基于老年营养风险指数构建老年医院获得性肺炎的预后模型

Prognostic model of hospital acquired pneumonia in the elderly based on geriatric nutritional risk index

目的探究老年营养风险指数(GNRI)在评估老年医院获得性肺炎(HAP)患者预后中的价值,并建立一个预测老年HAP患者预后的列线图模型。方法将2013年1月—2020年1月新疆维吾尔自治区人民医院急救中心411例患者纳入本研究,并采用简单随机法按照3∶1的比例将全部患者分为建模组(308例)与验证组(103例)。应用多因素logistic回归分析建模组数据,分析老年HAP患者预后不良的独立危险因素。应用R软件构建预测老年HAP患者预后不良风险的列线图模型,并对该模型进行验证。结果单因素回归分析结果示,年龄、性别、收缩压、舒张压、心率、充血性心力衰竭、慢性肾脏病、空腹血糖、低密度脂蛋白胆固醇、GNRI评分与老年HAP患者预后不良有关(P<0.05),多因素回归分析结果示,年龄、性别、充血性心力衰竭、GNRI评分是老年HAP患者预后不良的独立风险因素(OR>1,P<0.05)。对该模型进行内部和外部的验证,建模组中AUC值为0.854(95%CI:0.827~0.881);验证组中AUC值为0.841(95%CI:0.803~0.879)。Hosmer-Lemeshow检验结果提示该预测模型具有良好稳定性:建模组队列(P=0.739)、验证组队列(P=0.469)。结论本研究所构建的预测模型可较为准确地预测老年HAP患者出现预后不良的个体化风险,具有较高的潜在临床应用价值。

Objective To explore the value of the geriatric nutritional risk index(GNRI)in evaluating the prognosis of patients with hospital-acquired pneumonia(HAP)in geriatric patients,and a histogram model was established to predict the prognosis of elderly HAP patients.Methods A total of 411 patients in the Emergency Center of the People’s Hospital of Xinjiang Uygur Autonomous Region from January 2013 to January 2020 were included in this study.All patients were divided into modeling group(308 cases)and validation group(103 cases)in a 3∶1 ratio by simple randomization method.Multivariate logistic regression analysis was used to analyze the independent risk factors for poor prognosis in elderly HAP patients.R software was used to construct a histogram model to predict the risk of poor prognosis in elderly HAP patients,and the model was verified.Results Univariate regression analysis showed that age,gender,systolic blood pressure,diastolic blood pressure,heart rhythm,congestive heart failure,chronic kidney disease,fasting blood glucose,low density lipoprotein cholesterol,GNRI scores were associated with poor prognosis in elderly HAP patients(P<0.05),multivariate regression analysis showed that age,gender,congestive heart failure,and GNRI score were independent risk factors for poor prognosis in elderly HAP patients(P<0.05).The model was validated internally and externally,and the AUC value in the modeling group was 0.854(95%CI:0.827-0.881),AUC value in the validation group was 0.841(95%CI:0.803-0.879),Hosmer-Lemeshow test results indicate that the prediction model has good stability:modeling queue(P=0.739)and validation queue(P=0.469).Conclusion The prediction model established in this study can accurately predict the individual risk of poor prognosis in elderly HAP patients,and has a high potential clinical application value.