A traditional Chinese medicine diagnosis and treatment decision-making method and system based on evidence-based evidence

Abstract

The application provides a traditional Chinese medicine diagnosis and treatment decision-making method and system based on evidence-based evidence, relates to the technical field of diagnosis and treatment decision-making, and the method comprises the following steps: acquiring patient information, screening evidence-based evidence, classifying syndromes, generating a personalized treatment plan, collecting feedback and monitoring reactions, quantifying the curative effect, and dynamically adjusting the plan. The method is systematic and scientific, and has good application prospect and popularization value. Through the feedback mechanism and multi-round iterative learning, the application continuously optimizes the diagnosis and treatment plan, so that the diagnosis and treatment plan can adapt to the changing needs of the patient; by means of statistical analysis, the model is more interpretable, which helps doctors understand the decision-making process and improves the clinical confidence.

Description

Technical Field

[0001] This invention relates to the field of diagnostic and treatment decision-making technology, and in particular to a method and system for TCM diagnostic and treatment decision-making based on evidence-based principles. Background Technology

[0002] Traditional Chinese medicine (TCM), a medical system with a history of 5,000 years, is based on the core principles of holistic medicine and syndrome differentiation. TCM emphasizes personalized diagnosis and treatment according to the individual differences and characteristics of each patient's condition, boasting advantages such as definite clinical efficacy, relatively safe medication, flexible treatment methods, and relatively low cost. However, in actual clinical practice, doctors often need to adjust the combination and dosage of Chinese herbal medicines according to the patient's specific situation to achieve better therapeutic results. This process demands a high level of experience and professional knowledge from the doctor, especially for inexperienced young TCM doctors, mastering these complex adjustment methods is a challenge.

[0003] The diagnostic process in Traditional Chinese Medicine (TCM) typically relies on the doctor's subjective judgment, including traditional diagnostic methods such as observation, auscultation and olfaction, inquiry, and palpation. This inherent subjectivity and uncertainty make TCM diagnosis and treatment inherently subjective. Furthermore, a vast amount of TCM knowledge and experience is often scattered across ancient and modern literature and the minds of renowned veteran TCM doctors, awaiting further exploration and systematization. Therefore, transforming this tacit knowledge into explicit knowledge that can be understood and applied by computers has become a crucial issue in promoting personalized TCM treatment.

[0004] With the development of medical big data, artificial intelligence, and knowledge graph technologies, it has become possible to organize, analyze, and apply traditional Chinese medicine (TCM) knowledge using modern technological means. By constructing a TCM clinical knowledge graph and combining it with the processing and fusion of multimodal information, the accuracy and reliability of TCM diagnostic and treatment decisions can be effectively improved. This method not only helps doctors better understand and apply TCM theories but also, to some extent, addresses the knowledge sparsity and cold-start problems in TCM diagnosis and treatment.

[0005] Therefore, multimodal information fusion technology based on TCM knowledge graph can provide more scientific and systematic support for TCM diagnosis and treatment, promote the discovery and application of TCM knowledge, and drive TCM diagnosis and treatment towards a more precise and personalized direction.

[0006] The invention patent application number 202211281449.0 discloses a method and device for TCM auxiliary decision-making that combines clinical knowledge graphs with multimodal information fusion, involving the fields of medical big data, knowledge graphs, and artificial intelligence. Existing technologies address the characteristics of TCM, which emphasizes inheritance, practice, and is a complex system with a large amount of tacit knowledge existing in individual case studies. Existing technologies combine heterogeneous network embedding, graph neural network reasoning, and deep learning interpretability techniques to achieve an integrated framework for TCM four diagnostic methods multimodal data acquisition, information extraction, correction filtering, and information fusion decision-making under the construction of TCM knowledge graphs and representation learning. This approach has advantages such as small-sample learning and strong interpretability, and shows good application prospects in supporting the construction of personalized knowledge graphs for specialties / diseases / renowned doctors through human-machine collaboration, and promoting the transformation of TCM diagnostic knowledge discovery into diagnostic and treatment auxiliary decision-making. However, while this technology mentions enhancing the interpretability of the model through feedback iterative learning, it does not detail the specific implementation of the feedback mechanism and the method for evaluating its effectiveness. This deficiency may easily lead to inaccurate diagnostic and treatment decision results and insufficient confidence. Summary of the Invention

[0007] To overcome the shortcomings of existing technologies, the purpose of this invention is to provide an evidence-based TCM diagnosis and treatment decision-making method and system. Through feedback mechanisms and multi-round iterative learning, the diagnosis and treatment plan is continuously optimized to ensure that it adapts to the changing needs of patients. Statistical analysis is used to enhance the interpretability of the model, help doctors understand the decision-making process, and improve clinical confidence.

[0008] To achieve the above objectives, the present invention provides the following solution:

[0009] An evidence-based TCM diagnosis and treatment decision-making method includes:

[0010] Obtain basic information, traditional Chinese medicine diagnostic data, and modern medical examination data of the target patients;

[0011] Based on the established TCM evidence-based database, evidence is screened according to the basic information, TCM diagnostic data and modern medical examination data to obtain evidence-based evidence.

[0012] The TCM diagnostic data and the evidence-based evidence are input into the constructed syndrome identification model to classify TCM syndromes and obtain syndrome identification results.

[0013] Based on the results of syndrome differentiation, a personalized treatment plan is generated by combining classical Chinese medicine theories and evidence-based medicine.

[0014] Collect patients' subjective feedback and objective data, and monitor patients' treatment response in real time through follow-up questionnaires, electronic medical records, and wearable devices;

[0015] Based on statistical analysis methods, the efficacy is quantified according to the subjective feedback, the objective data, and the treatment response to obtain a quantitative efficacy value. The quantitative efficacy value is then compared with the personalized treatment plan to obtain a comparison result.

[0016] The parameters generated by the syndrome identification results and personalized treatment plans are dynamically adjusted based on the comparison results.

[0017] Through multiple rounds of iterative learning, the parameters of the syndrome identification results and personalized treatment plans are continuously adjusted to make the syndrome identification results and personalized treatment plans more closely match the individual needs of patients.

[0018] Preferably, the basic information includes age, gender, medical history, and lifestyle habits; the TCM diagnostic data includes chief complaint, symptoms, tongue appearance, and pulse appearance; and the modern medical examination data includes blood test results and imaging data.

[0019] Preferably, based on the constructed TCM evidence-based database, evidence is screened according to the basic information, the TCM diagnostic data, and the modern medical examination data to obtain evidence-based evidence, including:

[0020] A database of evidence-based traditional Chinese medicine (TCM) is constructed, comprising classical TCM literature, modern TCM research findings, and clinical trial data.

[0021] Natural language processing technology is used to perform semantic analysis on the basic information, the traditional Chinese medicine diagnostic data, and the modern medical examination data to extract evidence data related to the patient's symptoms;

[0022] Based on Bayesian networks, the most relevant evidence-based evidence is selected from the TCM evidence-based evidence database; the selection criteria for the evidence-based evidence include: evidence level, applicability, and safety.

[0023] Preferably, the symptom identification model is trained using a CNN-LSTM combined network model.

[0024] Preferably, the personalized treatment plan includes: a traditional Chinese medicine prescription, an acupuncture plan, and dietary recommendations.

[0025] Preferably, based on statistical analysis methods, the efficacy is quantified according to the subjective feedback, the objective data, and the treatment response to obtain a quantified efficacy value. The quantified efficacy value is then compared with the personalized treatment plan to obtain a comparison result, including:

[0026] Define quantitative indicators of therapeutic efficacy; these quantitative indicators include therapeutic efficacy scores and rates of change.

[0027] The subjective feedback and objective data are used to score the treatment effect;

[0028] The efficacy score is calculated based on the subjective feedback score and the objective data score; Efficacy score = (subjective feedback score + objective data score) / 2;

[0029] Calculate the rate of change; Rate of change = (Indicators of objective data after treatment - Indicators of objective data before treatment) / Indicators of objective data before treatment × 100%;

[0030] The value of the efficacy quantification index is determined as the efficacy quantification value;

[0031] The efficacy quantification value is compared with the personalized treatment plan to analyze the relationship between efficacy and treatment plan, and the comparison results are obtained; the comparison results include correlation analysis results and efficacy difference assessment results.

[0032] Preferably, the parameters generated by dynamically adjusting the syndrome identification results and personalized treatment plans based on the comparison results include:

[0033] The comparison results were statistically analyzed to calculate the differences in efficacy between different treatment regimens, and the key factors affecting efficacy were identified based on these differences; the key factors included the patient's individual characteristics and symptom changes.

[0034] The adjustment strategy will be determined based on the key factors and different treatment options;

[0035] The adjustment strategy was applied to the syndrome identification model, the model parameters were updated, and the model was retrained using the gradient descent algorithm to adapt to the new parameter settings.

[0036] Based on the updated syndrome identification results, a new personalized treatment plan is generated.

[0037] Preferably, through multiple rounds of iterative learning, the parameters of the syndrome identification results and the personalized treatment plan are continuously adjusted to make the syndrome identification results and the personalized treatment plan more closely match the individualized needs of the patient, including:

[0038] After implementing the new personalized treatment plan, continue to collect subjective feedback and objective data from the target patients, and return to the step "Based on statistical analysis methods, quantify the efficacy according to the subjective feedback, the objective data and the treatment response, obtain the efficacy quantification value, and compare the efficacy quantification value with the personalized treatment plan to obtain the comparison result";

[0039] A closed-loop control process is formed through a feedback mechanism to continuously monitor the efficacy and make dynamic adjustments.

[0040] An evidence-based TCM diagnosis and treatment decision-making system includes:

[0041] The data acquisition unit is used to acquire basic information, traditional Chinese medicine diagnostic data, and modern medical examination data of the target patient.

[0042] The evidence acquisition unit is used to screen evidence based on the constructed TCM evidence-based evidence database, according to the basic information, the TCM diagnostic data and the modern medical examination data, to obtain evidence-based evidence.

[0043] The result acquisition unit is used to input the TCM diagnostic data and the evidence-based evidence into the constructed syndrome identification model to classify TCM syndromes and obtain syndrome identification results.

[0044] The treatment plan generation unit is used to generate personalized treatment plans based on the syndrome identification results, combined with classical TCM theories and evidence-based medicine.

[0045] The data acquisition unit is used to collect patients' subjective feedback and objective data, and to monitor patients' treatment response in real time through follow-up questionnaires, electronic medical records and wearable devices;

[0046] The comparison unit is used to quantify the efficacy based on the subjective feedback, the objective data, and the treatment response using statistical analysis methods, obtain a quantitative efficacy value, and compare the quantitative efficacy value with the personalized treatment plan to obtain a comparison result.

[0047] The feedback unit is used to dynamically adjust the parameters generated by the syndrome identification results and the personalized treatment plan based on the comparison results.

[0048] The iterative learning unit is used to continuously adjust the parameters generated by the syndrome identification results and the personalized treatment plan through multiple rounds of iterative learning, so that the syndrome identification results and the personalized treatment plan are closer to the individualized needs of the patient.

[0049] According to specific embodiments provided by the present invention, the present invention discloses the following technical effects:

[0050] This invention provides a method and system for TCM diagnosis and treatment decision-making based on evidence-based principles. The method includes: acquiring basic information, TCM diagnostic data, and modern medical examination data of the target patient; filtering evidence based on the basic information, TCM diagnostic data, and modern medical examination data according to a pre-constructed TCM evidence-based evidence database to obtain evidence-based evidence; inputting the TCM diagnostic data and the evidence-based evidence into a pre-constructed syndrome identification model to classify TCM syndromes and obtain syndrome identification results; generating personalized treatment plans based on the syndrome identification results, combined with classical TCM theories and evidence-based evidence; and collecting subjective feedback from the patient. This invention provides a systematic and scientific TCM diagnostic and treatment decision-making method with good application prospects and promotional value. It utilizes objective data and monitors patients' treatment responses in real time through follow-up questionnaires, electronic medical records, and wearable devices. Based on statistical analysis methods, the efficacy is quantified according to the subjective feedback, objective data, and treatment responses to obtain a quantitative efficacy value. This quantitative efficacy value is then compared with the personalized treatment plan to obtain a comparison result. The parameters of the syndrome identification results and personalized treatment plan are dynamically adjusted based on the comparison result. Through multiple rounds of iterative learning, the parameters of the syndrome identification results and personalized treatment plan are continuously adjusted to better meet the individualized needs of patients. Attached Figure Description

[0051] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0052] Figure 1 A flowchart of the method provided in an embodiment of the present invention;

[0053] Figure 2 This is a schematic diagram of the system structure provided in an embodiment of the present invention. Detailed Implementation

[0054] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0055] The purpose of this invention is to provide an evidence-based TCM diagnosis and treatment decision-making method and system. Through feedback mechanisms and multi-round iterative learning, the diagnosis and treatment plan is continuously optimized to ensure that it adapts to the changing needs of patients. Statistical analysis is used to enhance the interpretability of the model, help doctors understand the decision-making process, and improve clinical confidence.

[0056] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0057] Figure 1 The method flowchart provided in the embodiments of the present invention is as follows: Figure 1 As shown, this invention provides an evidence-based TCM diagnosis and treatment decision-making method, comprising:

[0058] Step 100: Obtain the target patient's basic information, traditional Chinese medicine diagnostic data, and modern medical examination data;

[0059] Step 200: Based on the constructed TCM evidence-based database, evidence is screened according to basic information, TCM diagnostic data and modern medical examination data to obtain evidence-based evidence;

[0060] Step 300: Input TCM diagnostic data and evidence-based evidence into the constructed syndrome identification model to classify TCM syndromes and obtain syndrome identification results;

[0061] Step 400: Based on the syndrome identification results, generate a personalized treatment plan by combining classical TCM theories and evidence-based medicine.

[0062] Step 500: Collect patients' subjective feedback and objective data, and monitor patients' treatment response in real time through follow-up questionnaires, electronic medical records, and wearable devices;

[0063] Step 600: Based on statistical analysis methods, the efficacy is quantified according to subjective feedback, objective data and treatment response to obtain a quantitative efficacy value. The quantitative efficacy value is then compared with the personalized treatment plan to obtain the comparison results.

[0064] Step 700: Dynamically adjust the parameters generated from the syndrome identification results and personalized treatment plans based on the comparison results;

[0065] Step 800: Through multiple rounds of iterative learning, continuously adjust the parameters generated by the syndrome identification results and personalized treatment plans to make the syndrome identification results and personalized treatment plans more closely match the individual needs of patients.

[0066] Preferably, the basic information includes age, gender, medical history, and lifestyle habits; the TCM diagnostic data includes chief complaint, symptoms, tongue appearance, and pulse appearance; and the modern medical examination data includes blood test results and imaging data.

[0067] Optionally, step 100 in this embodiment is as follows:

[0068] Step 101: Collection of Basic Patient Information

[0069] When obtaining basic information about a target patient, data entry is first required through an Electronic Medical Record (EMR) system or a patient management system. During the patient's initial visit, medical staff collect basic information, including age, gender, medical history, and lifestyle habits, through face-to-face interviews or electronic questionnaires. This information can be collected using standardized questionnaires to ensure data consistency and completeness. Furthermore, the system should have data validation capabilities to ensure the accuracy of the entered information, for example, by setting age range and gender options to prevent incorrect input.

[0070] Step 102: Acquisition of TCM diagnostic data

[0071] The acquisition of TCM diagnostic data primarily relies on the clinical observation and diagnosis of TCM doctors. During the diagnosis and treatment process, doctors meticulously record information such as the patient's chief complaint, symptoms, tongue appearance, and pulse using traditional TCM diagnostic methods such as observation, auscultation and olfaction, inquiry, and palpation. To improve the accuracy and repeatability of the data, it is recommended to use standardized TCM diagnostic record forms to ensure that each doctor adheres to the same standards when recording. Furthermore, by incorporating modern technology, image recognition technology can be used to photograph and analyze tongue images, further improving the objectivity and accuracy of tongue image data.

[0072] Step 103: Integration of Modern Medical Examination Data

[0073] Modern medical examination data is typically acquired through hospital laboratories and radiology departments. After a patient's consultation, the doctor will order relevant tests based on the patient's condition, including blood tests and imaging examinations (such as X-rays, CT scans, and MRIs). The test results are automatically uploaded to the patient's electronic medical record through the hospital's information system (HIS). Medical staff need to regularly check and update this data to ensure that all relevant modern medical examination results are reflected in the patient's medical record in a timely and accurate manner. Furthermore, the system should have data integration capabilities to link traditional Chinese medicine diagnostic data with modern medical examination data for subsequent analysis and decision support.

[0074] Preferably, based on the constructed TCM evidence-based database, evidence is screened according to the basic information, the TCM diagnostic data, and the modern medical examination data to obtain evidence-based evidence, including:

[0075] A database of evidence-based traditional Chinese medicine (TCM) is constructed, comprising classical TCM literature, modern TCM research findings, and clinical trial data.

[0076] Natural language processing technology is used to perform semantic analysis on the basic information, the traditional Chinese medicine diagnostic data, and the modern medical examination data to extract evidence data related to the patient's symptoms;

[0077] Based on Bayesian networks, the most relevant evidence-based evidence is selected from the TCM evidence-based evidence database; the selection criteria for the evidence-based evidence include: evidence level, applicability, and safety.

[0078] Optionally, step 200 in this embodiment is as follows:

[0079] Step 201: Constructing a Traditional Chinese Medicine Evidence-Based Database

[0080] The first step in building a TCM evidence-based database is to collect and organize relevant classical TCM literature, modern TCM research findings, and clinical trial data. Classical TCM literature can include ancient medical books such as the *Huangdi Neijing* and *Shanghan Lun*, as well as research papers from modern TCM journals and databases. To ensure the comprehensiveness and accuracy of the data, researchers need to conduct a systematic literature review and screening to extract valid evidence relevant to TCM diagnosis and treatment. Simultaneously, clinical trial data should come from ethically reviewed studies to ensure their scientific validity and reliability. Finally, all collected data will be organized into a structured database for easy retrieval and analysis.

[0081] Step 202: Perform semantic analysis using natural language processing techniques.

[0082] After obtaining the patient's basic information, traditional Chinese medicine (TCM) diagnostic data, and modern medical examination data, Natural Language Processing (NLP) techniques are used to perform semantic analysis on this data. First, text preprocessing techniques are employed to clean the input data, including noise removal, word segmentation, and part-of-speech tagging. Then, Named Entity Recognition (NER) and relation extraction techniques are applied to extract key symptoms, signs, and related medical terms from the patient information. By constructing a semantic model, the extracted information is matched with data in a TCM evidence-based database to identify evidence data related to the patient's symptoms. This process not only improves the efficiency of data processing but also enhances the accuracy of information extraction.

[0083] Step 203: Evidence screening based on Bayesian networks

[0084] After semantic analysis, the next step is to screen the TCM evidence database using a Bayesian network. A Bayesian network is a graphical model used to represent conditional dependencies between variables, effectively handling uncertainty. In this stage, a Bayesian network model containing patient symptoms, diagnostic data, and relevant evidence is first constructed. By inputting the extracted evidence data, the posterior probability between each piece of evidence and the patient's symptoms is calculated. Based on established screening criteria, including evidence level (e.g., randomized controlled trials, observational studies), applicability (whether the evidence is suitable for the patient's specific condition), and safety (whether the evidence has potential side effects), the most relevant evidence-based evidence to the patient's symptoms is selected. This process ensures the scientific validity and clinical applicability of the selected evidence, providing a solid basis for subsequent TCM diagnosis and treatment decisions.

[0085] Preferably, the symptom identification model is trained using a CNN-LSTM combined network model.

[0086] Optionally, step 300 in this embodiment includes:

[0087] Step 301: Construct a CNN-LSTM combined network model

[0088] When constructing a syndrome identification model, the first step is to design a combined model that integrates Convolutional Neural Networks (CNNs) and Long Short-Term Memory Networks (LSTMs). The CNN part is primarily used to extract local features from TCM diagnostic data, such as image features of the tongue and pulse, as well as key symptoms and signs from textual data. Through multi-layer convolution and pooling operations, CNNs can effectively capture the spatial features of the input data. The LSTM part is used to process time-series data, capable of memorizing and analyzing patients' historical diagnostic information and treatment responses. Combining CNNs and LSTMs fully leverages the advantages of both, enhancing the model's ability to identify complex TCM syndromes.

[0089] Step 302: Data Preprocessing and Feature Extraction

[0090] Before model training, the TCM diagnostic data and evidence-based data need to be preprocessed and features extracted. First, the TCM diagnostic data is standardized, including converting tongue and pulse images to a uniform size and format, and performing data augmentation to improve the model's generalization ability. Simultaneously, the text data undergoes word segmentation, stop word removal, and word embedding to convert symptoms and signs into vector representations. For the evidence-based evidence, patient-related features are extracted and integrated with the TCM diagnostic data to form an input dataset containing multiple features. Finally, a feature matrix containing the TCM diagnostic data and evidence-based evidence is constructed for use in model training.

[0091] Step 303: Model Training

[0092] After data preprocessing, the constructed CNN-LSTM combined network model is trained. First, the preprocessed feature matrix is ​​divided into training and validation sets, typically 80% for training and 20% for validation. The cross-entropy loss function is used as the optimization objective, and the Adam optimizer is employed for model training. Through multiple iterations, the model parameters are gradually adjusted to minimize the loss function. After each training epoch, the model's performance is evaluated using the validation set, monitoring accuracy and loss values ​​to prevent overfitting. During training, early stopping can be used to determine the optimal training epochs, ensuring the model achieves its best performance on the validation set.

[0093] Step 304: Syndrome Identification and Result Output

[0094] After training, TCM diagnostic data and evidence-based data are input into the trained CNN-LSTM model for syndrome identification. The model extracts features from the input data through forward propagation and processes time-series information through LSTM layers, ultimately outputting syndrome classification results. The output includes the probability distribution of each syndrome, allowing doctors to determine the patient's primary syndrome type. To improve the model's interpretability, methods such as SHAP (SHapley Additive ex Planations) can be used to analyze the model's contribution to each input feature, helping doctors understand the model's decision-making process. Ultimately, the syndrome identification results will provide crucial information for generating subsequent personalized treatment plans.

[0095] Preferably, the personalized treatment plan includes: a traditional Chinese medicine prescription, an acupuncture plan, and dietary recommendations.

[0096] Specifically, step 400 in this embodiment includes:

[0097] Step 401: Analyze the results of syndrome identification

[0098] Before generating a personalized treatment plan, a detailed analysis of the syndrome identification results is necessary. Based on the syndrome classification results output by the model, doctors will identify the patient's main syndrome type and related symptoms. At this point, combining classical Chinese medicine theories, doctors can refer to classic texts such as the *Huangdi Neijing* and *Shanghan Lun* to understand the etiology, pathogenesis, and corresponding treatment principles of different syndromes. This process not only helps doctors confirm the nature of the syndrome but also provides a theoretical basis for subsequent treatment plans, ensuring the scientific validity and rationality of the treatment plan.

[0099] Step 402: Combining evidence-based approaches

[0100] Based on the analysis of the syndrome, doctors will further refine the treatment plan by incorporating previously selected evidence-based data. This evidence includes modern TCM research findings and clinical trial data, providing empirical support for herbal prescriptions, acupuncture treatments, and dietary adjustments. Doctors will select appropriate herbal medicines and acupuncture points based on the patient's specific condition (such as age, gender, and constitution) and the characteristics of their syndrome. For example, for the syndrome of "liver qi stagnation," doctors might choose liver-soothing and qi-regulating herbs such as Bupleurum and Cyperus, combined with corresponding acupuncture points such as Taichong and Ganshu, to achieve better therapeutic effects.

[0101] Step 403: Develop a personalized treatment plan

[0102] After comprehensively considering the syndrome identification results and evidence-based medicine, the doctor will develop a personalized treatment plan. This plan typically includes three main parts: a Chinese herbal medicine prescription, an acupuncture program, and dietary recommendations. The Chinese herbal medicine prescription will select appropriate herbs and dosages based on the patient's specific syndrome and constitution, ensuring the efficacy and safety of the medication. The acupuncture program will select appropriate acupuncture points and treatment frequency based on the syndrome type and the patient's specific condition. The dietary recommendations will incorporate traditional Chinese medicine dietary therapy theories to provide a suitable dietary plan for the patient, helping them to manage their health in daily life.

[0103] Step 404: Treatment plan recording and feedback mechanism

[0104] Finally, the doctor records the personalized treatment plan in the patient's electronic medical record for follow-up and adjustment. Simultaneously, a feedback mechanism is established to encourage patients to regularly report their subjective feelings and objective changes during treatment. Through follow-up and regular assessments, doctors can dynamically adjust the treatment plan based on patient feedback and treatment response, ensuring the continued effectiveness of the treatment. This personalized treatment plan, based on syndrome differentiation and evidence-based medicine, not only improves the targeting and effectiveness of TCM treatment but also enhances patient participation and satisfaction.

[0105] Optionally, step 500 in this embodiment includes:

[0106] Step 501: Design the accompanying volume

[0107] When collecting subjective feedback from patients, the first step is to design a structured follow-up questionnaire. The questionnaire should include questions across multiple dimensions, such as symptom improvement, treatment satisfaction, and quality of life. To ensure the validity and reliability of the questionnaire, existing standardized scales, such as the Visual Analogue Scale (VAS) or the SF-36 Health Survey, can be used as a reference. The questionnaire should use concise and clear language, avoiding technical jargon, so that patients can easily understand and answer accurately. The questionnaire can be distributed in paper or electronic form (such as a mobile application or online survey platform) to ensure convenient completion by patients during treatment.

[0108] Step 502: Updating Electronic Medical Records

[0109] During a patient's visit, doctors need to update the electronic medical record (EMR) in real time, recording the patient's subjective feedback and objective data. At each follow-up visit, doctors should inquire about changes in the patient's symptoms, treatment response, and any discomfort, and promptly input this information into the EMR system. The EMR system should have data integration and analysis capabilities, automatically generating the patient's treatment history and trend charts to help doctors better assess treatment effectiveness. Furthermore, the system should support secure data storage and privacy protection, ensuring the security and confidentiality of patient information.

[0110] Step 503: Application of Wearable Devices

[0111] To enable real-time monitoring of patient treatment responses, wearable devices (such as smartwatches and health monitors) can be introduced. These devices can collect patients' physiological data in real time, such as heart rate, blood pressure, sleep quality, and activity levels. Patients wear these devices during treatment, and the data is automatically uploaded to the cloud or the hospital's information system. Doctors can then use data analysis platforms to view changes in patients' physiological indicators in real time, promptly identifying potential health problems or treatment reactions. This real-time monitoring not only improves the efficiency of data collection but also enhances doctors' understanding of patients' health conditions.

[0112] Step 504: Data Integration and Analysis

[0113] After collecting patient feedback, electronic medical records, and wearable device data, the next step is to integrate and analyze this data. Data analytics tools aggregate data from different sources onto a unified platform, creating a comprehensive health record for each patient. Using statistical analysis methods, doctors can assess treatment effectiveness, identify key factors influencing treatment outcomes, and dynamically adjust treatment plans based on patient feedback and physiological data. Furthermore, regularly generated reports help doctors and patients jointly assess treatment progress, promoting patient engagement and treatment adherence. This comprehensive data collection and analysis approach ensures the scientific rigor and personalization of the treatment process.

[0114] Preferably, based on statistical analysis methods, the efficacy is quantified according to the subjective feedback, the objective data, and the treatment response to obtain a quantified efficacy value. The quantified efficacy value is then compared with the personalized treatment plan to obtain a comparison result, including:

[0115] Define quantitative indicators of therapeutic efficacy; these quantitative indicators include therapeutic efficacy scores and rates of change.

[0116] The subjective feedback and objective data are used to score the treatment effect;

[0117] The efficacy score is calculated based on the subjective feedback score and the objective data score; Efficacy score = (subjective feedback score + objective data score) / 2;

[0118] Calculate the rate of change; Rate of change = (Indicators of objective data after treatment - Indicators of objective data before treatment) / Indicators of objective data before treatment × 100%;

[0119] The value of the efficacy quantification index is determined as the efficacy quantification value;

[0120] The efficacy quantification value is compared with the personalized treatment plan to analyze the relationship between efficacy and treatment plan, and the comparison results are obtained; the comparison results include correlation analysis results and efficacy difference assessment results.

[0121] Specifically, step 600 in this embodiment includes:

[0122] Step 601: Define quantitative indicators of therapeutic effect

[0123] Before quantifying treatment efficacy, it is essential to first define the indicators used for efficacy quantification. These indicators typically include efficacy scores and rates of change. Efficacy scores are a comprehensive indicator that assesses treatment effectiveness by integrating subjective patient feedback with objective data, while rates of change reflect the degree of change in the patient's health status before and after treatment. By setting these indicators, physicians can more systematically evaluate treatment effectiveness and provide quantitative evidence for subsequent analysis.

[0124] Step 602: Calculation of Treatment Efficacy Score

[0125] After collecting patients' subjective feedback and objective data, the next step is to score the treatment effectiveness based on this data. Subjective feedback scores can be quantified using questions in a questionnaire, such as a 1-10 rating system, where patients rate their symptom improvement based on their own feelings. Objective data scores are based on medical examination results, such as changes in blood pressure and blood sugar levels. Doctors will calculate both subjective feedback and objective data scores for each patient based on this data, ensuring a standardized and consistent scoring process.

[0126] Step 603: Calculate the efficacy score

[0127] After obtaining subjective feedback scores and objective data scores, doctors will use the following formula to calculate the efficacy score:

[0128] Therapeutic effect score = (subjective feedback score + objective data score)^2 Therapeutic effect score = 2(subjective feedback score + objective data score)

[0129] This formula is designed to comprehensively consider both the patient's subjective feelings and objective medical indicators, ensuring that the efficacy score fully reflects the treatment effect. In this way, doctors can obtain a comprehensive efficacy score as the basis for subsequent analysis.

[0130] Step 604: Calculate the rate of change

[0131] Calculating the rate of change is a crucial step in evaluating treatment effectiveness. The formula for the rate of change is:

[0132] Change rate = (Post-treatment objective data index - Pre-treatment objective data index) × 100% Change rate = (Pre-treatment objective data index - Post-treatment objective data index) × 100%

[0133] This calculation method can quantify the degree of change in a patient's health status before and after treatment, reflecting the actual effect of the treatment. By calculating the rate of change, doctors can intuitively understand the impact of treatment on the patient's health and provide data support for subsequent efficacy analysis.

[0134] Step 605: Determine the quantitative value of therapeutic effect

[0135] After calculating the efficacy score and rate of change, doctors combine these two indicators to determine a quantitative efficacy value. This quantitative efficacy value not only reflects the patient's subjective experience but also incorporates changes in objective data, providing a comprehensive assessment of efficacy. Through this quantitative method, doctors can gain a clearer understanding of treatment effectiveness and provide a basis for adjusting subsequent treatment plans.

[0136] Step 606: Comparison of quantitative efficacy values ​​with personalized treatment plans

[0137] Finally, doctors compare the quantified efficacy values ​​with the personalized treatment plan to analyze the relationship between efficacy and the treatment plan. This comparison process includes correlation analysis and efficacy difference assessment. Correlation analysis uses statistical methods (such as the Pearson correlation coefficient) to assess the relationship between the quantified efficacy values ​​and the various components of the treatment plan (such as traditional Chinese medicine prescriptions, acupuncture procedures, etc.), helping doctors identify which treatment methods have the greatest impact on patient efficacy. Efficacy difference assessment compares the quantified efficacy values ​​of different patients to analyze the differences in the effects of different treatment plans, providing data support for subsequent optimization of personalized treatment plans. This systematic analysis method ensures the scientific validity and effectiveness of the treatment plan, ultimately improving the patient's treatment experience and satisfaction.

[0138] Preferably, the parameters generated by dynamically adjusting the syndrome identification results and personalized treatment plans based on the comparison results include:

[0139] The comparison results were statistically analyzed to calculate the differences in efficacy between different treatment regimens, and the key factors affecting efficacy were identified based on these differences; the key factors included the patient's individual characteristics and symptom changes.

[0140] The adjustment strategy will be determined based on the key factors and different treatment options;

[0141] The adjustment strategy was applied to the syndrome identification model, the model parameters were updated, and the model was retrained using the gradient descent algorithm to adapt to the new parameter settings.

[0142] Based on the updated syndrome identification results, a new personalized treatment plan is generated.

[0143] Optionally, step 700 in this embodiment includes:

[0144] Step 701: Analyze the comparison results

[0145] After comparing the quantified efficacy values ​​with the personalized treatment plans, the doctors and research team will conduct an in-depth analysis of the results. By evaluating the correlation analysis results, the team can identify which treatment methods have significant positive or negative correlations with the quantified efficacy values. This analysis not only helps doctors understand the effectiveness of the current treatment plan but also reveals potential areas for improvement. For example, if a certain traditional Chinese medicine prescription has a high correlation with the quantified efficacy value, the doctor may consider continuing to use the prescription in subsequent treatments; otherwise, adjustments or replacements may be necessary.

[0146] Step 702: Dynamically adjust parameters

[0147] Based on the comparison results, doctors will dynamically adjust the parameters of the syndrome identification model and the rules for generating personalized treatment plans. This process may involve re-evaluating the model's input features, such as adding new symptom features or adjusting the weights of existing features to improve the model's accuracy. Simultaneously, the rules for generating personalized treatment plans may also need to be adjusted based on patient feedback and efficacy metrics, such as changing the dosage of traditional Chinese medicine, adding new treatment methods, or optimizing acupuncture protocols. This dynamic adjustment ensures that treatment plans can respond promptly to changing patient needs, thus improving the level of personalization in treatment.

[0148] Step 703: Establishing a Feedback Mechanism

[0149] To achieve effective dynamic adjustments, establishing a feedback mechanism is crucial. This mechanism should include regular patient follow-ups and data collection to continuously monitor patients' treatment responses and health status. By incorporating patients' subjective feedback and objective data into the model's update process, physicians can better understand patients' individualized needs and adjust the parameters for generating syndrome identification results and treatment plans accordingly. This feedback mechanism not only improves the flexibility and adaptability of treatment but also enhances patient participation and satisfaction, ultimately promoting improved treatment outcomes.

[0150] Preferably, through multiple rounds of iterative learning, the parameters of the syndrome identification results and the personalized treatment plan are continuously adjusted to make the syndrome identification results and the personalized treatment plan more closely match the individualized needs of the patient, including:

[0151] After implementing the new personalized treatment plan, continue to collect subjective feedback and objective data from the target patients, and return to the step "Based on statistical analysis methods, quantify the efficacy according to the subjective feedback, the objective data and the treatment response, obtain the efficacy quantification value, and compare the efficacy quantification value with the personalized treatment plan to obtain the comparison result";

[0152] Specifically, step 800 in this embodiment includes:

[0153] Step 801: Framework for Iterative Learning

[0154] In implementing multi-round iterative learning, this embodiment establishes a systematic framework to collect and analyze data in each iteration cycle. At the start of each iteration, doctors update the input features and parameter settings of the syndrome identification model based on the latest patient feedback and treatment response data. In this way, the model can continuously learn and adapt to new data, thereby improving its responsiveness to individualized patient needs. The iterative learning framework should include four key components: data collection, model updating, result evaluation, and feedback mechanisms, ensuring that each component is effectively integrated.

[0155] Step 802: Model Self-Optimization

[0156] In each iteration, the syndrome identification model optimizes itself using newly collected data. By applying machine learning algorithms (such as random forests and support vector machines), the model identifies key factors influencing syndrome identification results and adjusts its parameter settings accordingly. For example, if certain symptoms show stronger predictive power in new data, the model automatically increases the weight of these features, thereby improving overall identification accuracy. Simultaneously, the rules for generating personalized treatment plans are adjusted based on the model's optimization results to ensure that the treatment plan better meets the patient's actual needs.

[0157] Step 803: Continuous Evaluation and Improvement

[0158] After each iteration, the doctors and research team evaluate the model's performance, analyzing its accuracy and effectiveness on new data. This evaluation process includes not only checking the accuracy of syndrome identification results but also assessing the efficacy of personalized treatment plans. By comparing the results of different iteration cycles, the team can identify areas for improvement and shortcomings in the model and develop corresponding improvement measures. This continuous evaluation and improvement process ensures that the model and treatment plans can continuously adapt to the individualized needs of patients, ultimately achieving higher treatment outcomes and patient satisfaction.

[0159] Corresponding to the above methods, such as Figure 2 As shown, this embodiment also provides an evidence-based TCM diagnosis and treatment decision-making system, including:

[0160] The data acquisition unit is used to acquire basic information, traditional Chinese medicine diagnostic data, and modern medical examination data of the target patient.

[0161] The evidence acquisition unit is used to screen evidence based on the constructed TCM evidence-based evidence database, according to the basic information, the TCM diagnostic data and the modern medical examination data, to obtain evidence-based evidence.

[0162] The result acquisition unit is used to input the TCM diagnostic data and the evidence-based evidence into the constructed syndrome identification model to classify TCM syndromes and obtain syndrome identification results.

[0163] The treatment plan generation unit is used to generate personalized treatment plans based on the syndrome identification results, combined with classical TCM theories and evidence-based medicine.

[0164] The data acquisition unit is used to collect patients' subjective feedback and objective data, and to monitor patients' treatment response in real time through follow-up questionnaires, electronic medical records and wearable devices;

[0165] The comparison unit is used to quantify the efficacy based on the subjective feedback, the objective data, and the treatment response using statistical analysis methods, obtain a quantitative efficacy value, and compare the quantitative efficacy value with the personalized treatment plan to obtain a comparison result.

[0166] The feedback unit is used to dynamically adjust the parameters generated by the syndrome identification results and the personalized treatment plan based on the comparison results.

[0167] The iterative learning unit is used to continuously adjust the parameters generated by the syndrome identification results and the personalized treatment plan through multiple rounds of iterative learning, so that the syndrome identification results and the personalized treatment plan are closer to the individualized needs of the patient.

[0168] The beneficial effects of this invention are as follows:

[0169] (1) This invention enhances the scientificity and accuracy of diagnosis and treatment decisions by combining classical Chinese medicine theories with modern medical data.

[0170] (2) The present invention generates personalized treatment plans based on the patient’s specific condition, which improves the pertinence and effectiveness of treatment.

[0171] (3) This invention continuously optimizes the diagnosis and treatment plan through feedback mechanism and multiple rounds of iterative learning to ensure that it adapts to the changing needs of patients.

[0172] (4) This invention utilizes statistical analysis to enhance the interpretability of the model, help doctors understand the decision-making process, and improve clinical confidence.

[0173] (5) The present invention can adapt to individual differences among different patients, and can still maintain a high accuracy rate, especially in small sample learning scenarios.

[0174] The various embodiments in this specification are described in a progressive manner, with each embodiment focusing on its differences from other embodiments. Similar or identical parts between embodiments can be referred to interchangeably. For the systems disclosed in the embodiments, since they correspond to the methods disclosed in the embodiments, the descriptions are relatively simple; relevant parts can be referred to the method section.

[0175] This document uses specific examples to illustrate the principles and implementation methods of the present invention. The descriptions of the above embodiments are only for the purpose of helping to understand the method and core ideas of the present invention. Furthermore, those skilled in the art will recognize that, based on the ideas of the present invention, there will be changes in the specific implementation methods and application scope. Therefore, the content of this specification should not be construed as a limitation of the present invention.

Claims

1. A traditional Chinese medicine diagnosis and treatment decision-making method based on evidence-based evidence, characterized in that, The method comprises the following steps: acquiring basic information, TCM diagnosis data and modern medical examination data of a target patient; performing evidence screening based on the constructed TCM evidence-based evidence database according to the basic information, the TCM diagnosis data and the modern medical examination data to obtain evidence-based evidence; inputting the TCM diagnosis data and the evidence-based evidence into a constructed syndrome identification model to classify TCM syndromes and obtain a syndrome identification result; generating an individualized treatment plan according to the syndrome identification result in combination with TCM classic theories and evidence-based evidence; collecting subjective feedback and objective data of the patient and monitoring treatment response of the patient in real time through questionnaires, electronic medical records and wearable devices; quantifying the treatment effect according to the subjective feedback, the objective data and the treatment response based on a statistical analysis method to obtain a treatment effect quantization value, and comparing the treatment effect quantization value with the individualized treatment plan to obtain a comparison result; dynamically adjusting parameters generated by the syndrome identification result and the individualized treatment plan according to the comparison result; continuously adjusting the parameters generated by the syndrome identification result and the individualized treatment plan through multiple rounds of iterative learning to make the syndrome identification result and the individualized treatment plan closer to individualized needs of the patient; quantifying the treatment effect according to the subjective feedback, the objective data and the treatment response based on a statistical analysis method to obtain a treatment effect quantization value, and comparing the treatment effect quantization value with the individualized treatment plan to obtain a comparison result, comprising: defining a treatment effect quantization index; the treatment effect quantization index comprises a treatment effect score and a change rate; scoring the subjective feedback and the objective data on treatment effect; calculating the treatment effect score according to a score value of the subjective feedback and a score value of the objective data; treatment effect score = (score value of subjective feedback + score value of objective data) / 2; calculating the change rate; change rate = (index of objective data after treatment - index of objective data before treatment) / index of objective data before treatment × 100%; determining a numerical value of the treatment effect quantization index as the treatment effect quantization value; comparing the treatment effect quantization value with the individualized treatment plan to analyze the relationship between treatment effect and treatment plan and obtain the comparison result; the comparison result comprises a correlation analysis result and a treatment effect difference evaluation result; dynamically adjusting parameters generated by the syndrome identification result and the individualized treatment plan according to the comparison result, comprising: performing statistical analysis on the comparison result to calculate the effect difference of different treatment plans and determine key factors affecting treatment effect according to the effect difference; the key factors comprise individual characteristics and symptom changes of the patient; determining an adjustment strategy according to the key factors and different treatment plans; applying the adjustment strategy to the syndrome identification model, updating model parameters and retraining the model using a gradient descent algorithm to adapt to new parameter settings; regenerating an individualized treatment plan according to the updated syndrome identification result; continuously adjusting the parameters generated by the syndrome identification result and the individualized treatment plan through multiple rounds of iterative learning to make the syndrome identification result and the individualized treatment plan closer to individualized needs of the patient, comprising: After the new individualized treatment plan is implemented, the subjective feedback and objective data of the target patient are continuously collected, and the step of "quantifying the efficacy based on statistical analysis method according to the subjective feedback, the objective data and the treatment reaction to obtain an efficacy quantification value, and comparing the efficacy quantification value with the individualized treatment plan to obtain a comparison result" is returned; A closed-loop control process is formed through the feedback mechanism to continuously monitor the efficacy and dynamically adjust.

2. The evidence-based traditional Chinese medicine diagnosis and treatment decision method according to claim 1, characterized in that, The basic information includes age, gender, medical history and living habits; the TCM diagnosis data includes chief complaint, symptoms, tongue and pulse; and the modern medical examination data includes blood test results and imaging data.

3. The evidence-based traditional Chinese medicine diagnosis and treatment decision method according to claim 1, characterized in that, Based on the constructed TCM evidence-based evidence database, evidence screening is performed according to the basic information, the TCM diagnosis data and the modern medical examination data to obtain evidence-based evidence, including: A TCM evidence-based evidence database is constructed; the TCM evidence-based evidence database includes TCM classic literature, modern TCM research results and clinical trial data; The semantic analysis of the basic information, the TCM diagnosis data and the modern medical examination data is performed using natural language processing technology to extract evidence data related to the patient's symptoms; Based on the Bayesian network, the most relevant evidence-based evidence is screened out from the TCM evidence-based evidence database; the screening criteria of the evidence-based evidence include evidence level, applicability and safety.

4. The evidence-based traditional Chinese medicine diagnosis and treatment decision method according to claim 1, characterized in that, The syndrome recognition model is obtained by training a CNN-LSTM combined network model.

5. The evidence-based traditional Chinese medicine diagnosis and treatment decision method according to claim 1, characterized in that, The individualized treatment plan includes a traditional Chinese medicine prescription, an acupuncture plan and diet conditioning suggestions.

6. A traditional Chinese medicine diagnosis and treatment decision system based on evidence-based evidence, characterized in that, It includes: A data acquisition unit is configured to acquire basic information, TCM diagnosis data and modern medical examination data of a target patient; An evidence acquisition unit is configured to perform evidence screening based on a constructed TCM evidence-based evidence database according to the basic information, the TCM diagnosis data and the modern medical examination data to obtain evidence-based evidence; A result acquisition unit is configured to input the TCM diagnosis data and the evidence-based evidence into a constructed syndrome recognition model to classify TCM syndromes and obtain a syndrome recognition result; A plan generation unit is configured to generate an individualized treatment plan according to the syndrome recognition result in combination with TCM classic theory and evidence-based evidence; A data acquisition unit is configured to collect subjective feedback and objective data of a patient and monitor treatment reaction of the patient in real time through a random access questionnaire, an electronic medical record and a wearable device; A comparison unit is configured to quantify the efficacy based on statistical analysis method according to the subjective feedback, the objective data and the treatment reaction to obtain an efficacy quantification value, and compare the efficacy quantification value with the individualized treatment plan to obtain a comparison result; A feedback unit is configured to dynamically adjust parameters of the syndrome recognition result and the individualized treatment plan generated according to the comparison result; An iterative learning unit is configured to continuously adjust the parameters of the syndrome recognition result and the individualized treatment plan generated through multiple rounds of iterative learning, so that the syndrome recognition result and the individualized treatment plan are closer to individualized needs of the patient; Based on the statistical analysis method, the therapeutic effect is quantified according to the subjective feedback, the objective data and the treatment response, and a therapeutic effect quantification value is obtained, and the therapeutic effect quantification value is compared with the individualized treatment plan to obtain a comparison result, including: Defining a therapeutic effect quantification index; the therapeutic effect quantification index includes a therapeutic effect score and a change rate; Scoring the therapeutic effect of the subjective feedback and the objective data; The therapeutic effect score is calculated according to the score value of the subjective feedback and the score value of the objective data; therapeutic effect score=(subjective feedback score value+objective data score value) / 2; The change rate is calculated; change rate=(objective data index after treatment-objective data index before treatment) / objective data index before treatment×100%; The value of the therapeutic effect quantification index is determined as the therapeutic effect quantification value; The therapeutic effect quantification value is compared with the individualized treatment plan to analyze the relationship between the therapeutic effect and the treatment plan, and the comparison result is obtained; the comparison result includes a correlation analysis result and a therapeutic effect difference evaluation result; According to the comparison result, the parameters generated by the syndrome identification result and the individualized treatment plan are dynamically adjusted, including: Statistical analysis is performed on the comparison result to calculate the effect difference of different treatment plans, and the key factors affecting the therapeutic effect are determined according to the effect difference; the key factors include the individual characteristics of the patient and the symptom changes; According to the key factors and different treatment plans, an adjustment strategy is determined; The adjustment strategy is applied to the syndrome identification model, the model parameters are updated, and the gradient descent algorithm is used to retrain the model to adapt to the new parameter settings; According to the updated syndrome identification result, the individualized treatment plan is regenerated; Through multiple rounds of iterative learning, the parameters generated by the syndrome identification result and the individualized treatment plan are continuously adjusted to make the syndrome identification result and the individualized treatment plan more close to the individual needs of the patient, including: After the implementation of the new individualized treatment plan, the subjective feedback and objective data of the target patient are continuously collected, and the step "based on the statistical analysis method, the therapeutic effect is quantified according to the subjective feedback, the objective data and the treatment response, and a therapeutic effect quantification value is obtained, and the therapeutic effect quantification value is compared with the individualized treatment plan to obtain a comparison result" is returned; A closed-loop control process is formed through a feedback mechanism to continuously monitor the therapeutic effect and dynamically adjust.

Images