| BackgroundComing into the 21st century, as the conflict between people's needs for health and its resources aggravate day by day, it's urgent to find a way to guarantee the sustainable development of health insurance according with China's actual conditions, while Traditional Chinese Medicine (TCM) is expected to play an important role. With the changes of disease spectrum, the arrival of an aging society, the reality of the grassroots of Chinese medicine and the growing potential demand, Chinese medicine will have a greater arena, in which TCM will contribute positively to maintaining people's health, easing the pressure of public health and economy. However, Chinese Medicine is rooted in Chinese culture, its efficacy evaluation standards is totally different from Western medicine. Current efficacy evaluation standards in used are mostly according to western standards. And the so-called modernization of TCM is the simple simulation of western medicine, applying its theories to efficacy assessment of TCM blindly.What is called "standard" is a major technology system to regulate economic and social development. It has gradually become one of the most important strategies of science and technology, economic development in various countries. To improve the international competitiveness, the formulation of "standard" will become the core of national competition. The criterion of Chinese Medicine is the significant manifestation of Chinese characteristics, and is an important basis for guidance of practice. It is also the intrinsic requirement of the development of the TCM discipline. However, the TCM criterion has not got international recognition and promotion. Through summarizing the safe and effective clinical diagnosis treatment method systematically, developing the TCM standard corresponding with international convention and establishing the standard system of Chinese medicine, are the important strategic task serving for the national health, which improve of the clinical efficacy of traditional Chinese medicine.The application of quality of life (QOL), patient reported outcomes (PRO) in the field of TCM is of great importance in the course of objectiveness and standardization of efficacy evaluation for Chinese medicine. As the embodiment of efficacy in TCM is mainly on the improvement of the patients'symptoms, the amelioration of patients'syndromes are reflected on the improving quality of life to some extent. And it also aims at bettering the patients'entire condition with the guidance of the holistic concept, through the regulation of internal and external environments in the human body, to improve the quality of life.Functional status and subjective symptoms are the most basic framework in the quality of life evaluation. Therefore, as a new evaluation indicator, introducing the concept of patient reported outcomes (PRO) to the TCM clinical evaluation (i. e. quantitative diagnosis of Chinese medicine, efficacy evaluation studies can take example by the method of scales evaluation; the information collected from observing, smelling, consulting and pulse-taking, can be quantified and standardized with modern statistical analysis, etc.) is of important significance. Many studies at home and abroad have begun to use patient reported outcomes as the efficacy evaluation index of the report, and China has also begun to attach importance to the application of the assessment method to Chinese medicine, such as, national and provincial Natural Science Foundation have taken the establishment of scale evaluation system as one of the important support items, which have greatly promoted the standardization process of Chinese medicine.Nevertheless, as the objects of medical research are the patients, the calculation of PRO items relies greatly on the completion and adherence of the patients, influencing the accuracy of the study. At present, most QOL studies apply classical statistical theories, which inevitably need a large item pool for different options to meet the needs of comprehensiveness. However, the compliance of the examinees and the measurement accuracy will reduce significantly due to the complicated test, resulting in the low test quality and feasibility. If the researchers want to reduce the quantity of items and improve accuracy, the coverage will be bound to lessen, which can not meet the needs for the test standard of all the examinees. Moreover, studies based on the traditional statistical theories rely heavily on the samples and their ability levels. If all the examinees can or cannot pass the test completely, it's unable to compute the difficulty and discrimination so as the intrinsic disparity not being reported in the samples of minor test differentiation. For the particular ability or knowledge tests, researchers must use the same set of test questions, or test results can not be compared and so on. The development of Item response theory (IRT) and computer programs provides a solid theoretical basis for the optimization of quality of life and settlement of the problems above. The application IRT and computerized adaptive testing (CAT) in the field of of quality of life, make it possible to optimize QOL researches, which could be more precise, comprehensive, simplified and standardized. Meanwhile, it can provide a reliable basis for the development of quality of life and patient reported outcomes in the TCM efficacy evaluation system.ObjectiveTo have the pilot study of development of the gastralgia PRO-CAT with the item response theory and computer programmes. The study aims at establishing the system not only for the clinical efficacy evaluation of TCM spleen and stomach diseases, but also for the quantified, precise, standardized and optimized assessment method of Chinese "syndrome" information. The research strives to acquire more and more accurate information of clinical efficacy evaluation with less items and more standardized procedures, to build TCM clinical efficacy evaluation system, facilitating the summary and comparison of different research centers, which is for the promotion and modernization of TCM.Method The study first determine the test structure, IRT model structure (two-parameter model) and the item category (dichotomous items) after passing through the digestive PRO scales, documents and the information collected from the interviews with patients and clinicians. After the discussion of the experts, the information of stomach enters the draft pool (with 103 items). Then small-range test is conducted to acquire the items responsiveness and difficulty. Modify the items not being understood well and low pertinence, according to the patients and experts opinion, and form the item banks.Eligible samples of 167 gastralgia patients and 33 healthy people are selected for the survey. Retrieve the scales and get rid of the unqualified ones. Test the unidimensionality and internal consistency, screen the items again according to the size of factor loading, and form item bank again (with 76 items).Assess the parameters of remaining items with two-parameter logistic models, compute their location and slope value, and get the estimation of item information functions of different ability parametersθwith the results above. Confirm the frequency analysis of Item slope values to evaluate the reasonableness of their distribution. Conduct the Monte Carlo simulation before CAT test. Build interface for testing, enter the value of the item location and slope into the FastTest Pro software, obtain the item information function graph. Define the first entry items, analysis models, item selection criteria and termination criteria.ResultsThrough the review of literature, inventory, and the information from physicians and patients,159 items were obtained. Conduct small-scale testing in 10 examinees, and test the items responsiveness and difficulty. Modify the items not being understood well, form an item bank of 103 items. Selected eligible sample of 167 stomach ache patients and 33 healthy people for the survey (213 scales were gathered,13 of them were rejected for their unqualifiedness). Then have the statistic analysis.Unidimensional tests suggest that:in gastralgia quality dimension, KMO test and Bartlett test showed suitable for factor analysis. And gradually remove the items (F8, F10, S3, S2, S23, F5, F10) of smaller factor loading and relatively high correlation between the items. The first feature value 5 times higher than the second Eigen value is observed, suggesting that the unidimensionality test is confirmative. In the gastralgia degree dimension:KMO test and Bartlett test showed suitable for factor analysis and the unidimensionality is confirmative. Dimension of anxiety:KMO test and Bartlett test showed suitable for factor analysis, gradually remove the items (A2, A4, A19, A6, A21, A3, A22) of smaller factor loading and relatively high correlation between the items, and the first feature value 5 times higher than the second Eigen value is observed, suggesting that the unidimension test is confirmative. Dimensions of social relations:in the first factor analysis, the first feature value has not been conducted 5 times larger than the second characteristic value. Gradually remove the items R13, R10, R5, R16, R15, R12, R9, according to factor loading, correlation between items, graded, and re-conduct the factor analysis. It suggests that the result meets the standards of unidimensionality. Dimensions of social function:the initial factor analysis has not passed through the unidimensionality test. Gradually remove the items FF1, FF10, FF2, FF11, FF14, FF9, FF15, FF3, FF16, according to factor loading, correlation between items, and re-conduct the factor analysis, and the unidomensionality is acquired. Conduct the personal separation index for each dimension, suggesting that separation index is generally greater than 0.8, which shows the good internal consistency. Personal separation index is conducted to test the internal consistency. The index of gastralgia property dimension is 0.910, the one of gastralgia degree is 0.932, and the anxiety dimension is 0.941, which are more than 0.9, indicating the good internal consistency. However, the indexes of social relation and function are 0.649 and 0.753, which are less than 0.8, indicating poor internal consistency.Then model fit is conducted among these 5 dimensions, which suggests that the dimensions of gastralgia property, gastralgia degree and anxiety show good comparative fit index and non-normed fit index, while the other two dimensions not. Item banks of these two dimensions should be further modified and enlarged in the future study.Item difficulty distribution analysis:the items slope (b) distribution of first three dimensions (gastralgia quality, extent, anxiety dimension) is uniform and reasonable, which meets the needs of examinees of different characteristics levels. And in the social relation and social function dimensions, difficulty distribution bias is shown, suggesting that items of some extent are lack and there may be ceiling or floor effects in tests. Referring to the two latter dimensions, item banks should be further enlarged and adjusted to obtain a more comprehensive item pool.Test the model fit and confirm the entry and end criterion with Monte Carlo Simulation. Generate 1000 simulees by the procedure. In the simulation of stomach property dimension, the maximum amount of item information (29.609) is acquired when the ability parametersθ=-0.50. As for the item bank standard error function, the item bank minimum standard error of 0.1838 is obtained, which meets the presupposed standard error of 0.3. In the stomach degree dimension, the maximum amount of information 14.897 can be obtained when the ability parameters are -0.25. As for the item bank standard error function, the item bank minimum standard error of 0.2591 is obtained, which meets the presupposed standard error of 0.3.In the anxiety dimension, the maximum amount of information 17.733 can be obtained when the ability parameters are -0.25. As for the item bank standard error function, the item bank minimum standard error of 0.2375 is obtained, which meets the presupposed standard error of 0.3.Referring to the remaining two dimensions, social relation and social function dimensions, the items have not been included in the CAT process, because of their poor item distribution. In the next stage study, further adjustment of the item pool, CAT simulation and testing should be operated.Take gastralgia quality dimension for example, establish the interface of human-computer interaction test, input the value of the item location, slope into the FastTest Pro software, and get the item information function graph. Define the first entry items, analysis model, item selection criteria, terminal criteria, and the gastralgia quality computerized adaptive tests program is achieved.Conclusion Gastralgia PRO-CAT is an instrument of reasonable distribution and effectiveness. The process of establishing CAT is in accordance with the relevant methods overseas. Its internal consistency is good and it is easy to use, can obtain accurate results with less items, saving lots of time and reducing the respondents burden. Therefore, it is suitable for clinical research applications.
|
PDF Full Text Request