Clinical manifestations of COVID-19; what have we learned from the global database?

Background: There is a need to analyze a worldwide database of the coronavirus disease of 2019 (COVID-19) pandemic.This may prove valuable to facilitate better strategies and planning on prevention, screening, surveillance, early diagnosis, containment and treatments. Method: We extracted 14,259 case reports of COVID-19 dated 11th November 2019 to 18th March 2020 from Johns Hopkins University Repository Online Databaseof 58 countries. After extensive data preprocessing, a multi-disciplinary expert researcherthen conducted series of vetting to categorizefree-text description of symptoms into discreet standardizedcategories.Continuous variables were presented by using median and inter-quartile range whereas categorical variables were presented by frequency and percentage. Result: A total of 2191 cases (15.4%) were included for demographic analysis. The median age was46 years (IQR26 years) with 787 (35.9%) cases involved patients aged of 60 and above while patients less than18 years of age were reported in 79 (3.6%) cases. Majority of the patients were males (n=1227, 56.7%). There were a total of 20standardized categories ofCOVID-19symptoms.The most prevalent were fever (74.8%), nonproductive cough (42.2%), fatigue (13.1%), sore throat (12.8%) and shortness of breath (11.7%). Other symptoms with frequency of more than 1% were chest discomfort, nasal congestion, muscular pain, chills and rigors, headache, diarrhoea, expectoration and joint pain. Other more uncommon symptoms reported include loss of appetite, conjunctivitis, toothache and abdominal pain. Asymptomatic manisfestations were reported in 8 cases (1.0%).All population are susceptible to COVID-19 especially the older age group. There were 20 standardized categories of symptoms wherefever, non-productive cough, fatigue, sore throat and shortness of breath were the most commonly reported. Conclusion: Findings of this study contribute to a deeper understanding on COVID-19 and may prove useful for researchers to better-design screening and surveillance strategies via more accurate risk-prediction modelling.