Continuous variables were expressed as the mean (SD) or median (interquartile range, or IQR). Student t test was used to compare data with a normal distribution, and the Mann-Whitney U test was used to analyze data with a skewed distribution. Categorical data were expressed as frequency distributions, and χ2 or Fisher test was used to identify statistically significant differences. To control potential confounding factors, we included variables other than antibiotic exposure associated with a drug-resistant infection at a significance level of P ≤ .2 in the univariate analysis,and variables of potential clinical importance were included in a multivariable logistic regression model. Antibiotic exposures were examined in 2 ways, as qualitative data (ie,exposed or not exposed) or quantitative data (ie, number of days of therapy with each antibiotic). Because several antibiotics are often concomitantly or sequentially used to treat infections, antibiotic groups showing a significant association with outcome were included in the multivariate analysis. All statistical tests were 2 tailed. A P value of less than .05 was considered statistically significant. All data analyses were performed using the SPSS software package(SPSS Inc, Chicago, Ill).