| Background:Relevant description of coronary artery disease(CAD)in traditional Chinese medicine(TCM)documents dates back thousands of years.By records and through practice,practitioners of TCM have proven that TCM therapies can relieve the symptoms of CAD patients and improve their quality of life.TCM bases its therapy on syndrome differentiation,and integrates it with nowadays diagnosis of diseases,so that patients receive individualized therapies.This characteristic of TCM can be considered as an early form of precision medicine.However,syndrome differentiation in TCM is mostly based on a practitioner’s subjective opinion,with a list of indicators of certain syndrome to be considered.This massive quantity of indicators partly restricts a broader clinical application of TCM theory.For nearly half of a century,many researchers have explored the correlation between TCM syndromes and different objective indicators,expecting to broaden the connotation of TCM syndromes and improve their objectivity;some researchers even succeeded in building diagnostic criteria of blood stasis syndrome that integrated TCM and objective indicators.Then here comes the question:is it possible to construct a simplified diagnostic solution of basic TCM syndromes of stable coronary artery disease(SCAD)that combines both TCM and objective indicators,so that the threshold of applying TCM syndrome differentiation is lowered?To work out the answer to this question,we first did literature researches,and constructed a preliminary syndrome differentiation solution through two rounds of Delphi consultation.Then we evaluated and improved this preliminary syndrome differentiation solution using real world clinical data from a large sample,in the hope of providing experience and evidence for the construction of simplified syndrome differentiation criteria in the future.Study 1:Delphi consultation of a simplified,integrative solution for SCAD syndrome differentiationObjective:To screen TCM and objective indicators relevant to each basic syndrome of SCAD,and construct a preliminary simplified,integrative syndrome differentiation solution through Delphi consultation.Methods:Based on literature researches,we did an initial screen of indicators based on their frequency of appearance and designed the questionnaires.Through two rounds of Delphi consultation,we invited experts to grade each indicator.The questionnaires’ credibility was assessed by calculating the coefficient of experts’ authority(Cr)and Cronbach’s alpha;coordination between experts’ opinions was assessed by calculating Kendall’s W;indicators were screened based on their average score,full score ratio and coefficient of variation.Results:96.9%and 100%of questionnaires were returned in the first and second round of consultation,all questionnaires returned were valid.The coefficient of experts’ authority was 0.975,indicating a high level of authority of the experts.Coefficients of Cronbach’s alphas were 0.967 and 0.959 respectively in the first round and second round,indicating that the result’s credibility was high.Coefficients of Kendall’s W were 0.327 and 0.366(P<0.001)in the first round and second round,indicating that the experts’ opinions were coordinated.Through two rounds of Delphi consultation,a simplified,integrative solution for SCAD syndrome differentiation with 41 indicators in all was formed.In 41 indicators,8 were for blood stasis syndrome,5 for turbid phlegm syndrome,6 for qi stagnation syndrome,7 for qi asthenia syndrome,6 for yin asthenia syndrome and 9 for yang asthenia syndrome.Conclusion:The credibility of the results and authority of the experts were relevant high.It was shown through the results that experts’ opinions towards TCM indicators were more coordinated than that of objective indicators,which reflected that achievements and consensus in this field were lacking.Study 2:Construction and evaluation of a simplified,integrative solution for SCAD syndrome differentiationObjective:To examine the relevance between indicators and TCM syndromes in the simplified,integrative solution for SCAD syndrome differentiation,and to evaluate the diagnostic performance of a simpler solution on this basis.Methods:Data from Capital’s Funds for Health Improvement and Research(No.2018-1-4171)program was used.TCM terms were unified using TCM clinical terms platform developed by Institute of Information on Traditional Chinese Medicine.Syndrome differentiation was based on syndrome diagnostic criteria for coronary heart angina.That whether a patient was diagnosed of a certain TCM syndrome was treated as a binomial dependent variable,and logistic regression analysis was adopted to construct models of each syndrome’s simplified differentiation solution.Relevance between indicators and syndromes and the models’ accuracy,sensitivity,specificity,consistency in diagnosis and clinical application values were analyzed.Results:1.The area under receiver operating characteristic(ROC)curve(AUC)of the logistic model for simplified blood stasis syndrome differentiation solution was 0.904.The sensitivity and specificity of the model were 88.9%and 83.6%respectively.The Kappa index for consistency check was 0.52.Independent variables that were included by the final model were:fixed chest pain,dark purple tongue or tongue with ecchymosis,and imaging indicators or medical history of thrombosis or infarction in regions other than coronary arteries.2.The AUC of the logistic model for simplified turbid phlegm syndrome differentiation solution was 0.906.The sensitivity and specificity of the model were 80.7%and 87.7%respectively.The Kappa index for consistency check was 0.59.Independent variables that were included by the final model were:heavy coating on the tongue,excessive phlegm and hyperlipidemia.3.The AUC of the logistic model for simplified qi stagnation syndrome differentiation solution was 0.811.The sensitivity and specificity of the model were 89.5%and 66.9%respectively.The Kappa index for consistency check was 0.06.Independent variables that were included by the final model were:swelling pain below the costal region,irritable temper,and anxiety.4.The AUC of the logistic model for simplified qi asthenia syndrome differentiation solution was 0.954.The sensitivity and specificity of the model were 87.5%and 92.3%respectively.The Kappa index for consistency check was 0.73.Independent variables that were included by the final model were:lacking in strength,shortness of breath,and exertional angina.5.The AUC of the logistic model for simplified yin asthenia syndrome differentiation solution was 0.817.The sensitivity and specificity of the model were 79.2%and 78.5%respectively.The Kappa index for consistency check was 0.19.Independent variables that were included by the final model were:thirst,night perspire and diabetes mellitus.6.The AUC of the logistic model for simplified yang asthenia syndrome differentiation solution was 0.801.The sensitivity and specificity of the model were 72.7%and 76.6%respectively.The Kappa index for consistency check was 0.17.Independent variables that were included by the final model were:intolerance of cold,lacking in strength,and elevated BNP or NT-proBNP.7.The positive predictive value and negative predictive value for simplified blood stasis syndrome differentiation solution were 46.2%and 97.9%respectively,57.4%and 95.6%for turbid phlegm,11.5%and 94.3%for qi stagnation,70.6%and 97.2%for qi asthenia,24.5%and 93.2%for yin asthenia and 21.9%and 92.3%for yang asthenia.Conclusion:Most logistic regression models for the simplified syndrome differentiation solution performed well in diagnostic accuracy,sensitivity,and negative predictive value,but differed greatly in specificity and positive predictive value.This indicated that TCM indicators were sensitive for diagnosis of certain syndrome,and that oversimplifying the syndrome differentiation solution would lower the specificity of the model.Appropriate objective indicators for syndromes like qi stagnation and yin asthenia were lacking,and called for further researches.Overall model performance of simplified syndrome differentiation solutions for blood stasis,turbid phlegm and qi asthenia was good and robust,and these models were worth further evaluation and examination in clinical practice.In all,we have constructed a preliminary simplified,integrative solution for SCAD syndrome differentiation based on literature researches and Delphi consultation,and evaluated and improved this solution with logistic regression analysis using large sample data from real world study.We have explored the feasibility of constructing a simplified syndrome differentiation solution and the potential value of applying it in clinical practice,thus providing experience and evidences for its further evaluation and application in clinical environment. |
PDF Full Text Request