Analysis Of Functional Dyspepsia Scale Of Traditional Chinese Medicine Based On Syndrome Differentiation Using Multidimensional Item Response Theory Method

ObjectiveThe purpose of this study is to extract potential features that can reflect the nature of syndrome essence from symptom items by using multidimensional item response theory, and analyze the contribution of different symptoms to their relevant syndromes; And based on the cross relationship of syndrome and its corresponding symptoms, combine the scale dimension structure with TCM syndrome models, explore the correlation between the syndrome types; Develope a clinical dialectical scale can be applied to functional dyspepsia, and research methods of quantitative syndrome differentiation of traditional Chinese medicine from the Angle of mathematics method.MethodsThis study adopted a cross section sampling methods and used Traditional Chinese Medicine initial scale in functional gastrointestinal disease to collect samples. First, core symptoms, syndrome elements and their combination rules were extracted based on literature analysis and through several rounds of expert consultations and discussions. Then TCM syndrome model was set up bottomed on the foundation of single syndrome. Finally, Functional Dyspepsia quantitative syndrome differentiation rules were established through the following steps:analyzing contribution of different symptoms to syndromes by using Structural Equation Model combined and Multidimensional Item Response Theory, calculating weighted value of related symptoms according to contributions to different syndromes and using Receiver Operating Characteristic curve analysis method to determine syndrome diagnosis threshold.ResultIn this study, syndromes of FD were divided into the following eight types:the spleen and stomach Qi deficiency syndrome, stagnation of liver Qi syndrome, stagnation of Qi syndrome, stagnation of phlegm dampness syndrome, cold syndrome, spleen Yang deficiency syndrome, heat syndrome and stomach Yin deficiency syndrome. Because of the lack of samples collected, syndromes that did not meet the minimal sample level were excluded. Finally, there were only five syndromes were kept in the study.These syndromes were the spleen and stomach Qi deficiency syndrome, stagnation of liver Qi syndrome, stagnation of Qi syndrome, stagnation of phlegm dampness syndrome and heat syndrome.1.SamplesIn this study,507 qualified samples were collected, among which 193 were men(38.07%) and 314 were women(61.93%), the proportion of women was about 1.5 times as much as men. In these patients,216 were young people(age≤35), which accounted for 42.60% and 247 were middle-aged people (age between 36 to 55).which accounted for 48.72%. These results indicated that FD were more common in young and middle-aged people and the incidence in women was slightly higher than that in men,which were consistent with the FD epidemiological survey results. The statistical results of FD progress showed that cases with a course under 1 year accounted for 28.80% and cases with a longer course accounted for 71.20%. These agreed with the characteristics of FD as a long-course and easy to recurrent disease. This study included seven syndromes, the spleen and stomach Qi deficiency syndrome, dampness-heat of spleen and stomach syndrome and liver-stomach disharmony syndrome accounted for about three-quarters of the total collected samples.2. Frequency statistical resultsThe questionnaire used in this study was designed for functional gastrointestinal diseases, which covered a variety of diseases, but this study was only aimed to analyze FD. Therefore in order to avoid the scale was too complex, items were first analyzed by using frequency analysis and then screened by combining with the theory of TCM. At first, the scale contained 104 items, including 66 symptom items,27 tongue-appearance items and 11 pulse-condition items. After item screening,49 items were deleted and 55 were reserved. As for the spleen and stomach Qi deficiency syndrome, items referring to sweating, expectoration and relaxed pulse were deleted. In heat syndrome, items referring to xerostomia and diarrhoea were deleted. Items referring to discomfort in flank was kept because it was significant for diagnosing stagnation of liver Qi syndrome, although it occurred with low frequency. In stagnation of phlegm dampness syndrome, items referring to heavy sensation of head, diarrhoea and tongue with glossy coating were deleted.In stagnation of Qi syndrome, items referring to chest tightness, have loose bowels when having a stomachache and heavy feeling in abdomen were deleted.3. SEM analysis resultsSEM model were set up including five potential variables:spleen qi deficiency syndrome, liver depression syndrome, and qi stagnation, dampness syndrome, heat syndrome, also contains four paths:liver depression and qi stagnation, spleen deficiency with dampness retention,, dampness resistance and qi stagnation, retention of dampness generating interior heat. Using SEM to analyze the model, the main model fitting index as chi-square/DF= 1.42, CFI=0.87, GFI=0.93, NFI=0.83, NNFI=0.85, RMSEA=0.046, ECVI=2.37, RMR= 0.052, the above data show that the fit of the described syndrome differentiation model is ideal, the model is acceptable. Considering factor loading coefficient of each factor, the smaller loading coefficient of items concluded:irritability0.27, pantothenic acid0.23, urine yellow short 0.30, vomiting and nausea0.15, lazy words0.28, tastelessness-0.25, Stool dry and then sloppy 0.13, stool dry rot 0.09, spotted tongue 0.06 and plump tongue 0.26. And the significance test results (t> 1.96 at 0.05 significant level) show that items not passed including:pantothenic acid, urine yellow short, tastelessness, Stool dry and then sloppy, stool dry rot, spotted tongue and plump tongue. While all path coefficients reached significant level. Combining chinese medicine practical theory, tastelessness,]teeth-marked tongue, Stool dry and then sloppy, stool dry rot,4 items were removed. Again after correction, model tests, the results show that the model fitting index: chi square/DF=1.50, CFI=0.92, GFI=0.88, NFI=0.93, NFI=0.86, RMSEA-0.042, ECVI=2.41, SRMR=0.049, the fit of the model is improved from the previous, the model is acceptable. SEM model analysis results illustrate functional dyspepsia sydrome conversion direction:under the development of liver depression syndrome with qi deficiency or qi stagnation syndrome, spleen and stomach syndrome and phlegm wet resistance, the phlegm dampness syndrome and qi stagnation syndrome, and relationship between heat syndromes for direct effect; Spleen qi deficiency syndrome and qi stagnation syndrome, and the relationship between the heat syndromes for indirect effect.4. MIRT analysis resultsBy using MIRT analysis methods and according to the overall degree of differentiation in items, item spotted-tongue was deleted.Weighted value of symptom items were calculated according to contributions to syndromes. After the calculation,507 samples were scored again. Then a ROC curve was drawn with syndrome differentiation diagnosed by experts as state variables and new scores of syndromes as test variable.Finally, diagnosis threshold of the spleen and stomach Qi deficiency syndrome, stagnation of liver Qi syndrome, heat syndrome, stagnation of phlegm dampness syndrome and stagnation of Qi syndrome were determined. The diagnosis threshold of these syndromes and weighted value of symptoms were as follows:(1)The diagnosis threshold of spleen and stomach Qi deficiency syndrome was 18,with easy to feel tired (1), fatigue (1), anorexia (1) poor appetite (1), early satiety (1), dull pain (1), pressure-relieved symptoms (1), worse after eating symptoms (1), loose stools (1), pale tongue or fat and tender tongue (1) and thin pulse (1).(2)The diagnosis threshold of stagnation of liver Qi syndrome was 11, with symptoms worse with depression(4), discomfort in flank (2), easy to be agitated(2) and stringy pulse(2).(3)The diagnosis threshold of stagnation of phlegm dampness syndrome was 11, with feeling of a heavy body(3), nausea(2), heavily coated tongue (2), greasy coated tongue(2) and slippery pulse(2).(4)The diagnosis threshold of heat syndrome was 15, with burning sensation(1),acid regurgitation(1), dry and bitter taste in mouth(2), constipation(1), tongue with yellow coating(2), reddened tongue(2) and rapid pulse(1).(5)The diagnosis threshold of stagnation of Qi syndrome was 13, with poor appetite (1),obstructivedefecation(1), belching(2), distending pain(2), uncomfortable fullness feel ing(2), pressure-aggravated symptoms (1).Although the five syndrome classification diagnosis thresholds were different, the sensitivity of the scale was between 80%to 90% and specificity was between 84% to 100%. The area under the ROC curve of these syndromes were more than 0.85, which met the minimum requirement of diagnostic test. The area under the ROC curve of the spleen and stomach Qi deficiency syndrome, stagnation of liver Qi syndrome, heat syndrome and stagnation of phlegm dampness syndrome were more than 0.9, which indicated that the scale possessed good syndrome diagnosis capability and accuracyConclusionThis study was conducted based on the existing scale and with a more comprehensive and standard symptoms collection,so the results are both scientific and reliable.The scale has been simplified by using Structural Equation Model and Multidimensional Item Response Theory methods. The dimensional model obtained after analyzed by using Multidimensional Item Response Theory is close to the experts-diagnosed model, which confirms that the theory can be used in quantitative research of syndromes.The five scales for FD patients established in the study possess good sensitivity and specificity in syndrome diagnosis, which are suitable for using in case screening and choosing. The ideas and methods of this study will provide as a useful model for quantitative research in other diseases.