Using Artificial Intelligence in the COVID-19 Pandemic: A Systematic Review

Abstract

Artificial intelligence applications are known to facilitate the diagnosis and treatment of COVID-19 infection. This research was conducted to investigate and systematically review the studies published on the use of artificial intelligence in the COVID-19 pandemic. The study was conducted between April 25 and May 6, 2020 by scanning national and international studies accessed in "Web of Science, Google Scholar, Pubmed, and Scopus" databases with the keywords ("Coronavirus" or "COVID-19") and ("artificial intelligence" or "deep learning" or "machine learning"). As a result of the scanning process, 1495 (Google Scholar: 1400, Pubmed: 58, Scopus: 30, WOS: 7) studies were accessed. The studies were first examined according to their titles, and 1385 studies, which were not related to the research topic, were not included in the scope of the research. 50 articles, which did not meet the inclusion criteria, were excluded. The abstract and complete texts of the remaining 60 studies were scanned for the study's inclusion and exclusion criteria. A total of 10 studies, consisting of reviews, letters to the editor, meta-analysis studies, animal studies, conference presentations, studies not related to COVID-19, and incomplete studying protocols, were excluded. There were 50 studies left. 9 articles with duplication were identified and excluded. The remaining 41 studies were examined in detail. A total of 26 studies were found to meet the criteria for the systematic review study. In this systematic review, AI applications were found to be effective in COVID-19 diagnosis, classification, epidemiological estimates, mode of transmission, distribution, the density of lesions, case increase estimation, mortality/mortality risk, and early scans.