Abstract: Objective　To construct a model of "heart failure-hypertension" comorbidity risk assessment in elderly people based on medical big data and analyze its feasibility. Methods　A total of 500 hypertensive patients treated in The Second Medical Center, Chinese PLA General Hospital from January 1, 2020 to June 1, 2021 were selected as subjects and were divided into a heart failure group (124 cases) and a non-heart failure group (376 cases) according to whether combined with heart failure. In the heart failure group, there were 88 males and 36 females, with an age of (74.25±4.38) years old; in the non-heart failure group, there were 234 males and 142 females, with an age of (73.66±5.25) years old. The clinical data of the two groups were analyzed retrospectively. Univariate and multivariate logistic regression analysis were used to analyze the comorbidity risk of "heart failure-hypertension" and model it, and the receiver operating characteristic curve (ROC) was used to test the model effect. χ² test was used for the count data and independent sample t test was used for the measurement data. Results　Blood pressure level, body mass index (BMI), concomitant infection, concomitant hyperlipidemia, concomitant hyperuricemia, concomitant left ventricular hypertrophy, and smoking were influence factors for heart failure-hypertension comorbidity (OR=1.071, 0.852, 1.835, 1.978, 2.002, 2.552, and 1.988; all P<0.05). The area under the curve (AUC) was 0.852, the sensitivity was 0.685, the specificity was 0.835, and the Youden index was 0.521. The AUC was over 0.7, which indicated a moderate predictive efficiency. Conclusions　The prevalence of heart failure in elderly hypertensive patients is high, blood pressure level, BMI, concomitant infection, concomitant hyperlipidemia, concomitant hyperuricemia, concomitant left ventricular hypertrophy, and smoking are influence factors for heart failure-hypertension comorbidity, and the risk assessment model has a good predictive efficiency.

Key words: Hypertension, Heart failure, Comorbidity, Risk assessment

摘要： 目的　在医疗大数据背景下构建老年“心衰-高血压”共病风险评估模型，并对其可行性进行分析。方法　选取2020年1月1日至2021年6月1日解放军总医院第二医学中心收治的高血压患者500例作为研究对象，按照是否合并心衰分为合并心衰组（124例）和未合并心衰组（376例）。合并心衰组男88例、女36例，年龄（74.25±4.38）岁；未合并心衰组男234例、女142例，年龄（73.66±5.25）岁。对临床资料进行回顾性分析。采用单因素和logistic多因素回归分析对“心衰-高血压”共病风险进行分析和模型建立，采用受试者工作特征曲线（ROC）检验模型效果。计数资料采用χ²检验，计量资料采用独立样本t检验。结果　血压水平、体质量指数（BMI）、合并感染、合并高脂血症、合并高尿酸血症、合并左心室肥厚、吸烟是心衰-高血压共病的影响因素（OR=1.071、0.852、1.835、1.978、2.002、2.552、1.988；均P<0.05）。曲线下面积（AUC）为0.852，灵敏度为0.685，特异度为0.835，约登指数为0.521，AUC >0.7，预测效能中等。结论　高血压患者心衰的患病率较高，血压水平、BMI、合并感染、合并高脂血症、合并高尿酸血症、合并左心室肥厚、吸烟是心衰-高血压共病的影响因素，风险评估模型预测效能较好。

关键词: 高血压, 心衰, 共病, 风险预测