there was no eschar and skin rash in hepatitis A population, and we believed it was unreasonable to include these variables in the analysis because odd ratios could not be obtained. Thus, we excluded these variables in final multivariate logistic regression.