论文部分内容阅读
目的构建手术部位感染风险预警模型,为筛查高危人群、发现疑似感染病例提供帮助。方法回顾性收集2013年1月—2015年12月国内6所医院5 067例腹部外科手术患者病例资料,将所有病例按照6︰4的比例随机分为建模组和验证组,采用logistic回归建立预警模型,以ROC曲线下面积(AUC)评价模型的判别能力,以约登指数最大作为最佳截断点。结果高危患者预警模型AUC为0.823,灵敏度为78.81%,特异度为74.33%,阳性预测值为19.67%,阴性预测值为97.78%。疑似感染病例判别模型AUC为0.978,灵敏度93.38%,特异度95.62%,阳性预测值62.95%,阴性预测值为99.45%。结论此研究建立的预警模型判别能力较好,可以为医院感染信息系统的预警和判别开发提供参考。
Objective To establish an early warning model of the risk of surgical site infection and provide help for screening high-risk population and finding suspected cases of infection. Methods The data of 5 067 cases of abdominal surgery in 6 hospitals in China from January 2013 to December 2015 were retrospectively collected. All cases were randomly divided into model group and validation group according to the ratio of 6: 4, and were established by logistic regression Early warning model to assess the model under the area under the curve (AUC) ROC ability to discriminate, the Youden index maximum as the best cut-off point. Results The AUC of high-risk patients was 0.823, the sensitivity was 78.81%, the specificity was 74.33%, the positive predictive value was 19.67% and the negative predictive value was 97.78%. The AUC of suspected infection cases was 0.978, the sensitivity was 93.38%, the specificity was 95.62%, the positive predictive value was 62.95% and the negative predictive value was 99.45%. Conclusion The early warning model established by this study has a good ability of discriminating, which can provide a reference for early warning and discriminant development of hospital infection information system.