279例更年期综合征常见病性类证候要素Logistic回归分析

Logistic Regression Analysis of the Important Element of Syndrome of Common Disease Nature on Menopausal Syndrome,A Report of 279 Cases

目的:基于更年期综合征患者临床调查,探讨更年期综合征常见病性类证候要素与症状之间的相关性。方法:通过临床流行病学调查,收集妇科门诊且符合疾病诊断和纳入标准的更年期综合征患者,在对专家经验辨证结果规范化的基础上,提取病性类证候要素,运用二分类Logistic回归分析方法,研究常见病性类证候要素与症状之间的相关性。结果:共获取更年期综合征门诊患者279例,经规范专家经验辨证结果后,提取常见病性类证候要素7个,其回归模型显示:病性类证候要素与症状之间的相关性与传统中医的认识吻合程度较高,也有个别症状对证候要素的判断意义上存在一定的争议。结论:以更年期综合征临床流行病学调查为数据基础,借助二分类Logistic回归分析方法探讨分析常见病性类证候要素与症状之间的对应关系,是建立基于病症结合模式下以证候要素为核心的证候诊断标准体系的有效途径,如何规避二分类Logistic回归分析对多元共线的敏感性是今后研究的重点。

Objective： To investigate the associativity of the menopausal important element of syndrome and symptoms on syndrome according to clinical survey. Methods： Through clinical epidemiological investigation we collected the cases with menopausal syndrome according with diagnostic code and selected standard in department of gynecdogy. To normalize the differentiation results which many specialists concluded, we get out the important dement of syndrome of the disease nature to study their associativity on menopausal syndrome with Logistic regression analysis. Results ：279 cases with menopausal syndrome were selected, and 7 important elements of syndrome were obtained according to the specialists＇ differentiation results. The regression model showed that the associativity between the important element of syndrome and symptoms coincided with the traditional theory in Chinese medicine on menopausal syndrome, but some individual symptoms went out its way in deciding the meaning of the important element of syndrome, and may cause debate. Conclusion：Investigating and analyzing the corresponding relationship between the important element of syndrome of common disease nature and symptoms with Logistic regression analysis on the basis of clinical epidemiological investigation on menopausal syndrome is the effective way in founding the syndrome diagnostic code system around the important element of syndrome in the model of combination of disease and symptoms. And how to avoid the sensitivity of multiple collinear in Logistic regression analysis is the emphasis in future study.