高级检索

老年高血压合并糖尿病患者医院感染的Nomogram预测模型建立

A Nomogram model establishment for noscomial infection in elderly patients with hypertension and diabetes mellitus

    摘要
  • 摘要:
      目的  分析老年高血压合并糖尿病患者发生医院感染（HAI）的危险因素，并建立预测HAI的Nomogram模型。
      方法  回顾性分析148例老年高血压合并糖尿病患者的临床资料，根据是否发生HAI将患者分为HAI组和非HAI组。通过单因素、多因素Logistic回归分析筛选出HAI发生的独立危险因素，同时应用Nomogram法对各个因素进行评分，构建预测模型。应用受试者工作特征（ROC）曲线评估模型的预测价值，并对模型进行内部验证，计算一致性指数（C-index），绘制校准图。
      结果  148例患者平均年龄为（64.21±12.84）岁，其中32例（21.62%）患者发生HAI。单因素分析显示，年龄、吸烟、共患病程、血压控制情况、血糖控制情况、合并其他基础疾病、急性生理学与慢性健康状况评分系统Ⅱ（APACHEⅡ）评分、意识障碍、白蛋白水平均是HAI发生的影响因素（P < 0.10）。多因素Logistic回归分析显示，共患病程≥10年（OR=3.589，95% CI为1.056~12.193，P=0.041），血糖控制不达标（OR=4.538，95% CI为1.287~16.002，P=0.019），合并其他基础疾病（OR=8.893，95% CI为2.624~30.132，P < 0.001），APACHEⅡ评分≥20分（OR=6.259，95% CI为1.934~20.256，P=0.002），存在意识障碍（OR=9.365，95% CI为2.744~34.477，P=0.001）均是HAI发生的独立危险因素。基于上述预测因子建立Nomogram预测模型，经验证，该模型贴合度、C-index、ROC曲线下面积均良好，提示该模型具有良好的预测效能和区分度。
      结论  共患病程≥10年、血糖控制不达标、合并其他基础疾病、APACHE Ⅱ评分≥20分、存在意识障碍均是老年高血压合并糖尿病患者发生HAI的独立危险因素，基于上述危险因素建立的Nomogram模型具有良好的预测效能，可为HAI防控工作提供参考依据。

     

    Abstract:
      Objective  To investigate the risk factors of hospital-associated infection (HAI) in elderly patients with hypertension and diabetes mellitus, and to establish a nomogram model for HAI.
      Methods  A retrospective study was performed to analyze the clinical data of 148 elderly patients with hypertension complicated with diabetes mellitus. The patients were divided into the HAI group and non-HAI group according to occurrence of HAI. Univariate analysis and multivariate Logistic regression analysis were used to screen out the independent risk factors of HAI occurrence. Then, each factor was scored by Nomogram method to construct the prediction model. Receiver operating characteristic (ROC) curve was drawn to assess the predictive value of the established Nomogram. Furthermore, the predictive ability of the Nomogram model was internally validated by calculating the C-index and the calibration plot was drawn.
      Results  The mean age of 148 patients was (64.21±12.84) years, and 32 patients (21.62%) developed HAI. Univariate analysis showed that the occurrence of HAI was correlated with age, smoking, disease duration of comorbidities, blood pressure and blood glucose control state, other underlying diseases, APAHEⅡ scores, consciousness state and albumin levels (P < 0.05). The multivariate Logistic regression analysis showed that disease duration of comorbidities ≥10 years (OR=3.589, 95%CI, 1.056~12.193, P=0.041), blood glucose control substandard (OR=4.538, 95%CI, 1.287~16.002, P=0.019), other underlaying diseases (OR=8.893, 95%CI, 2.624~30.132, P < 0.001), APACHEⅡ score ≥20 (OR=6.259, 95%CI, 1.934~20.256, P=0.002), consciousness disorder (OR=9.365, 95%CI, 2.744~34.477, P=0.001) were independent risk factors for HAI occurrence. Based on above risk factors in Nomogram model, statistical analysis showed that this model had a good discrimination, C-index value and the area under the ROC curve, indicating that the nomogram model had better predictive performance and differentiation.
      Conclusion  The disease duration of comorbidities ≥10 years, substandard blood glucose control, other underlying diseases, APACHEⅡscore ≥20, consciousness disorder are independent risk factors for HAI occurrence in in elderly patients with hypertension and diabetes mellitus. Nomogram model based on these risk factors has good predictive efficacy and important clinical value, and can provide reference for prevention and control of HAI.

     

    • HTML全文
    • 参考文献(25)
    • 相关文章
    • 施引文献
  • 资源附件(0)

/

返回文章
返回