Statistical analysis
The repositive rate of SARS-CoV-2 was estimated by dividing the number of positive retest patients by the number of COVID-19 patients. Categorical variables are described as absolute numbers and percentages (%). Skewed and normally distributed continuous variables are described as the median (interquartile range [IQR] or range) and mean (standard deviation [SD]), respectively. Chi-square tests and t-tests were used to compare characteristics between positive retest patients and negative retest patients.
Univariate and multivariable logistic regression models(Tripepi, Jager, Dekker, & Zoccali, 2008) were performed to estimate odds ratios (ORs) and 95% confidence intervals (95% CIs) for associations of potential risk factors with retested positivity. Age (0-17, 18-44, 45-59, or ≥60 years), sex (male or female), continent (Asia, Africa or others), severity (asymptomatic, mild, moderate, or severe) and clinical symptoms (fever, dry cough, expectoration, myalgia, diarrhea or shortness of breath) at the first admission were included in the multivariable model.
Analyses were all performed with SAS software (version 9.4 for Windows, SAS Institute, Inc., Cary, NC, USA). Statistical tests were two-sided, and P values of less than 0.05 were considered to indicate statistical significance.