吸取中医元素的高血压病风险预警系统构建

Establishment of an early warning system for hypertension based on elements of traditional Chinese medicine

目的:建立新的涵盖中医元素的高血压病风险预警系统模型。方法:收集2015年3月至2016年2月期间于郑州市中医院就诊的高血压病患者(200例)和无高血压病体检者(100名),综合运用线性、非线性数据处理方法,筛选出涵盖中医元素的相关危险因素,构建高血压病中医风险预警模型。通过BP人工神经网络预测模型与回归模型进行比较,评价高血压病中医风险预警模型的预测能力。结果:(1)饮食具有偏酸、偏清淡爱好、体质偏向阳虚质、沉脉的人群患高血压病的风险相对较低;(2)饮食具有偏咸、偏腌制爱好,体质偏向痰湿质、气虚质及气郁质,有齿痕、舌下脉络怒张,脉长中脉及脉搏有力、弦/硬脉,D型人格以及负性事件值得分>0的人群患高血压病的风险较高。(3)BP人工神经网络预测模型的预测能力优于回归模型。结论:吸取中医元素的高血压病BP人工神经网络预警模型精度和效度优于回归模型,能够更好的预测和防治高血压病。

Objective： To establish a new model of hypertension warning system with TCM elements. Methods： The patients with （200 cases） or without （100 cases） hypertension who visited Zhengzhou TCM Hospital from March 2015 to February 2016 were collected. The related risk factors with TCM elements were selected to construct the hypertension risk early warning model by the integrated use of linear and non-linear data processing method. The predicted ability of hypertension risk early warning model was evaluated through the comparison between BP artificial neural network predicted model and the regression model. Results： ①These people who had meta-acid diet, light diet, yang deficiency constitution and deep pulse suffered from low risk of hypertension. ②These people who had high-salt diet, pickled foods, phlegm-dampness constitution, qi deficiency constitution, qi-stagnation constitution, teeth-printed tongue, sublingual vein engorgement, the strong, stringy and hard radial pulse, type D personality and the greater than zero point of negative event relatively suffered from high risk of hypertension. ③The predictive ability of neural network predicted model was better than the regression model. Conclusion： The BP artificial neural network early warning system model of hypertension which contains TCM elements is better than the regression model in the accuracy and validity, and it can batter predict and prevent hypertension.