冠心病心绞痛血瘀证的药效学解构与动态演变的多维度诠释
|
摘要
证候具有“内实外虚”的特征。所谓“实”,是指最能反映该病机的权重最大的关键内容,是群体在某一特定病变过程中所具有的共性规律,是干预的依据。“虚”则指具体某一患者所表现出的一系列个性化症状信息,它涵盖了所有能够表达个性化的内容,对干预原则和方法具有一定的影响作用。因此,药物干预证候的症状靶位不同,对证候的疗效判定也具有不同的意义。尼莫地平法是临床研究中评价证候疗效的常用方法。采用这种方法评价两种药物治疗冠心病心绞痛血瘀证的证候疗效,治疗前后两种药物的证候改善率均为50%,认为两种药物的证候疗效是相等的,但如果一种药物改善的主要是“胸痛”,另一种药物改善的主要是“脉细涩”。根据证候“内实外虚”理论,第一种药物治疗冠心病心绞痛血瘀证的效果要明显优于第二种药物。因此,本研究提出评价证候疗效,首先要进行证候的解构,明确证候的“内实”,分析证候“内实”的动态变化趋势。
证候是随着时空的变迁而演化的过程流,随着时间的推移和状态的变化,证候有可能发生轻重程度的变化以及由此及彼的改变。证候的“内实外虚”也不是固定不变的,也具有随着时间的变化而不断变化的趋势。分析证候整体、证候“内实”以及单一症状的动态变化趋势,有利于探索证候变化的拐点,评价证候疗效,分析药物干预的症状靶位。
因此,本研究在理解王永炎院士证候“内实外虚”理论的基础上,以心血康胶囊治疗冠心病心绞痛血瘀证的随机、双盲、多中心临床试验为研究模型,重视证候的整体性、结构性、动态性和非线性特征,为探索中医证候疗效评价方法进行了初步的尝试。
1、研究目的：通过对证候的药效学解构,分析证候整体及其结构的动态变化趋势,探索随着时间的变化药物对证候和证候“内实”动态变化的影响,为进一步探索证候状态变化的拐点、评价中医证候临床疗效提供方法学方面的支撑。
2、研究方法：本研究以心血康胶囊治疗冠心病心绞痛血瘀证的随机、双盲、多中心临床试验的数据作为研究对象,采用潜在类别模型对冠心病心绞痛血瘀证进行解构,分析冠心病心绞痛血瘀证的潜在类别及其“内实”部分。根据数据类型,针对有序分类变量,采用可以处理非正态、非独立数据的分析重复测量资料的常用方法——广义估计方程对不同药物干预下冠心病心绞痛血瘀证潜在类别的变化和构成冠心病心绞痛血瘀证的单一症状的变化进行分析;针对证候、证候“内实”等多维数据的变化,采用相似度分析方法,通过相似度的变化探索不同药物干预下的证候和证候“内实”动态变化趋势,在此基础上,以非线性混合效应模型拟合相似度随着时间变化的过程。
3、研究结果
3.1基线分析
通过对入组时冠心病心绞痛轻重程度、血瘀证的轻重程度、理化检查指标等的基线水平比较分析,发现治疗组和对照组两组之间具有可比性。
3.2一般统计分析
分别采用t检验和单变量重复测量资料方差分析探索不同时间点和不同药物干预下中医证候的变化情况,t检验结果发现除第2周外,其余时间点的中医证候总分与基线的差异有统计学意义,3个月随访中医证候总分在两组间的差异是有统计学意义的。
单变量重复测量资料方差分析结果发现：治疗后2周至随访3个月各时点,证候证候总分与基线的差异有统计学意义,治疗后4周、6周、8周及3个月随访的中医证候积分与基线的差异在A、B’两组之间有统计学意义。
采用t检验和单变量重复测量资料方差分析不同时间点和不同药物干预下中医证候的变化情况,发现两种统计分析方法得到的结果存在一定的差异,后者是分析重复测量资料常用的方法之一,得到的结果比较可靠。
3.3冠心病心绞痛血瘀证的解构
采用潜在类别模型对冠心病心绞痛血瘀证进行解构,分析冠心病心绞痛血瘀证证候的潜在分类以及证候的“内实”部分。结果发现：冠心病心绞痛血瘀证的潜在分类包括两类,一类为轻证,一类为重证,其中轻证的概率是0.6320,重证的概率是0.3680。胸痛、胸闷、唇色紫暗、舌质紫暗是冠心病心绞痛血瘀证的“内实”部分。
3.4证候的动态变化与证候状态变化的拐点
采用夹角余弦法、Dice系数法、Jaccard系数法分别计算不同时间点不同组别的冠心病心绞痛血瘀证相似度值,以相似度值代替证候积分分析证候的动态变化,以非线性混合效应模型,探索证候动态变化趋势,分析证候状态变化的拐点。结果发现：
不同方法计算的证候相似度值总体变化趋势基本一致,即随着治疗时间的延长,A、B两组证候相似度值均逐渐下降；
证候的变化受到心绞痛积分、证候“内实”、试验中心的影响,说明地域不同,A、B两种药物对“血瘀证”的疗效存在一定的差异。心绞痛程度越重,药物对“血瘀证”的疗效越差；
A、B两组证候变化存在差异,A药物治疗冠心病心绞痛血瘀证的疗效略优于B药物；
以相似度值等于0.8作为界点,计算相此时药物使用时间,结果显示群体临床试验过程中,A药使用7.28周,B药使用10.72周时,入组时的冠心病心绞痛血瘀证状态发生变化。
3.5证候潜在类别的动态变化
通过潜在类别分析和广义估计方程分析,探索证候潜在类别的变化,结果显示：第2、4、6、8周和3个月随访的潜在类别与基线相比,均有统计学意义,并且相邻两个时间点潜在类别的差异也具有统计学意义。同时,在第4周时,时间与组别存在交互作用。说明A、B两种药物均可以缓解冠心病心绞痛血瘀证,并且随着治疗时间的延长,冠心病心绞痛血瘀证缓解越明显。A、B两组相比,在治疗后的第4周时潜在类别的变化有统计学差异。
3.6证候“内实”的动态变化
采用夹角余弦法、Dice系数法、Jaccard系数法分别计算不同时间点不同组别的冠心病心绞痛血瘀证“内实”相似度值,以相似度值代替证候积分分析证候“内实”的动态变化,以非线性混合效应模型,探索证候“内实”动态变化趋势。结果发现：
随着治疗时间的延长,A、B两组证候“内实”的相似度值均逐渐下降,说明A、B两种药物都可以干预证候的“内实”部分。
A药物干预下证候“内实”相似度的变化受到试验中心的影响,与心绞痛积分无关,说明无论心绞痛程度轻重,A药物均能较好的改善血瘀证的“内实”部分,并且在不同的研究中心,药物对“证候”内实部分的影响不同,说明地域不同,A药物对“血瘀证”内实的疗效存在一定的差异。
B药物干预下证候“内实”相似度的变化受到心绞痛积分和试验中心的影响,说明地域不同,B药物对“血瘀证”内实的疗效存在一定的差异。同时,心绞痛程度越重,B药物对“血瘀证”内实的疗效越差。
A、B两组证候“内实”的变化存在差异,A药物干预冠心病心绞痛血瘀证“内实”的疗效略优于B药物。
3.7单一症状的动态变化
分别采用秩和检验和广义估计方程对冠心病心绞痛血瘀证单一症状分析,秩和检验结果发现：除2周、4周B组的脉细涩外,其他时间点各项中医症状评分与其自身前后的差异有统计学意义。4周、6周、8周时舌质紫暗、脉细涩在两组间的差异有统计学意义,随访3个月时胸痛、舌质紫暗与脉细涩在两组间的差别有统计学意义。
广义估计方程分析显示：6个单一症状在治疗后2周、4周、6周、8周、随访3个月与基线相比以及相邻2个时间点相比,症状积分变化有统计学意义,说明A药和B药都具有改善冠心病心绞痛血瘀证相关症状的作用,并且治疗时间越长,药物的疗效越好。同时发现,第20周胸痛的变化、2周、4周、8周及3个月随访心悸的变化、4周起舌质紫暗的变化在两组之间存在差异。
采用秩和检验和广义估计方程分析不同时间点和不同药物干预下中医症状的变化情况,发现两种统计分析方法得到的结果存在一定的差异。后者是分析重复测量资料常用的方法之一,得到的结果比较可靠。
4、结论
4.1症状、证候潜在类别、证候的“内实”以及证候整体都具有时空演变的特征。证候疗效可通过分析证候“内实”的动态变化、证候整体的动态变化和单一症状的动态变化进行评价。
4.2采用潜在类别模型等方法分析不同症状对证候的贡献度,可以用来探索证候的“内实”,潜在类别模型是诠释证候“内实外虚”特征的方法之一
4.3研究资料属于重复测量数据时,要充分考虑不同观测时间之间的相关性,注重采用适合于重复测量数据的统计方法进行分析。
4.4采用相似度分析结合非线性混合效应模型可有效地分析证候的高维性和非线性特征,探索证候状态变化的拐点。
4.5本研究揭示了证候具有结构性和临界性的特征,为辨证论治疗效评价方法学提供了部分的理论依据和技术支撑。
本研究首次提出对证候进行解构,指出证候“内实”的变化是证候疗效评价的主要指标,并采用潜在类别模型探索证候的“内实外虚”,利用相似度分析和非线性混合效应模型探索药物干预下证候的动态变化趋势及证候状态变化的拐点。课题采用相似度分析和NONMEM分析证候的动态变化趋势,是符合证候特征的研究方法,但由于证候的复杂性,我们的研究还需深入,开展进一步的研究。证候“外虚”的变化与证候疗效评价的关系,并未探讨,是本课题下一步主要研究的方向。
Syndrome has the characteristics of "internal stability and external flexibility". The "internal stability" means the key syndrome that reflects pathogenesis and the common character in the group, which is the basis of intervention. The "external flexibility" means the individual syndrome, which can affects the principle and method of intervention. Therefore, if the syndrome target is different, the significance to evaluate efficacy is different. Nimodipine is the common method to evaluate the syndrome efficacy, which doesn't conform to the characteristics of "internal stability and external flexibility". So the study proposed that deconstruction to syndrome and analysis to "internal stability" dynamic changes is the key part to evaluate syndrome efficacy.
Syndrome is the dynamic process which varies with time, and so are the"internal stability" and "external flexibility" parts of Syndrome. The recognization of the dynamic change trend of syndrome especially "internal stability" and "external flexibility" part of Syndrome is the important foundation for the application into the clinical practice and enhancement of the treatment effect of TCM, and the critical process to evaluate the efficacy of the syndrome. Based on the theory of "internal stability and external flexibility" proposed by Pro. Wang Yongyan, a randomized, double-blind, multi-center trial of Xinxuekang Capsule for blood stasis syndrome of angina pectoris of coronary heart disease was conducted to find an approach to evaluating the effect in TCM syndrome.
Object:To discuss the effect of the drug for Syndrome and the "internal stability" part of Syndrome under the change of time by clinical annotation to "internal stability and external flexibility" of Syndrome and its dynamic change, which would provide new method to evaluate the efficacy of the syndrome.
Method:Based on the randomized, blind, multi-center clinical trial data of Xinxuekang capsule for the treatment of blood stasis syndrome in angina pectoris of coronary heart disease (CHD-AP-BSS),the study determined the latent class and the"internal stability" part of CHD-AP-BSS according to the latent class modeling (LCM). According to the class of the data, the study analyzed the change of latent class and single symptom of CHD-AP-BSS by Generalized Estimating Equations (GEE).And then the study discussed the dynamic change trends of syndrome and "internal stability" part of syndrome by similarity analysis under the different interventions based on the dynamic changes of dimensional data such as syndrome and "internal stability" part of syndrome. At last, the study applied NONMEM into the analysis of the changes in similarities with time.
Results:
1. According to the comparisons of the baseline in A and B2 groups, there were not any significant differences in CHD-AP, BSS and other test results between the two groups.
2. By t-test analysis, the study found that the changes of the syndrome score between 4w,6w,8w or 3m after treating and baseline were significant, respectively. The changes of the syndrome score between A and B group at the 3 m after treating were significant, respectively.
By single factor analysis of variance of repetitive measurement, the study found that syndrome score decreased by time. The changes of the syndrome score between 2w,4w,6w,8w or 3m after treating and baseline were significant, respectively. The changes of the syndrome score between A and B group at the 4w,6w,8w or 3m after treating were significant, respectively.
The results of t-test analysis and single factor analysis of variance of repetitive measurement is different. The latter is the common method to analyze the repetitive measurement data, which the results is dependable.
3. Based on the results of LCM to CHD-AP-BSS, the study found that the patients can be classified to "the mild" and "the serious". The symptoms like chest pain, chest distress, dark purple lip, dark tongue were relatively more stable than other symptoms in CHD-AP-BSS, which is the "internal stability" part of CHD-AP-BSS.
4. Based on the results of similarity analyses to CHD-AP-BSS at different times and in different groups by Cosine method, Dice method and Jaccard method, the study found that the total change trend of the similarity values was similar by different methods, although the details of similarity values were different in different methods, which shows that the similarity values were decreased gradually in both A and B groups with the time of treatment. It suggested that both A and B group can improve the blood stasis syndrome.
Analyzing syndrome changing trend with time by nonlinear mixed effected model, the research found that similarity of syndrome changed with time in both A and B groups. The change of the similarity was influenced by angina pectoris scores, "internal stability" part of syndrome, et al. At the same time, the factor "group" was related to the change of syndrome, which suggested that the change of syndrome between A group and B group was different significantly.
Take the similarity values equal to 0.8 as the boundary point, the study found that at that time, A is used 7.28w and B is used 10.72w, which is considered to the change of CHD-AP-BSS.
5. By LCM and GEE, the research found that the latent class in 2w,4w,6w,8w and 3m was significantly different compared with the baseline, respectively. At the 4w, the factor "time" and "group" existed interaction. The results illustrated that CHD-AP-BSS alleviates gradually. The change of latent class in 4w after treatment between A group and B group is different significantly.
6. Based on the results of the similarity values of "internal stability" part of CHD-AP-BSS at different time and in different groups by Cosine method, Dice method and Jaccard method, the research found that although the similarity values were different by different methods, its total change trend was similar in each method. It suggested that the similarity values were decreased gradually in both A and B groups with the time of treatment, which illustrated that both A and B group can improve the "internal stability" part of CHD-AP-BSS.
According to the analysis of the correlation between the change of "internal stability" part of CHD-AP-BSS and time by NONMEM, the results showed that similarity of "internal stability" part of syndrome changed with time. The factor "test centre" was an influence factor for the similarity change of "internal stability" part in both A and B group. The factor "angina pectoris scores" affected the similarity change of "internal stability" part of CHD-AP-BSS only in B group. The difference of similarity change of "internal stability" part of CHD-AP-BSS between group A and group B was significant.
7. According to the analysis of the single symptoms of CHD-AP-BSS by rank test, the study found that the changes of symptom scores of chest pain, chest distress, palpitation, dark purple lip, tongue and pulse manifestations in 2w,4w,6w,8w and 3m were significantly different significantly compared with the baseline except pulse manifestations in 2w,4w in B group. The differences of chest pain in 20w, palpitation in 2w,4w,8w and 20w, tongue picture from 4w to 20w between group A and group B were significant.
According to the analysis of the single symptoms of CHD-AP-BSS by GEE, the research found that the changes of symptom scores of chest pain, chest distress, palpitation, dark purple lip, tongue and pulse manifestations in 2w,4w,6w,8w and 3m were significantly different significantly compared with the baseline, and the differences between the symptoms scores at the two adjacent time points were also different significantly. It indicated that both drug A and B could improve the symptoms of CHD-AP-BSS and the longer the patients were treated, the better the efficacy would be got. The differences of chest pain in 20w, palpitation from 2w to 20w, tongue picture from 4w to 20w between group A and group B were significant.
The results of rank test analysis and single factor analysis of variance of repetitive measurement is different. The latter is the common method to analyze the repetitive measurement data, which the results is dependable.
Conclusion
1. Symptoms contributing to the syndrome, the latent class and the "internal stability" part of the syndrome and the syndrome itself are in constant change with space and time. The effect in TCM syndrome can be evaluated by analyzing the change of contributing symptoms, the "internal stability" part of the syndrome and the syndrome itself。
2. LCM can analyze the contribution of symptom to syndrome, which can find out the "internal stability" part of the syndrome. LCM is one of the effective methods to annotate, the "internal stability" part of the syndrome.
3. The data type should be considered carefully during analysis. When the data belongs to repeated measurement data, the relationship between data in different observation point must be considered. The analysis method should be fit for repeated measurement data.
4. The integration method of similarity analysis and NONMEM can better display the high dimensionality and nonlinear character of syndrome and find out the inflection point of syndrome change.
5. The study suggested that syndrome has the character of constitutive property and criticality, which provides part theoretical foundation and technical support for efficacy evaluation of treatment based on syndrome differentiation.
The study proposed that the analysis to the dynamic change of "internal stability and external flexibility" of syndrome was the entry point to evaluate the effect in TCM syndrome for the first time. Determine the "internal stability and external flexibility" of syndrome by LCM. Determine the dynamic change trend and the knee point of the change in syndrome by similarity analysis and NONMEM. The approach of analysis to dynamic change trend by similarity analysis and NONMEM is suitable for syndrome study, which is one of the entry points to evaluate the effect in TCM syndrome. But due to the complicated property of the syndrome, the further study is needed in the future.2. The study suggested that the change of "internal stability" part of syndrome was the main indicator to evaluate the effect in TCM syndrome. However, the relationship between the change in "external flexibility" part of syndrome and the effect of syndrome isn't discussed in this study, which is the direction to further study.
证候是随着时空的变迁而演化的过程流,随着时间的推移和状态的变化,证候有可能发生轻重程度的变化以及由此及彼的改变。证候的“内实外虚”也不是固定不变的,也具有随着时间的变化而不断变化的趋势。分析证候整体、证候“内实”以及单一症状的动态变化趋势,有利于探索证候变化的拐点,评价证候疗效,分析药物干预的症状靶位。
因此,本研究在理解王永炎院士证候“内实外虚”理论的基础上,以心血康胶囊治疗冠心病心绞痛血瘀证的随机、双盲、多中心临床试验为研究模型,重视证候的整体性、结构性、动态性和非线性特征,为探索中医证候疗效评价方法进行了初步的尝试。
1、研究目的：通过对证候的药效学解构,分析证候整体及其结构的动态变化趋势,探索随着时间的变化药物对证候和证候“内实”动态变化的影响,为进一步探索证候状态变化的拐点、评价中医证候临床疗效提供方法学方面的支撑。
2、研究方法：本研究以心血康胶囊治疗冠心病心绞痛血瘀证的随机、双盲、多中心临床试验的数据作为研究对象,采用潜在类别模型对冠心病心绞痛血瘀证进行解构,分析冠心病心绞痛血瘀证的潜在类别及其“内实”部分。根据数据类型,针对有序分类变量,采用可以处理非正态、非独立数据的分析重复测量资料的常用方法——广义估计方程对不同药物干预下冠心病心绞痛血瘀证潜在类别的变化和构成冠心病心绞痛血瘀证的单一症状的变化进行分析;针对证候、证候“内实”等多维数据的变化,采用相似度分析方法,通过相似度的变化探索不同药物干预下的证候和证候“内实”动态变化趋势,在此基础上,以非线性混合效应模型拟合相似度随着时间变化的过程。
3、研究结果
3.1基线分析
通过对入组时冠心病心绞痛轻重程度、血瘀证的轻重程度、理化检查指标等的基线水平比较分析,发现治疗组和对照组两组之间具有可比性。
3.2一般统计分析
分别采用t检验和单变量重复测量资料方差分析探索不同时间点和不同药物干预下中医证候的变化情况,t检验结果发现除第2周外,其余时间点的中医证候总分与基线的差异有统计学意义,3个月随访中医证候总分在两组间的差异是有统计学意义的。
单变量重复测量资料方差分析结果发现：治疗后2周至随访3个月各时点,证候证候总分与基线的差异有统计学意义,治疗后4周、6周、8周及3个月随访的中医证候积分与基线的差异在A、B’两组之间有统计学意义。
采用t检验和单变量重复测量资料方差分析不同时间点和不同药物干预下中医证候的变化情况,发现两种统计分析方法得到的结果存在一定的差异,后者是分析重复测量资料常用的方法之一,得到的结果比较可靠。
3.3冠心病心绞痛血瘀证的解构
采用潜在类别模型对冠心病心绞痛血瘀证进行解构,分析冠心病心绞痛血瘀证证候的潜在分类以及证候的“内实”部分。结果发现：冠心病心绞痛血瘀证的潜在分类包括两类,一类为轻证,一类为重证,其中轻证的概率是0.6320,重证的概率是0.3680。胸痛、胸闷、唇色紫暗、舌质紫暗是冠心病心绞痛血瘀证的“内实”部分。
3.4证候的动态变化与证候状态变化的拐点
采用夹角余弦法、Dice系数法、Jaccard系数法分别计算不同时间点不同组别的冠心病心绞痛血瘀证相似度值,以相似度值代替证候积分分析证候的动态变化,以非线性混合效应模型,探索证候动态变化趋势,分析证候状态变化的拐点。结果发现：
不同方法计算的证候相似度值总体变化趋势基本一致,即随着治疗时间的延长,A、B两组证候相似度值均逐渐下降；
证候的变化受到心绞痛积分、证候“内实”、试验中心的影响,说明地域不同,A、B两种药物对“血瘀证”的疗效存在一定的差异。心绞痛程度越重,药物对“血瘀证”的疗效越差；
A、B两组证候变化存在差异,A药物治疗冠心病心绞痛血瘀证的疗效略优于B药物；
以相似度值等于0.8作为界点,计算相此时药物使用时间,结果显示群体临床试验过程中,A药使用7.28周,B药使用10.72周时,入组时的冠心病心绞痛血瘀证状态发生变化。
3.5证候潜在类别的动态变化
通过潜在类别分析和广义估计方程分析,探索证候潜在类别的变化,结果显示：第2、4、6、8周和3个月随访的潜在类别与基线相比,均有统计学意义,并且相邻两个时间点潜在类别的差异也具有统计学意义。同时,在第4周时,时间与组别存在交互作用。说明A、B两种药物均可以缓解冠心病心绞痛血瘀证,并且随着治疗时间的延长,冠心病心绞痛血瘀证缓解越明显。A、B两组相比,在治疗后的第4周时潜在类别的变化有统计学差异。
3.6证候“内实”的动态变化
采用夹角余弦法、Dice系数法、Jaccard系数法分别计算不同时间点不同组别的冠心病心绞痛血瘀证“内实”相似度值,以相似度值代替证候积分分析证候“内实”的动态变化,以非线性混合效应模型,探索证候“内实”动态变化趋势。结果发现：
随着治疗时间的延长,A、B两组证候“内实”的相似度值均逐渐下降,说明A、B两种药物都可以干预证候的“内实”部分。
A药物干预下证候“内实”相似度的变化受到试验中心的影响,与心绞痛积分无关,说明无论心绞痛程度轻重,A药物均能较好的改善血瘀证的“内实”部分,并且在不同的研究中心,药物对“证候”内实部分的影响不同,说明地域不同,A药物对“血瘀证”内实的疗效存在一定的差异。
B药物干预下证候“内实”相似度的变化受到心绞痛积分和试验中心的影响,说明地域不同,B药物对“血瘀证”内实的疗效存在一定的差异。同时,心绞痛程度越重,B药物对“血瘀证”内实的疗效越差。
A、B两组证候“内实”的变化存在差异,A药物干预冠心病心绞痛血瘀证“内实”的疗效略优于B药物。
3.7单一症状的动态变化
分别采用秩和检验和广义估计方程对冠心病心绞痛血瘀证单一症状分析,秩和检验结果发现：除2周、4周B组的脉细涩外,其他时间点各项中医症状评分与其自身前后的差异有统计学意义。4周、6周、8周时舌质紫暗、脉细涩在两组间的差异有统计学意义,随访3个月时胸痛、舌质紫暗与脉细涩在两组间的差别有统计学意义。
广义估计方程分析显示：6个单一症状在治疗后2周、4周、6周、8周、随访3个月与基线相比以及相邻2个时间点相比,症状积分变化有统计学意义,说明A药和B药都具有改善冠心病心绞痛血瘀证相关症状的作用,并且治疗时间越长,药物的疗效越好。同时发现,第20周胸痛的变化、2周、4周、8周及3个月随访心悸的变化、4周起舌质紫暗的变化在两组之间存在差异。
采用秩和检验和广义估计方程分析不同时间点和不同药物干预下中医症状的变化情况,发现两种统计分析方法得到的结果存在一定的差异。后者是分析重复测量资料常用的方法之一,得到的结果比较可靠。
4、结论
4.1症状、证候潜在类别、证候的“内实”以及证候整体都具有时空演变的特征。证候疗效可通过分析证候“内实”的动态变化、证候整体的动态变化和单一症状的动态变化进行评价。
4.2采用潜在类别模型等方法分析不同症状对证候的贡献度,可以用来探索证候的“内实”,潜在类别模型是诠释证候“内实外虚”特征的方法之一
4.3研究资料属于重复测量数据时,要充分考虑不同观测时间之间的相关性,注重采用适合于重复测量数据的统计方法进行分析。
4.4采用相似度分析结合非线性混合效应模型可有效地分析证候的高维性和非线性特征,探索证候状态变化的拐点。
4.5本研究揭示了证候具有结构性和临界性的特征,为辨证论治疗效评价方法学提供了部分的理论依据和技术支撑。
本研究首次提出对证候进行解构,指出证候“内实”的变化是证候疗效评价的主要指标,并采用潜在类别模型探索证候的“内实外虚”,利用相似度分析和非线性混合效应模型探索药物干预下证候的动态变化趋势及证候状态变化的拐点。课题采用相似度分析和NONMEM分析证候的动态变化趋势,是符合证候特征的研究方法,但由于证候的复杂性,我们的研究还需深入,开展进一步的研究。证候“外虚”的变化与证候疗效评价的关系,并未探讨,是本课题下一步主要研究的方向。
Syndrome has the characteristics of "internal stability and external flexibility". The "internal stability" means the key syndrome that reflects pathogenesis and the common character in the group, which is the basis of intervention. The "external flexibility" means the individual syndrome, which can affects the principle and method of intervention. Therefore, if the syndrome target is different, the significance to evaluate efficacy is different. Nimodipine is the common method to evaluate the syndrome efficacy, which doesn't conform to the characteristics of "internal stability and external flexibility". So the study proposed that deconstruction to syndrome and analysis to "internal stability" dynamic changes is the key part to evaluate syndrome efficacy.
Syndrome is the dynamic process which varies with time, and so are the"internal stability" and "external flexibility" parts of Syndrome. The recognization of the dynamic change trend of syndrome especially "internal stability" and "external flexibility" part of Syndrome is the important foundation for the application into the clinical practice and enhancement of the treatment effect of TCM, and the critical process to evaluate the efficacy of the syndrome. Based on the theory of "internal stability and external flexibility" proposed by Pro. Wang Yongyan, a randomized, double-blind, multi-center trial of Xinxuekang Capsule for blood stasis syndrome of angina pectoris of coronary heart disease was conducted to find an approach to evaluating the effect in TCM syndrome.
Object:To discuss the effect of the drug for Syndrome and the "internal stability" part of Syndrome under the change of time by clinical annotation to "internal stability and external flexibility" of Syndrome and its dynamic change, which would provide new method to evaluate the efficacy of the syndrome.
Method:Based on the randomized, blind, multi-center clinical trial data of Xinxuekang capsule for the treatment of blood stasis syndrome in angina pectoris of coronary heart disease (CHD-AP-BSS),the study determined the latent class and the"internal stability" part of CHD-AP-BSS according to the latent class modeling (LCM). According to the class of the data, the study analyzed the change of latent class and single symptom of CHD-AP-BSS by Generalized Estimating Equations (GEE).And then the study discussed the dynamic change trends of syndrome and "internal stability" part of syndrome by similarity analysis under the different interventions based on the dynamic changes of dimensional data such as syndrome and "internal stability" part of syndrome. At last, the study applied NONMEM into the analysis of the changes in similarities with time.
Results:
1. According to the comparisons of the baseline in A and B2 groups, there were not any significant differences in CHD-AP, BSS and other test results between the two groups.
2. By t-test analysis, the study found that the changes of the syndrome score between 4w,6w,8w or 3m after treating and baseline were significant, respectively. The changes of the syndrome score between A and B group at the 3 m after treating were significant, respectively.
By single factor analysis of variance of repetitive measurement, the study found that syndrome score decreased by time. The changes of the syndrome score between 2w,4w,6w,8w or 3m after treating and baseline were significant, respectively. The changes of the syndrome score between A and B group at the 4w,6w,8w or 3m after treating were significant, respectively.
The results of t-test analysis and single factor analysis of variance of repetitive measurement is different. The latter is the common method to analyze the repetitive measurement data, which the results is dependable.
3. Based on the results of LCM to CHD-AP-BSS, the study found that the patients can be classified to "the mild" and "the serious". The symptoms like chest pain, chest distress, dark purple lip, dark tongue were relatively more stable than other symptoms in CHD-AP-BSS, which is the "internal stability" part of CHD-AP-BSS.
4. Based on the results of similarity analyses to CHD-AP-BSS at different times and in different groups by Cosine method, Dice method and Jaccard method, the study found that the total change trend of the similarity values was similar by different methods, although the details of similarity values were different in different methods, which shows that the similarity values were decreased gradually in both A and B groups with the time of treatment. It suggested that both A and B group can improve the blood stasis syndrome.
Analyzing syndrome changing trend with time by nonlinear mixed effected model, the research found that similarity of syndrome changed with time in both A and B groups. The change of the similarity was influenced by angina pectoris scores, "internal stability" part of syndrome, et al. At the same time, the factor "group" was related to the change of syndrome, which suggested that the change of syndrome between A group and B group was different significantly.
Take the similarity values equal to 0.8 as the boundary point, the study found that at that time, A is used 7.28w and B is used 10.72w, which is considered to the change of CHD-AP-BSS.
5. By LCM and GEE, the research found that the latent class in 2w,4w,6w,8w and 3m was significantly different compared with the baseline, respectively. At the 4w, the factor "time" and "group" existed interaction. The results illustrated that CHD-AP-BSS alleviates gradually. The change of latent class in 4w after treatment between A group and B group is different significantly.
6. Based on the results of the similarity values of "internal stability" part of CHD-AP-BSS at different time and in different groups by Cosine method, Dice method and Jaccard method, the research found that although the similarity values were different by different methods, its total change trend was similar in each method. It suggested that the similarity values were decreased gradually in both A and B groups with the time of treatment, which illustrated that both A and B group can improve the "internal stability" part of CHD-AP-BSS.
According to the analysis of the correlation between the change of "internal stability" part of CHD-AP-BSS and time by NONMEM, the results showed that similarity of "internal stability" part of syndrome changed with time. The factor "test centre" was an influence factor for the similarity change of "internal stability" part in both A and B group. The factor "angina pectoris scores" affected the similarity change of "internal stability" part of CHD-AP-BSS only in B group. The difference of similarity change of "internal stability" part of CHD-AP-BSS between group A and group B was significant.
7. According to the analysis of the single symptoms of CHD-AP-BSS by rank test, the study found that the changes of symptom scores of chest pain, chest distress, palpitation, dark purple lip, tongue and pulse manifestations in 2w,4w,6w,8w and 3m were significantly different significantly compared with the baseline except pulse manifestations in 2w,4w in B group. The differences of chest pain in 20w, palpitation in 2w,4w,8w and 20w, tongue picture from 4w to 20w between group A and group B were significant.
According to the analysis of the single symptoms of CHD-AP-BSS by GEE, the research found that the changes of symptom scores of chest pain, chest distress, palpitation, dark purple lip, tongue and pulse manifestations in 2w,4w,6w,8w and 3m were significantly different significantly compared with the baseline, and the differences between the symptoms scores at the two adjacent time points were also different significantly. It indicated that both drug A and B could improve the symptoms of CHD-AP-BSS and the longer the patients were treated, the better the efficacy would be got. The differences of chest pain in 20w, palpitation from 2w to 20w, tongue picture from 4w to 20w between group A and group B were significant.
The results of rank test analysis and single factor analysis of variance of repetitive measurement is different. The latter is the common method to analyze the repetitive measurement data, which the results is dependable.
Conclusion
1. Symptoms contributing to the syndrome, the latent class and the "internal stability" part of the syndrome and the syndrome itself are in constant change with space and time. The effect in TCM syndrome can be evaluated by analyzing the change of contributing symptoms, the "internal stability" part of the syndrome and the syndrome itself。
2. LCM can analyze the contribution of symptom to syndrome, which can find out the "internal stability" part of the syndrome. LCM is one of the effective methods to annotate, the "internal stability" part of the syndrome.
3. The data type should be considered carefully during analysis. When the data belongs to repeated measurement data, the relationship between data in different observation point must be considered. The analysis method should be fit for repeated measurement data.
4. The integration method of similarity analysis and NONMEM can better display the high dimensionality and nonlinear character of syndrome and find out the inflection point of syndrome change.
5. The study suggested that syndrome has the character of constitutive property and criticality, which provides part theoretical foundation and technical support for efficacy evaluation of treatment based on syndrome differentiation.
The study proposed that the analysis to the dynamic change of "internal stability and external flexibility" of syndrome was the entry point to evaluate the effect in TCM syndrome for the first time. Determine the "internal stability and external flexibility" of syndrome by LCM. Determine the dynamic change trend and the knee point of the change in syndrome by similarity analysis and NONMEM. The approach of analysis to dynamic change trend by similarity analysis and NONMEM is suitable for syndrome study, which is one of the entry points to evaluate the effect in TCM syndrome. But due to the complicated property of the syndrome, the further study is needed in the future.2. The study suggested that the change of "internal stability" part of syndrome was the main indicator to evaluate the effect in TCM syndrome. However, the relationship between the change in "external flexibility" part of syndrome and the effect of syndrome isn't discussed in this study, which is the direction to further study.
引文
1郭蕾.证候概念渊源及现代对证候的研究简析[J].中医药学刊,2006,21(6):947-948
2封静,张志斌.现代证候概念研究发展趋势[J].山西中医学院学报,2006,7(1)：54-56
3郭蕾,王永炎,张志斌等.从辩证逻辑角度探寻证候概念的形成轨[J].中医杂志,2007,48(2)：101-103
4郭蕾,王永炎,张志斌等.证候概念发展轨迹探源[J].中西医结合学报,2006,4(4)：335-338
5匡调元.中医病理研究进展[J].重庆医药,1979.3:4-14,23.
6杨维益,王天芳,陈家旭等.关于中医证的概念及其定义的思考[J].中医杂志,1996,37(6)：370
7印会河.中医基础理论[M].上海：上海科学技术出版社,1984
8邓铁涛.邓铁涛医学文集[M].北京：人民卫生出版社,2001：1740.
9张志斌,王永炎,吕爱平等.论证候要素与证候靶点应证组合辨证[J].中医杂志,2006,27(7)：483-485
10王永炎,张志斌.再议完善辨证方法体系的几个问题[J].天津中医药,2007,24(1)：1-4
11张志斌,王永炎.辨证方法新体系的建立[J].北京中医药大学学报,2005,28(1)：1-3
12 Qu K, Wang TF, Zhao Y, et al. Common syndrome factors of traditional Chinese medicine in chronic renal failure based on the questionnaire investigation among experts[J]. Zhong Xi Yi Jie He Xue Bao.2010 Dec;8(12):1147-52.
13王嘉麟,郭蓉娟,张允岭等.基于因子分析的高血压病证候要素研究[J].天津中医药,2010,27(5)：434-436
14谢颖桢,马洪明,程平荣等.急性缺血性中风病不同OCSP分型证候要素的特征[J].中国中医基础医学杂志.2008,14(1)：44-45
15靳冬.证候要素的提出及应用近况[J].光明中医,2010,25(2):339-340
16 Chen MS, et al. Data mining:an overview from database perspective[J]. IEEE Transactions on Knowledge and data Engineering, 1996,8(6):866-883.
17李瑞娟,杨文明.中医证候研究方法学探讨[J].中医药临床杂志.2010,22(2):109-112
18 Han Jia-wei, Kamber M. Data mining:Concepts and techniques[J]. Beijing:Higher Education Press,2001.
19黄粤,高颖,马斌.中医证候研究常用数据挖掘方法述评[J].中医药学报,2010,38(3)：6-10
20艾军,汪受传,杨宏宝等.583例小儿肺炎证候病机学关联规则介析[J].世界科学技术一中医药现代化,2009,11(60)：810-814
21陈明,杨慧芳,余蕾.基于关联规则的肝硬变辨证数据挖掘研究[J].河南中医,2009,29(3)：258-260
22汪受传,艾军,杨燕等.基于关联规则的小儿肺炎热郁痰瘀相关病机分析[J].南京中医药大学学报,2010,26(2)：97-101
23贾运滨,魏江磊.数据挖掘技术在中医证候研究中的应用述评[J].中国中医急症,2010,19(7)：1184-1186
24沈亚诚,王小云,蔡永铭.数据挖掘技术在中医证治研究中的应用与展望[J].中国中西医结合杂志.2008,28(9)：847-850
25李斌,徐蓉,李福伦等.基于联系数的痛风性关节炎血瘀证辨证因子研究[J].中西医结合学报,2009,7(8):724-728
26孟庆刚,郭书文,王永炎等.如何将数学方法应用于现代中医药学研究[J].中国中医药信息杂志,2000,7(8):21.
27吴焕林,阮新民,罗文杰等.319例冠心病患者证型分布聚类分析及证型诊断条件的确立[J].中国中西医结合杂志,2007,27(7):616-618
28张纾难,疏欣杨,韩春生等.131例特发性肺纤维化患者中医证候聚类分析[J].环球中医药,2011,4(1)：20-22
29高永刚,马惠荣,尤昭玲等.功能失调性子宫出血1016例症状聚类分析[J].河北中医药学报,2009,24(3)：9-11
30龚燕冰,王永炎,倪青.中医证候研究的现代方法学综述(一)——中医证候数据挖掘技术[J].北京中医药大学学报,2006,29(12)：797-801
31 Pawlak Z.Rough Set theoretical Aspects of Reasoning aboutData[J]. Dordrecht:Kluwer Academic Publishers,1991.
32 Slowinski R. Intelligent Decision Support-handbook of Applications and Advances of the Rough Sets Theory[J].Dordrecht:Kluwer Academic Publishers,1992.
33秦中广,毛宗源,邓兆智.粗糙集在中医类风湿证候诊断中的应用[J].中国生物医学工程学报,2001,20(4)：357-363
34杨小波,张钢,梁兆晖.基于粗糙集理论的溃疡性结肠炎中医主症分析[J].辽宁中医杂志.2008,35(5)：687-689
35 Jensen FV. An Introduction to Bayesian Networks [J]. New York:Springer, 1996.
36 Heckermand, Mamdania, Wellmanm P. Realworld application of Bayesian network [J]. Communication of ACM,1995,38:24-26.
37张新峰,沈兰荪,刘垚巍等.基于粗糙集与SVM概率输出的中医舌象特征融合方法研究[J].世界科学技术—中医药现代化,2007,9(5):122-128
38吴昊,段禅伦,熊志伟等.粗糙集理论在中医诊断学中的应用研究[J].内蒙古大学学报(自然科学版),2006,37(3)：351-355
39 Jensen FV. An Introduction to Bayesian Networks [J]. New York:Springer, 1996.
40李志更,王天芳,任婕等.中医科研中几种常用数据挖掘方法浅析[J].中医药学报,2008,36(2)：29-32.
41唐启盛,曲淼,包祖晓等.抑郁症中医证候的贝叶斯网络研究[J].中医杂志,2008,49(11)：1013-1015
42龚燕冰,倪青,高思华等.2型糖尿病不同并发症中医证候与血糖相关性的贝叶斯网络分析[J].北京中医药大学学报,2009,32(12)：815-818
43许朝霞,王忆勤,颜建军.贝叶斯网络在中医辨证研究中的应用[J].上海中医药大学学报,2009,23(2)：77-79
44陈淑慧,梁伟雄.现代方法学在中医证候诊断规律研究中的应用现状[J].中国中医基础医学杂志,2009,15(4):312-313,320
45朱文锋,晏峻峰,黄碧群.贝叶斯网络在中医证素辨证体系中的应用[J].中西医结合学报,2006,4(6)：567-571
46周利锋,原少勤,高尔生.人工神经网络的概念[J].医学信息.1998,11(11)：8-10
47孙贵香,郜文辉,袁肇凯等.构建基于人工神经网络的证候研究方法学[J].中华中医药学刊,2009,27(11)：2316-2317
48白云静,申洪波,孟庆刚.基于共轭梯度下降算法的类风湿关节炎BP神经网络证候模型研究[J].中国中医药信息杂志,2010,17(3):96-97
49白云静,申洪波,孟庆刚,王永炎.基于人工神经网络的中医证候非线性建模研究[J].中国中医药信息杂志,2007,14(7)：3-4,83
50孙贵香,袁肇凯.人工神经网络在中医证候研究中的应用[J].中华中医药学刊,2007,25(7)：1450-1452
51孙振球,徐勇勇.医学统计学[M].北京：人民卫生出版社,2002:26.
52高怀林,吴以岭,贾振华等.中医证候量化诊断研究概况[J].江苏中医药.2007,39(9)：78-80
53张晓刚,吕志平,吴国祥.隐证型慢性乙型肝炎证候相关因素的logistic回归分析[J].中华中医药学刊,2009,27(8):1761-1762
54周德生,马成瑞.223例多发性硬化患者中医辨证分型研究[J].中国中医药信息杂志,2009,16(12)：21-23
55杜彩凤,王天芳,辛意等.绝经后期更年期综合征常见证候要素的Logistic回归分析[J].北京中医药大学学报,2009,32(4):238-243
56王芳.主成分分析与因子分析的异同比较及应用[J].统计教育,2003(5)：14
57牟新,周旦阳,赵进喜.中医证候学研究常用统计方法学[J].中国中医药科技,2008,15(1)：56-57
58孔林,霍清萍,李金菩等.137例缺血性脑血管病患者中医证候研究[J].浙江中医药大学学报,2010,34(6):852-856
59聂莉芳,韩东彦,余仁欢.363例IgA肾病气阴两虚证类证候的分布研究[J].??中国中西医结合肾病杂志,2008,9(5)：426-42960王雪花,夏春明,颜建军等.中医证候分类中常用多元统计分析方法及应用评析[J].世界科学技术——中医药现代化
思路与方法,2008,10(2):15-2061毛竹君,董颖,栾洁.180例2型糖尿病患者中医证候调查及证型研究[J].中华中医药杂志,2009,24(8)：1064-106762李丛煌,花宝金.283例首治中晚期非小细胞肺癌患者证候要素研究[J].辽宁中医杂志,2009,36(11)：1844-184563王阶,何庆勇,李海霞等.815例不稳定型心绞痛中医证候的因子分析[J].中西医结合学报,2008,6(8)：788-79264 Anderson JC, Gerbing DW. Structural equation modeling in practice:A review and recommended two-step approach[J]. Psychological Bulletin, 1988,103:411-42365郦永平,陈启光,张菊萍等.脑梗死中医“证”的结构方程[J].四川中医,2006,24(8)：27-3066陈启光,申春悌,张华强等.结构方程模型在中医证候规范标准研究中的应用[J].中国卫生统计,2005,22(1)：2-467 ZHANG N L. Hierarchical latentclassmodels for cluster analysis[J]. Journalof Machine Learning Research,2004,5(6):697-723.68张连文,袁世宏,陈弢等.隐结构模型与中医辨证研究(Ⅱ)——肾虚数据分析[J].北京中医药大学学报,2008,31(9)：584-58769王阶,李海霞,王燚等.基于隐结构模型的血赛证症状体征研究[J].世界科学技术一中医药现代化,2006,8(6):23-26,6270袁世宏,张连文,陈弢等.隐结构模型与中医辨证研究(Ⅲ)——模型辨证与专家辨证[J].北京中医药大学学报,2008,31(10)：659-66371简维雄,袁肇凯.中医证候的组学研究进展[J].上海中医药大学学报,2008,22(2)：69-7272郭芳宏,魏嵋.运用蛋白质组学技术研究中医证候本质进展[J].现代中西医结合杂志,2010,19(3)：1678-167973 Ma T, Tan C, Zhang H, et al. Bridging the gap between traditional Chinese medicine and systems biology:the connection of Cold Syndrome and NEI network. [J]. Mol Biosyst.2010 Apr;6(4):613-9.
74 Ding WJ, Zeng YZ, Li WH, et al. Identification of Linkage Disequilibrium SNPs from a Kidney-Yang Deficiency Syndrome Pedigree[J]. Am J Chin Med.2009;37(3):427-38.
75马民.闭塞性动脉硬化症血瘀证患者bcl-2和HsP70基因的表达[J].云南大学学报(自然科学版),2004,26(6A):1-5
76张明,赵英日,李强.从基因角度研究中医证候实质的思路[J].江西中医药,2009,39(9)：68-69
77 Lu CL, Qv XY, Jiang JG. Proteomics and syndrome of Chinese medicine [J]. J Cell Mol Med.2010 Dec;14(12):2721-8
78胡元会,褚瑜光,吴华琴.高血压病中医不同证候患者血清蛋白质组学研究[J].中医药中青年科技创新与成果展示论坛论文集,2009
79郎笑梅,徐珊.痰瘀证的血清蛋白质指纹图谱的临床研究[J].浙江中医药大学学报.2010,34(3)：314-315
80曾年菊,梁清华,陈疆等.蛋白质组学方法研究TPx在肝风内动证中的表达及关系[J].中华中医药学刊,2008,26(4):750-752
81郭芳宏,魏嵋.运用蛋白质组学技术研究中医证候本质进展[J].现代中西医结合杂志,2010,19(13)：1678-1679
82杨元宵,李昌煜.运用蛋白质组学研究中医证候的思考[J].中国实验方剂学杂志,2007,13(1)：69-72
83杜武勋,朱明丹,姜民等.代谢组学与中医证候研究[J].中国循证心血管医学杂志,2010,2(3)：142-145
84董飞侠,黄迪,何立群等.Ⅲ期慢性肾病肾阳虚证患者尿液代谢组学特征的研究[J].中华中医药杂志,2008,23(12):1109-1113
85郑海生,蒋健,贾伟等.慢性心力衰竭肾阳虚证患者代谢组学研究[J].中华中医药杂志,2009,S1：16-18
86任荣政,周永明,朱文伟.骨髓增生异常综合征中医证型的代谢组学分析[J].上海中医药杂志,2009,43(8):28-31
87罗和古,陈家旭.代谢组学技术与中医证候的研究[J].中国中医药信息杂志,2007,14(5)：3-5
88简维雄.中医“证候”与代谢组学研究[J].中华中医药学刊,2009,27(2):351-352
89刘凤斌,方积乾,王建华.中医药临床疗效评价的探讨[J].中药新药与临床药理,2004,4(7)：290-292.
90梁茂新,王雪峰.中医疗效评价评价指标和方法研究需要解决的认识问题[J].世界科学技术——中药现代化,2006,8(1)：31-35.
91张伯礼,康立源,项耀祖.中药新药临床试验中有关疗效评价若干问题的思考[J].中国新药与临床杂志,2007,26(11)：861-863
92申春悌,陈启光,张华强,等.建立病证结合疗效评价体系的思路和方法[J].中国中医基础医学杂志,2006,12(10):749-751
93熊宁宁,罗玫,蒋萌等.中药临床试验的适应证候设计与疗效评价[J].中国临床药理学与治疗学,2003,8(6)：715-717.
94郭新峰,赖世隆,梁伟雄.中医药临床疗效评价中结局指标的选择与应用[J].广州中医药大学学报,2002,19(4):251-255.
95王文萍,李晓斌,李晓等.临床痊愈比例超高时中医症状赋分方法对证候疗效评价的影响[J].世界科学技术—中医药现代化,2008,10(5)：33-37
96王文萍,李晓斌,李晓.探讨三种不同中医症状赋分方法对中医证候疗效评价的影响[J].世界科学技术—中医药现代化,2008,10(1)：35-38
97刘建平.循证医学方法与中医疗效评价[J].首都医科大学学报,2007,28(2):212-215
98江锋.关于中医肿瘤临床疗效评价的思考[J].中华中医药学刊,2008,26(3)：525-527
99梁晓春.循证医学与中医临床疗效评价[J].中华全科医师杂志,2004,3(1):7-9
100倪量,李峰.循证医学在中医药临床疗效评价中的应用[J].山西中医,2008,24(10)：57-58
101商洪才,王保和,张伯礼.中药新药证候及疗效评价[J].中药新药与临床药 理,2004,15(5)：365-368
102胡学军,商洪才,张伯礼等.生存质量及其量表在中医药疗效评价中的应用[J].天津中医药,2004,21(3):191-193
103汪受传,赵霞,任现志等.基于证候动态变化的病毒性肺炎疗效评价方法研究[J].世界科学技术-中医药现代化.2008,10(5)：10-15
104汪受传,赵霞,任现志等.基于主症动态变化的病毒性肺炎疗效评价方法研究[J].中华中医药杂志,2008,23(8)：675-679
105王雪峰,董丹,梁茂新等.数据挖掘技术在小儿肺炎中医临床疗效评价研究中应用的思路与方法[J].中国中西医结合杂志,2007,27(10):949-951
106沈亚诚,王小云,傅昊阳等.基于数据挖掘的绝经综合征中医药临床疗效评价研究[J].中国卫生统计,2008,25(4):387-390
107李东涛,凌昌全,郎庆波等.以证候为内容的原发性肝癌中医疗效评价体系研究[J].中西医结合学报,2007,5(1)：15-22
108郭蕾,王永炎,张俊龙.论证候的内实外虚[J].中国医药学报,2004,19(11):645-648
109申春悌.结构方程模型在异病同证研究中的运用[J].中国中医基础医学杂志,2008,14(12).953-956
110王蕾,王永炎,张志斌等.证候动态时空特征的复杂性及相应的研究思路[J].中医研究,2006,19(3):1-3
111马斌,高颖.中风病发病第7天和第14天证候要素演变规律初步研究[J].辽宁中医杂志,2006,33(12)：1561-1563.
112 Shao Yuan Chuang, Chen Huan Chen, Pesus Chou. Hypertension and the Insulin-related Metabolic Syndrome:Factor Analysis in Taiwanese. Acta Cardiol Sin,2004;20:229-36
113 Liang Han CHang. A Study on Bank Employees'Job Stress, Leisure Coping, and Physical and Mental Health. Asian Journal of Management and Humanity Sciences,2006, (1):161-174,
114潘国清.受限语境下文本语义相似度计算研究及应用[J].东南大学硕士学位论文,2007,23
115 Tian Runtao, XIE Peishan. Study on the standardization of similarity evaluation method of chromatographic fingerprints(Part D[J]. Traditional Chinese Drug Research & Clinical Pharmacology,2006, 17(1):40-42.
116王巧,许磊,冯超.金莲花药材高效液相色谱法指纹图谱分析.时珍国医国药,2010,21(2)：331-332.
117林文津,张亚敏,徐榕青,等.尿微量蛋白毛细管电泳指纹图谱的初步分析.福建医药杂志,2009,31(5)：73-75
118马欣,孙毓庆.银杏叶提取物的多维指纹图谱研究.色谱,2003,21(6)：562-567
119曹洪欣.浅谈证候动态变化的形式[J].中医药学报,1987,(3):6-7
120邱皓政.潜在类别模型的原理与技术[M].教育科学出版社,2008
121 Rost J. A latent class Model for rating data. Psychometrika,1985, 50:37-49
122李翠娟,烟建华,巩振东.对中医证本质研究的反思[J].中国中医基础医学杂志,2005,11(4)：271—272.
123张连文,袁世宏.隐结构模型与中医辨证研究——隐结构法的基本思想及隐结构分析工具[J].北京中医药大学学报,2006,29(6):365-369
124姚奎武,王阶,朱翠玲等.不同计量诊断对于冠心病血瘀证症状体征分析比较[J].世界科学技术—中医药现代化,2009,11(5)：684-688
125李稍.从维度与阶度探讨中医证候的特征及标准化方法[J].北京中医药大学学报,2003,26(3)：1-4
126郭蕾,王学伟,王永炎.论高维高阶与证候的复杂性[J].中华中医药杂志.2006,21(2)：76-78
127苏越,王磊,张华,等.肝炎肝硬化中医证候与疾病信息相似度分析[J].中国中西医结合杂志,2009,29(5)：398—402.
128操牡丹,何前锋,王柏.中医药方剂相似度模型[J].计算机工程,2009,35(16)：275—277.
129孙继佳,陆奕宇,苏式兵.基于相似匹配测度的中医证候疗效评价方法研究 [J].中国中医药信息杂志.2010,17(11):6-9
130谢平,肖婵,雷红富等.基于向量相似度原理的湖泊富营养化评价方法及其验证[J].安全与环境学报.2008,8(4):93-94
131刘素萍,仁立学,胡广春等.夹角余弦法用于辐射源一致性判定的评估[J].第十四届全国核电子学与核探测技术学术年会论文集,2008
132陶跃华.基于向量的相似度计算方案[J].云南师范大学学报.2001,21(5)：17-19
133 Liang KY, Zeger SL. Longitudinal data analysis using generalized linear models. Biometrics,1986,73:13-22
134 Lu JC, Chen D, Zhou WX. Quasi-likelihood estimation for GLM with random scales. Journal of Statistical Planningand Inference,2006,136(2): 401-429
135 Lipsitz SR, Kim K, Zhao L. Analysis of repeated categorical data using generalized estimating equations Statistics in Medicine,1994,13(1 1):1149-1163
136 Bishop J, Venables WN, Wang YG. Analysing commercial catch and effort data from a Penaeid trawl fishery:A comparison of linear models, mixed models, and generalised estimating equations approaches. Fisheries Research,2004,70(2-3):179-193
137 Pendergast JF, Gange SJ, Newton MA, el al. A Survey of Methods for Analyzing Clustered Binary Response Data. International Statistical Review,1996,64(1):89-118
138 Carr GJ, Chi EM. Analysis of variance for repeated measures Data: a generalized estimating equations approach. Statistics in Medicine, 1992,11(8):1033-1140
139黄子辉.地奥心血康胶囊治疗冠心病心绞痛的临床分析[J].医药论坛,2009,(1)：24-25
140赵晋芳,罗天娥,刘桂芬.分类重复测量数据的非线性混合效应模型[J].现代预防医学,2007,34(20)：3866-3868
141罗天娥,刘桂芬.重复测量资料非线性混合效应模型应用与实现[J].中国卫生统计,2006,23(2)：104-107
2封静,张志斌.现代证候概念研究发展趋势[J].山西中医学院学报,2006,7(1)：54-56
3郭蕾,王永炎,张志斌等.从辩证逻辑角度探寻证候概念的形成轨[J].中医杂志,2007,48(2)：101-103
4郭蕾,王永炎,张志斌等.证候概念发展轨迹探源[J].中西医结合学报,2006,4(4)：335-338
5匡调元.中医病理研究进展[J].重庆医药,1979.3:4-14,23.
6杨维益,王天芳,陈家旭等.关于中医证的概念及其定义的思考[J].中医杂志,1996,37(6)：370
7印会河.中医基础理论[M].上海：上海科学技术出版社,1984
8邓铁涛.邓铁涛医学文集[M].北京：人民卫生出版社,2001：1740.
9张志斌,王永炎,吕爱平等.论证候要素与证候靶点应证组合辨证[J].中医杂志,2006,27(7)：483-485
10王永炎,张志斌.再议完善辨证方法体系的几个问题[J].天津中医药,2007,24(1)：1-4
11张志斌,王永炎.辨证方法新体系的建立[J].北京中医药大学学报,2005,28(1)：1-3
12 Qu K, Wang TF, Zhao Y, et al. Common syndrome factors of traditional Chinese medicine in chronic renal failure based on the questionnaire investigation among experts[J]. Zhong Xi Yi Jie He Xue Bao.2010 Dec;8(12):1147-52.
13王嘉麟,郭蓉娟,张允岭等.基于因子分析的高血压病证候要素研究[J].天津中医药,2010,27(5)：434-436
14谢颖桢,马洪明,程平荣等.急性缺血性中风病不同OCSP分型证候要素的特征[J].中国中医基础医学杂志.2008,14(1)：44-45
15靳冬.证候要素的提出及应用近况[J].光明中医,2010,25(2):339-340
16 Chen MS, et al. Data mining:an overview from database perspective[J]. IEEE Transactions on Knowledge and data Engineering, 1996,8(6):866-883.
17李瑞娟,杨文明.中医证候研究方法学探讨[J].中医药临床杂志.2010,22(2):109-112
18 Han Jia-wei, Kamber M. Data mining:Concepts and techniques[J]. Beijing:Higher Education Press,2001.
19黄粤,高颖,马斌.中医证候研究常用数据挖掘方法述评[J].中医药学报,2010,38(3)：6-10
20艾军,汪受传,杨宏宝等.583例小儿肺炎证候病机学关联规则介析[J].世界科学技术一中医药现代化,2009,11(60)：810-814
21陈明,杨慧芳,余蕾.基于关联规则的肝硬变辨证数据挖掘研究[J].河南中医,2009,29(3)：258-260
22汪受传,艾军,杨燕等.基于关联规则的小儿肺炎热郁痰瘀相关病机分析[J].南京中医药大学学报,2010,26(2)：97-101
23贾运滨,魏江磊.数据挖掘技术在中医证候研究中的应用述评[J].中国中医急症,2010,19(7)：1184-1186
24沈亚诚,王小云,蔡永铭.数据挖掘技术在中医证治研究中的应用与展望[J].中国中西医结合杂志.2008,28(9)：847-850
25李斌,徐蓉,李福伦等.基于联系数的痛风性关节炎血瘀证辨证因子研究[J].中西医结合学报,2009,7(8):724-728
26孟庆刚,郭书文,王永炎等.如何将数学方法应用于现代中医药学研究[J].中国中医药信息杂志,2000,7(8):21.
27吴焕林,阮新民,罗文杰等.319例冠心病患者证型分布聚类分析及证型诊断条件的确立[J].中国中西医结合杂志,2007,27(7):616-618
28张纾难,疏欣杨,韩春生等.131例特发性肺纤维化患者中医证候聚类分析[J].环球中医药,2011,4(1)：20-22
29高永刚,马惠荣,尤昭玲等.功能失调性子宫出血1016例症状聚类分析[J].河北中医药学报,2009,24(3)：9-11
30龚燕冰,王永炎,倪青.中医证候研究的现代方法学综述(一)——中医证候数据挖掘技术[J].北京中医药大学学报,2006,29(12)：797-801
31 Pawlak Z.Rough Set theoretical Aspects of Reasoning aboutData[J]. Dordrecht:Kluwer Academic Publishers,1991.
32 Slowinski R. Intelligent Decision Support-handbook of Applications and Advances of the Rough Sets Theory[J].Dordrecht:Kluwer Academic Publishers,1992.
33秦中广,毛宗源,邓兆智.粗糙集在中医类风湿证候诊断中的应用[J].中国生物医学工程学报,2001,20(4)：357-363
34杨小波,张钢,梁兆晖.基于粗糙集理论的溃疡性结肠炎中医主症分析[J].辽宁中医杂志.2008,35(5)：687-689
35 Jensen FV. An Introduction to Bayesian Networks [J]. New York:Springer, 1996.
36 Heckermand, Mamdania, Wellmanm P. Realworld application of Bayesian network [J]. Communication of ACM,1995,38:24-26.
37张新峰,沈兰荪,刘垚巍等.基于粗糙集与SVM概率输出的中医舌象特征融合方法研究[J].世界科学技术—中医药现代化,2007,9(5):122-128
38吴昊,段禅伦,熊志伟等.粗糙集理论在中医诊断学中的应用研究[J].内蒙古大学学报(自然科学版),2006,37(3)：351-355
39 Jensen FV. An Introduction to Bayesian Networks [J]. New York:Springer, 1996.
40李志更,王天芳,任婕等.中医科研中几种常用数据挖掘方法浅析[J].中医药学报,2008,36(2)：29-32.
41唐启盛,曲淼,包祖晓等.抑郁症中医证候的贝叶斯网络研究[J].中医杂志,2008,49(11)：1013-1015
42龚燕冰,倪青,高思华等.2型糖尿病不同并发症中医证候与血糖相关性的贝叶斯网络分析[J].北京中医药大学学报,2009,32(12)：815-818
43许朝霞,王忆勤,颜建军.贝叶斯网络在中医辨证研究中的应用[J].上海中医药大学学报,2009,23(2)：77-79
44陈淑慧,梁伟雄.现代方法学在中医证候诊断规律研究中的应用现状[J].中国中医基础医学杂志,2009,15(4):312-313,320
45朱文锋,晏峻峰,黄碧群.贝叶斯网络在中医证素辨证体系中的应用[J].中西医结合学报,2006,4(6)：567-571
46周利锋,原少勤,高尔生.人工神经网络的概念[J].医学信息.1998,11(11)：8-10
47孙贵香,郜文辉,袁肇凯等.构建基于人工神经网络的证候研究方法学[J].中华中医药学刊,2009,27(11)：2316-2317
48白云静,申洪波,孟庆刚.基于共轭梯度下降算法的类风湿关节炎BP神经网络证候模型研究[J].中国中医药信息杂志,2010,17(3):96-97
49白云静,申洪波,孟庆刚,王永炎.基于人工神经网络的中医证候非线性建模研究[J].中国中医药信息杂志,2007,14(7)：3-4,83
50孙贵香,袁肇凯.人工神经网络在中医证候研究中的应用[J].中华中医药学刊,2007,25(7)：1450-1452
51孙振球,徐勇勇.医学统计学[M].北京：人民卫生出版社,2002:26.
52高怀林,吴以岭,贾振华等.中医证候量化诊断研究概况[J].江苏中医药.2007,39(9)：78-80
53张晓刚,吕志平,吴国祥.隐证型慢性乙型肝炎证候相关因素的logistic回归分析[J].中华中医药学刊,2009,27(8):1761-1762
54周德生,马成瑞.223例多发性硬化患者中医辨证分型研究[J].中国中医药信息杂志,2009,16(12)：21-23
55杜彩凤,王天芳,辛意等.绝经后期更年期综合征常见证候要素的Logistic回归分析[J].北京中医药大学学报,2009,32(4):238-243
56王芳.主成分分析与因子分析的异同比较及应用[J].统计教育,2003(5)：14
57牟新,周旦阳,赵进喜.中医证候学研究常用统计方法学[J].中国中医药科技,2008,15(1)：56-57
58孔林,霍清萍,李金菩等.137例缺血性脑血管病患者中医证候研究[J].浙江中医药大学学报,2010,34(6):852-856
59聂莉芳,韩东彦,余仁欢.363例IgA肾病气阴两虚证类证候的分布研究[J].??中国中西医结合肾病杂志,2008,9(5)：426-42960王雪花,夏春明,颜建军等.中医证候分类中常用多元统计分析方法及应用评析[J].世界科学技术——中医药现代化
思路与方法,2008,10(2):15-2061毛竹君,董颖,栾洁.180例2型糖尿病患者中医证候调查及证型研究[J].中华中医药杂志,2009,24(8)：1064-106762李丛煌,花宝金.283例首治中晚期非小细胞肺癌患者证候要素研究[J].辽宁中医杂志,2009,36(11)：1844-184563王阶,何庆勇,李海霞等.815例不稳定型心绞痛中医证候的因子分析[J].中西医结合学报,2008,6(8)：788-79264 Anderson JC, Gerbing DW. Structural equation modeling in practice:A review and recommended two-step approach[J]. Psychological Bulletin, 1988,103:411-42365郦永平,陈启光,张菊萍等.脑梗死中医“证”的结构方程[J].四川中医,2006,24(8)：27-3066陈启光,申春悌,张华强等.结构方程模型在中医证候规范标准研究中的应用[J].中国卫生统计,2005,22(1)：2-467 ZHANG N L. Hierarchical latentclassmodels for cluster analysis[J]. Journalof Machine Learning Research,2004,5(6):697-723.68张连文,袁世宏,陈弢等.隐结构模型与中医辨证研究(Ⅱ)——肾虚数据分析[J].北京中医药大学学报,2008,31(9)：584-58769王阶,李海霞,王燚等.基于隐结构模型的血赛证症状体征研究[J].世界科学技术一中医药现代化,2006,8(6):23-26,6270袁世宏,张连文,陈弢等.隐结构模型与中医辨证研究(Ⅲ)——模型辨证与专家辨证[J].北京中医药大学学报,2008,31(10)：659-66371简维雄,袁肇凯.中医证候的组学研究进展[J].上海中医药大学学报,2008,22(2)：69-7272郭芳宏,魏嵋.运用蛋白质组学技术研究中医证候本质进展[J].现代中西医结合杂志,2010,19(3)：1678-167973 Ma T, Tan C, Zhang H, et al. Bridging the gap between traditional Chinese medicine and systems biology:the connection of Cold Syndrome and NEI network. [J]. Mol Biosyst.2010 Apr;6(4):613-9.
74 Ding WJ, Zeng YZ, Li WH, et al. Identification of Linkage Disequilibrium SNPs from a Kidney-Yang Deficiency Syndrome Pedigree[J]. Am J Chin Med.2009;37(3):427-38.
75马民.闭塞性动脉硬化症血瘀证患者bcl-2和HsP70基因的表达[J].云南大学学报(自然科学版),2004,26(6A):1-5
76张明,赵英日,李强.从基因角度研究中医证候实质的思路[J].江西中医药,2009,39(9)：68-69
77 Lu CL, Qv XY, Jiang JG. Proteomics and syndrome of Chinese medicine [J]. J Cell Mol Med.2010 Dec;14(12):2721-8
78胡元会,褚瑜光,吴华琴.高血压病中医不同证候患者血清蛋白质组学研究[J].中医药中青年科技创新与成果展示论坛论文集,2009
79郎笑梅,徐珊.痰瘀证的血清蛋白质指纹图谱的临床研究[J].浙江中医药大学学报.2010,34(3)：314-315
80曾年菊,梁清华,陈疆等.蛋白质组学方法研究TPx在肝风内动证中的表达及关系[J].中华中医药学刊,2008,26(4):750-752
81郭芳宏,魏嵋.运用蛋白质组学技术研究中医证候本质进展[J].现代中西医结合杂志,2010,19(13)：1678-1679
82杨元宵,李昌煜.运用蛋白质组学研究中医证候的思考[J].中国实验方剂学杂志,2007,13(1)：69-72
83杜武勋,朱明丹,姜民等.代谢组学与中医证候研究[J].中国循证心血管医学杂志,2010,2(3)：142-145
84董飞侠,黄迪,何立群等.Ⅲ期慢性肾病肾阳虚证患者尿液代谢组学特征的研究[J].中华中医药杂志,2008,23(12):1109-1113
85郑海生,蒋健,贾伟等.慢性心力衰竭肾阳虚证患者代谢组学研究[J].中华中医药杂志,2009,S1：16-18
86任荣政,周永明,朱文伟.骨髓增生异常综合征中医证型的代谢组学分析[J].上海中医药杂志,2009,43(8):28-31
87罗和古,陈家旭.代谢组学技术与中医证候的研究[J].中国中医药信息杂志,2007,14(5)：3-5
88简维雄.中医“证候”与代谢组学研究[J].中华中医药学刊,2009,27(2):351-352
89刘凤斌,方积乾,王建华.中医药临床疗效评价的探讨[J].中药新药与临床药理,2004,4(7)：290-292.
90梁茂新,王雪峰.中医疗效评价评价指标和方法研究需要解决的认识问题[J].世界科学技术——中药现代化,2006,8(1)：31-35.
91张伯礼,康立源,项耀祖.中药新药临床试验中有关疗效评价若干问题的思考[J].中国新药与临床杂志,2007,26(11)：861-863
92申春悌,陈启光,张华强,等.建立病证结合疗效评价体系的思路和方法[J].中国中医基础医学杂志,2006,12(10):749-751
93熊宁宁,罗玫,蒋萌等.中药临床试验的适应证候设计与疗效评价[J].中国临床药理学与治疗学,2003,8(6)：715-717.
94郭新峰,赖世隆,梁伟雄.中医药临床疗效评价中结局指标的选择与应用[J].广州中医药大学学报,2002,19(4):251-255.
95王文萍,李晓斌,李晓等.临床痊愈比例超高时中医症状赋分方法对证候疗效评价的影响[J].世界科学技术—中医药现代化,2008,10(5)：33-37
96王文萍,李晓斌,李晓.探讨三种不同中医症状赋分方法对中医证候疗效评价的影响[J].世界科学技术—中医药现代化,2008,10(1)：35-38
97刘建平.循证医学方法与中医疗效评价[J].首都医科大学学报,2007,28(2):212-215
98江锋.关于中医肿瘤临床疗效评价的思考[J].中华中医药学刊,2008,26(3)：525-527
99梁晓春.循证医学与中医临床疗效评价[J].中华全科医师杂志,2004,3(1):7-9
100倪量,李峰.循证医学在中医药临床疗效评价中的应用[J].山西中医,2008,24(10)：57-58
101商洪才,王保和,张伯礼.中药新药证候及疗效评价[J].中药新药与临床药 理,2004,15(5)：365-368
102胡学军,商洪才,张伯礼等.生存质量及其量表在中医药疗效评价中的应用[J].天津中医药,2004,21(3):191-193
103汪受传,赵霞,任现志等.基于证候动态变化的病毒性肺炎疗效评价方法研究[J].世界科学技术-中医药现代化.2008,10(5)：10-15
104汪受传,赵霞,任现志等.基于主症动态变化的病毒性肺炎疗效评价方法研究[J].中华中医药杂志,2008,23(8)：675-679
105王雪峰,董丹,梁茂新等.数据挖掘技术在小儿肺炎中医临床疗效评价研究中应用的思路与方法[J].中国中西医结合杂志,2007,27(10):949-951
106沈亚诚,王小云,傅昊阳等.基于数据挖掘的绝经综合征中医药临床疗效评价研究[J].中国卫生统计,2008,25(4):387-390
107李东涛,凌昌全,郎庆波等.以证候为内容的原发性肝癌中医疗效评价体系研究[J].中西医结合学报,2007,5(1)：15-22
108郭蕾,王永炎,张俊龙.论证候的内实外虚[J].中国医药学报,2004,19(11):645-648
109申春悌.结构方程模型在异病同证研究中的运用[J].中国中医基础医学杂志,2008,14(12).953-956
110王蕾,王永炎,张志斌等.证候动态时空特征的复杂性及相应的研究思路[J].中医研究,2006,19(3):1-3
111马斌,高颖.中风病发病第7天和第14天证候要素演变规律初步研究[J].辽宁中医杂志,2006,33(12)：1561-1563.
112 Shao Yuan Chuang, Chen Huan Chen, Pesus Chou. Hypertension and the Insulin-related Metabolic Syndrome:Factor Analysis in Taiwanese. Acta Cardiol Sin,2004;20:229-36
113 Liang Han CHang. A Study on Bank Employees'Job Stress, Leisure Coping, and Physical and Mental Health. Asian Journal of Management and Humanity Sciences,2006, (1):161-174,
114潘国清.受限语境下文本语义相似度计算研究及应用[J].东南大学硕士学位论文,2007,23
115 Tian Runtao, XIE Peishan. Study on the standardization of similarity evaluation method of chromatographic fingerprints(Part D[J]. Traditional Chinese Drug Research & Clinical Pharmacology,2006, 17(1):40-42.
116王巧,许磊,冯超.金莲花药材高效液相色谱法指纹图谱分析.时珍国医国药,2010,21(2)：331-332.
117林文津,张亚敏,徐榕青,等.尿微量蛋白毛细管电泳指纹图谱的初步分析.福建医药杂志,2009,31(5)：73-75
118马欣,孙毓庆.银杏叶提取物的多维指纹图谱研究.色谱,2003,21(6)：562-567
119曹洪欣.浅谈证候动态变化的形式[J].中医药学报,1987,(3):6-7
120邱皓政.潜在类别模型的原理与技术[M].教育科学出版社,2008
121 Rost J. A latent class Model for rating data. Psychometrika,1985, 50:37-49
122李翠娟,烟建华,巩振东.对中医证本质研究的反思[J].中国中医基础医学杂志,2005,11(4)：271—272.
123张连文,袁世宏.隐结构模型与中医辨证研究——隐结构法的基本思想及隐结构分析工具[J].北京中医药大学学报,2006,29(6):365-369
124姚奎武,王阶,朱翠玲等.不同计量诊断对于冠心病血瘀证症状体征分析比较[J].世界科学技术—中医药现代化,2009,11(5)：684-688
125李稍.从维度与阶度探讨中医证候的特征及标准化方法[J].北京中医药大学学报,2003,26(3)：1-4
126郭蕾,王学伟,王永炎.论高维高阶与证候的复杂性[J].中华中医药杂志.2006,21(2)：76-78
127苏越,王磊,张华,等.肝炎肝硬化中医证候与疾病信息相似度分析[J].中国中西医结合杂志,2009,29(5)：398—402.
128操牡丹,何前锋,王柏.中医药方剂相似度模型[J].计算机工程,2009,35(16)：275—277.
129孙继佳,陆奕宇,苏式兵.基于相似匹配测度的中医证候疗效评价方法研究 [J].中国中医药信息杂志.2010,17(11):6-9
130谢平,肖婵,雷红富等.基于向量相似度原理的湖泊富营养化评价方法及其验证[J].安全与环境学报.2008,8(4):93-94
131刘素萍,仁立学,胡广春等.夹角余弦法用于辐射源一致性判定的评估[J].第十四届全国核电子学与核探测技术学术年会论文集,2008
132陶跃华.基于向量的相似度计算方案[J].云南师范大学学报.2001,21(5)：17-19
133 Liang KY, Zeger SL. Longitudinal data analysis using generalized linear models. Biometrics,1986,73:13-22
134 Lu JC, Chen D, Zhou WX. Quasi-likelihood estimation for GLM with random scales. Journal of Statistical Planningand Inference,2006,136(2): 401-429
135 Lipsitz SR, Kim K, Zhao L. Analysis of repeated categorical data using generalized estimating equations Statistics in Medicine,1994,13(1 1):1149-1163
136 Bishop J, Venables WN, Wang YG. Analysing commercial catch and effort data from a Penaeid trawl fishery:A comparison of linear models, mixed models, and generalised estimating equations approaches. Fisheries Research,2004,70(2-3):179-193
137 Pendergast JF, Gange SJ, Newton MA, el al. A Survey of Methods for Analyzing Clustered Binary Response Data. International Statistical Review,1996,64(1):89-118
138 Carr GJ, Chi EM. Analysis of variance for repeated measures Data: a generalized estimating equations approach. Statistics in Medicine, 1992,11(8):1033-1140
139黄子辉.地奥心血康胶囊治疗冠心病心绞痛的临床分析[J].医药论坛,2009,(1)：24-25
140赵晋芳,罗天娥,刘桂芬.分类重复测量数据的非线性混合效应模型[J].现代预防医学,2007,34(20)：3866-3868
141罗天娥,刘桂芬.重复测量资料非线性混合效应模型应用与实现[J].中国卫生统计,2006,23(2)：104-107