset. The use of ML demonstrated an AUROC of 0.65 to predict a COVID-19 diagnosis in children. The absence of high-grade fever was the major predictor of COVID-19 in younger children, whereas loss of taste or smell was the most determinant symptom in older children. Although the accuracy of the models was lower than expected, they can be used to provide a diagnosis when epidemiological data on the risk of exposure to COVID-19 is unknown.