The Preparation And Application Of The Test Cases Of TCM Syndrome Elements Diagnosis Model

Objective:Preparing contain normative syndrome and syndrome elements diagnostic information test case, and ensure the reliability of syndrome elements diagnosis, provide test cases for testing the accuracy, sensitivity, special effects, precision and other indexes of syndrome element diagnostic mode (such as support vector machines and grade of syndrome elements), which is trained by machine learning.Methods:1. Choose 50 cases from large Chinese medicine clinical case database of the regulated four diagnostic information.2. Adopt two-round Delphi consultation, invite 4 field experts to make the syndrome element diagnosis targeting each case, the interval of the two rounds should be over one week. In the first round, only provide the four diagnostic information of 50 clinic cases, invite experts diagnose one by one; in the second round, in the premise of supporting reference of grade of syndrome element and interrupting the case order, invite experts to diagnose the above 50 cases.3. Adopt Kenders Harmonious Coefficient as the consistency reliability evaluation index of two-round expert consultation, uses the kappa coefficient as the test-retest reliability evaluation index of two-round experts’ syndrome element diagnosis.4. In order to avoid the subjective bias of the syndrome, the identification accuracy of the constructed model can be negatively affected. Our study invited 4 experts in the field of traditional Chinese medicine to make a diagnosis in terms of each case, use expert voting to gather the syndrome diagnosis results of experts, the diagnosis for each syndrome elements is established, each expert were cast a vote, more votes syndrome elements were the final results of the case syndrome elements diagnosis.5. Select the higher Kendall harmonious coefficient average of syndrome elements diagnose 50 cases of a round to be traditional Chinese medicine identification of the initial screening case, in each case, use the expert vote method to generated the result of syndrome element diagnosis. if the result of the case passes through the reliability evaluation, the case is the required Chinese medicine syndrome element diagnosis testing case. If not, exclude it.Result:1. The expert syndrome element diagnosis results in the first round have 37 syndrome elements, after the consistency reliability evaluation, the consistency of 23 syndrome diagnosis is good (account for 62%) including water retention, phlegm and blood stasis, has the statistic significance(P<0.05).2. The expert syndrome element diagnosis results in the second round have 29 syndrome elements, after the consistency reliability evaluation, the consistency of 19 syndrome diagnosis is good (account for 66%) including blood deficiency, qi stagnation and liver, has the statistic significance(P<0.05).3. Comparison of the consistency of different syndrome elements in two rounds of expert consultation, the mean value of Kendall’s harmony coefficient in the first round of syndrome elements diagnosis was 0.388, in the second round of syndrome elements diagnosis was 0.451.4. The syndrome elements of two-round expert consultation have 38, after the test-retest reliability evaluation, the test-retest reliability evaluation of 22 syndrome elements is passable including cause blood disorder, liver and sputum, has the statistic significance(P<0.05).5. The syndrome element diagnosis passing reliability evaluations(2 Consistency reliability evaluations and one test-retest reliability evaluation) have 16, namely, syndrome of phlegm, blood stasis, qi depression, heart, yin deficiency, Lung, liver, kidney, qi deficiency, heat, external wind, dampness, spleen, stomach, cause blood disorder and uterus.6. Use Chinese medicine syndrome element diagnosis mode test case preparation method, through two-round expert consultation, evaluate the results with consistency reliability evaluation and test-retest reliability evaluation, there are 43 Chinese medicine syndrome element diagnosis mode test case in the 50 clinic cases.Conclusion:1. In this experiment that four experts on 50 cases of clinical cases of two rounds of expert consultation prompt:accounting for more than 60% of the syndrome elements diagnostic consistency is better.2. In this experiment that four experts on 50 cases of clinical cases of two rounds of expert consultation prompt:the proportion of the 58% syndrome elements of the test-retest reliability is passable.3. Under the premise at lack of standard of syndrome elements diagnosis, Several TCM experts make the syndrome elements diagnostic of the same clinical case have inconsistency, this will affect the performance of the syndrome elements diagnostic model evaluation results.4. This study is under the premise of ensuring the consistency and reliability, the preparation method of a set of test cases of TCM syndrome diagnosis model was explored.