Correlation And Constitution Between Hyperuricemia And Constitution Of Traditional Chinese Medicine Study On The Construction Of Discriminant Model

Objective:On the one hand,this study aims to explore the characteristics of TCM constitution distribution and related influencing factors of hyperuricemia,so as to provide clinical basis for the intervention and prevention of hyperuricemia.On the other hand,the role of machine learning in the construction of TCM constitution discrimination model was preliminarily discussed to provide a new idea for improving the accuracy of TCM constitution discrimination.Methods:In this study,a case-control study was used to select patients who had undergone traditional Chinese medicine constitution identification and filled out the health risk factor assessment form from the medical examination personnel of the First Affiliated Hospital of Heilongjiang University of Chinese Medicine Center for Prevention and Treatment of Disease as the research objects.In this study,a total of 1,942 patients with physical examination data(including the constitution identification questionnaire and health risk factor assessment form)were selected,and 158 patients with incomplete data were excluded.Among them,258 patients met the inclusion criteria of the high uric acid group and 294 patients met the inclusion criteria of the non-high uric acid group.The contents of TCM constitution,biochemical indexes,body index,dietary habits and previous history were selected to study the common biased constitution types of hyperuricemia and the risk factors affecting the elevated blood uric acid level,and to analyze the effect of machine learning to construct the TCM constitution discrimination model.Results:1.The occurrence frequency of TCM constitution types in the high uric acid group from high to low was peaceful quality(37.98%),qi-deficiency quality(25.19%),Yang deficiency quality(12.79%),damp-heat quality(8.53%),phlegm-dampness quality(7.36%),yin-deficiency quality(3.49%),qi-stagnation quality(2.71%)and intrinsic quality(1.16)%),blood stasis(0.78%);In the non-hyperuric group,The occurrence frequency of TCM constitution types from high to low in order is peaceful quality(31.97%),Yang deficiency quality(28.91%),qi deficiency quality(16.67%),damp-heat quality(6.46%),qi stagnation quality(5.44%),blood stasis quality(4.42%),Yin deficiency quality(2.38%),phlegm-dampness quality(2.04%),and intrinsic quality(1.70%).2.There were 193 males(74.81%)and 65 females(25.19%)in the hyperuric acid group,and 26 males(8.84%)and 268 females(91.16%)in the non-hyperuric acid group.There was a significant difference between the two groups.3.There were significant differences in body index(weight,BMI,abdominal circumference,blood pressure),biochemical indexes(uric acid,blood glucose,blood lipid),previous history of hypertension,smoking history,drinking history,psychological stress,water intake,meat intake,staple food intake and meal time between the high uric acid group and the non-high uric acid group;4.In machine learning,based on the personal history,bad living habits,biochemical indices,body index,the dietary habits of the KNN algorithm to build physical discrimination model for the traditional Chinese medicine of traditional Chinese medicine constitution classification accuracy of 31.6%,32.1%,31%,28.9%,27.4%,random forest,support vector machine(SVM),neural network model and neural network model for 1 to 2 Chinese medicine constitution classification accuracy of 100%,47.64%,34% and 35.8%respectively.Conclusion:1.In this study,the male patients with hyperuricemia were more than the female ones,and their TCM constitutions were mainly composed of placid quality,qi-deficiency quality,yang-deficiency quality,phlegm-dampness quality and damp-heat quality,among which qi-deficiency quality,damp-heat quality and phlegm-dampness quality were the common biased constitutions.2.Increased blood pressure,blood glucose and lipid levels,overweight,abdominal obesity,smoking history,drinking history,high psychological stress,irregular meal time and staple food,and excessive meat intake may be the risk factors affecting the elevated blood uric acid level.3.In machine learning,the random forest algorithm has a high accuracy for constructing the TCM constitution discrimination model,and the algorithm may improve the accuracy of TCM constitution discrimination.