Intelligent diagnosis and treatment management method and system for chronic respiratory disease based on large model

Abstract

The invention relates to the technical field of data processing, and discloses an intelligent diagnosis and treatment management method and system for chronic respiratory diseases based on a large model. The method comprises the following steps: performing multi-dimensional data acquisition processing on a chronic respiratory disease patient to form a structured data set; constructing a large model and a knowledge graph by using a dual coding framework; performing risk hierarchical assessment based on the large model; generating a personalized diagnosis and treatment scheme according to an evaluation result; constructing a medical collaborative framework to execute curative effect monitoring; and customizing health education content and continuously optimizing the system. According to the application, accurate layering, personalized treatment scheme making, multi-stage medical collaborative management and dynamic curative effect monitoring of the chronic respiratory disease patients are realized, so that the defects of insufficient diagnosis and treatment accuracy, low resource allocation efficiency and discontinuous patient management in the prior art are overcome, and the overall efficiency of chronic respiratory disease diagnosis and treatment management is improved.

Description

Technical Field

[0001] The present application relates to the field of data processing technology, and in particular to a method and system for intelligent diagnosis and treatment management of chronic respiratory diseases based on a large model. Background Art

[0002] Respiratory chronic diseases include chronic obstructive pulmonary disease (COPD), asthma, pulmonary fibrosis, etc., which have the characteristics of long course, repeated attacks, and long-term management, and have become a major public health problem in the world. Traditional management of chronic respiratory diseases mainly relies on regular outpatient follow-up and patient self-reporting. The uneven distribution of medical resources has led to a significant gap in diagnosis and treatment capabilities between primary medical institutions and specialized hospitals. Existing technologies attempt to improve the monitoring and management of chronic respiratory diseases through telemedicine and wearable devices, such as remote monitoring systems of physiological parameters based on the Internet of Things, medication reminder systems with simple rule engines, and disease risk prediction tools based on statistical models. Some studies have also explored patient education platforms and primary artificial intelligence-assisted diagnosis systems based on mobile Internet, trying to improve disease management efficiency and patient compliance.

[0003] However, existing technologies have obvious deficiencies in practical applications. First, the traditional respiratory chronic disease management model lacks accuracy and personalization, and cannot dynamically adjust the treatment plan according to the individual patient's condition. Secondly, although the existing remote monitoring system can collect data, it lacks effective data analysis and knowledge mining capabilities, and it is difficult to extract valuable clinical decision-making information from massive multimodal medical data. Furthermore, the problem of information islands between medical institutions is serious, and tiered diagnosis and treatment is difficult to implement in a substantial manner, resulting in the inefficient allocation of high-quality medical resources. In addition, existing patient education tools generally adopt a "one-size-fits-all" approach, ignoring the differences in patients' cognitive levels and acceptance abilities, and the educational effect is limited. Most importantly, existing technologies lack an effective way to integrate medical expertise with advanced artificial intelligence technology, and cannot achieve intelligent decision support driven by medical knowledge. Summary of the invention

[0004] The present application provides a method and system for intelligent diagnosis and treatment management of chronic respiratory diseases based on a large model, which is used to achieve accurate stratification of patients with chronic respiratory diseases, formulation of personalized treatment plans, multi-level medical collaborative management, and dynamic efficacy monitoring, thereby overcoming the defects of insufficient diagnosis and treatment accuracy, inefficient resource allocation, and discontinuous patient management in the prior art, and improving the overall effectiveness of diagnosis and treatment management of chronic respiratory diseases.

[0005] In a first aspect, the present application provides a method for intelligent diagnosis and treatment management of respiratory chronic diseases based on a large model, and the method for intelligent diagnosis and treatment management of respiratory chronic diseases based on a large model includes: collecting and processing multi-dimensional data of patients with respiratory chronic diseases to obtain a structured patient data set; inputting the structured patient data set and the professional knowledge base of respiratory diseases into a dual encoding pre-training and contrastive learning fine-tuning framework to obtain a large model of respiratory chronic diseases and a knowledge graph of respiratory chronic diseases; based on the large model of respiratory chronic diseases and the knowledge graph of respiratory chronic diseases, semantic analysis and multi-dimensional calculation are performed on patient symptom descriptions and answers to evaluation scales to generate patient risk stratification assessment results; based on the patient risk stratification assessment results, the large model of respiratory chronic diseases is used to retrieve clinical guideline rules and combine individual characteristics of patients to generate personalized diagnosis and treatment plans and matching scores; based on the personalized diagnosis and treatment plans and the matching scores, a multi-level medical institution collaborative framework is constructed and dynamic efficacy monitoring is performed to generate efficacy indicators and treatment response indexes; based on the efficacy indicators and the treatment response indexes, patient health education content and self-management tools are customized, and system operation data is collected to update the model and optimize the process.

[0006] In a second aspect, the present application provides a large model-based intelligent diagnosis and treatment management system for chronic respiratory diseases, which includes: The acquisition module is used to collect and process multi-dimensional data of patients with chronic respiratory diseases to obtain a structured patient data set; An input module, used to input the structured patient dataset and the professional knowledge base of respiratory diseases into a dual encoding pre-training and contrastive learning fine-tuning framework to obtain a large respiratory chronic disease model and a respiratory chronic disease knowledge graph; A calculation module, for performing semantic analysis and multi-dimensional calculation on the patient's symptom description and the evaluation scale answer based on the respiratory chronic disease big model and the respiratory chronic disease knowledge graph, to generate a patient risk stratification evaluation result; A generation module is used to generate a personalized diagnosis and treatment plan and a matching score based on the patient risk stratification assessment results, using the respiratory chronic disease big model to retrieve clinical guideline rules and combining the patient's individual characteristics; A construction module is used to construct a multi-level medical institution collaboration framework and perform dynamic efficacy monitoring based on the personalized diagnosis and treatment plan and the matching score, and generate efficacy indicators and treatment response indexes; The customization module is used to customize patient health education content and self-management tools according to the efficacy indicators and the treatment response index, and collect system operation data for model updating and process optimization.

[0007] In the third aspect, a big model-based intelligent diagnosis and treatment management device for respiratory chronic diseases is provided, comprising: a memory and at least one processor, the memory storing instructions; the at least one processor calls the instructions in the memory to enable the big model-based intelligent diagnosis and treatment management device for respiratory chronic diseases to execute the above-mentioned big model-based intelligent diagnosis and treatment management method for respiratory chronic diseases.

[0008] In a fourth aspect, a computer-readable storage medium is provided, wherein instructions are stored in the computer-readable storage medium, and when the computer-readable storage medium is run on a computer, the computer executes the above-mentioned large-model-based intelligent diagnosis and treatment management method for chronic respiratory diseases.

[0009] In the technical solution provided in this application, a structured patient data set is formed by multi-dimensional data collection and processing, thereby realizing comprehensive acquisition and standardized processing of information related to respiratory chronic diseases; secondly, a dual encoding pre-training and contrastive learning fine-tuning framework is used to process structured patient data sets and a professional knowledge base of respiratory diseases. The generated large model of respiratory chronic diseases and knowledge graph of respiratory chronic diseases have deep medical understanding and professional reasoning capabilities, which are significantly better than traditional machine learning models, and can understand complex medical contexts and extract key clinical features from them; thirdly, semantic analysis and multi-dimensional calculation are performed on patient symptom descriptions and evaluation scale answers based on the large model of respiratory chronic diseases and the knowledge graph, and the generated patient risk stratification assessment results are highly accurate, can capture subtle differences in disease characteristics, and achieve accurate stratification, thereby optimizing the allocation of medical resources; fourthly According to the results of patient risk stratification assessment, the respiratory chronic disease big model is used to retrieve clinical guideline rules and generate personalized diagnosis and treatment plans and matching scores based on the individual characteristics of patients, which fully considers the individual differences of patients, so that the treatment plan is both in line with the principles of evidence-based medicine and highly personalized, thereby improving the accuracy and compliance of treatment; Fifth, the multi-level medical institution collaboration framework and dynamic efficacy monitoring system constructed based on personalized diagnosis and treatment plans and matching scores have broken the medical information island, realized the rational allocation of high-quality medical resources and continuous medical services, and reduced repeated examinations and treatment errors; patient health education content and self-management tools customized according to efficacy indicators and treatment response indexes, combined with the continuous collection of system operation data, model updates and process optimization, have formed a closed-loop quality improvement system, enabling the system to continuously evolve and improve. It is particularly worth emphasizing that the large model algorithm used in this solution has made great contributions to the field of diagnosis, treatment and management of chronic respiratory diseases. Its contextual understanding ability enables the model to extract key clinical information from unstructured medical texts, and the knowledge-enhanced reasoning mechanism realizes the automated support of complex medical decision-making. The multimodal data fusion algorithm can integrate patient information from different sources to form a comprehensive view, and the adaptive learning characteristics enable the system to continuously optimize diagnosis and treatment strategies based on new data. These specific algorithm features jointly build a highly specialized intelligent decision-making support system for respiratory diseases, which greatly improves the efficiency and accuracy of chronic respiratory disease management. BRIEF DESCRIPTION OF THE DRAWINGS

[0010] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other accompanying drawings can be obtained based on these accompanying drawings without paying creative work.

[0011] Figure 1 This is a schematic diagram of an embodiment of a method for intelligent diagnosis and treatment of chronic respiratory diseases based on a large model in an embodiment of the present application; Figure 2 This is a schematic diagram of an embodiment of an intelligent diagnosis and treatment management system for chronic respiratory diseases based on a large model in an embodiment of the present application; Figure 3 It is a schematic block diagram of the structure of a large model-based intelligent diagnosis and treatment management device for chronic respiratory diseases in an embodiment of the present invention. DETAILED DESCRIPTION

[0012] The embodiments of the present application provide a method and system for intelligent diagnosis and treatment management of chronic respiratory diseases based on a large model. The terms "first", "second", "third", "fourth", etc. (if any) in the specification and claims of this application and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchangeable where appropriate, so that the embodiments described here can be implemented in an order other than that illustrated or described here. In addition, the terms "including" or "having" and any variations thereof are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device that includes a series of steps or units is not necessarily limited to those steps or units that are clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0013] For ease of understanding, the specific process of the embodiment of the present application is described below. Figure 1 In the embodiment of the present application, an embodiment of the intelligent diagnosis and treatment management method for chronic respiratory diseases based on a large model includes: Step S101, collecting and processing multi-dimensional data of patients with chronic respiratory diseases to obtain a structured patient data set; Step S102: input the structured patient data set and the professional knowledge base of respiratory diseases into the dual encoding pre-training and contrastive learning fine-tuning framework to obtain a large model of respiratory chronic diseases and a knowledge graph of respiratory chronic diseases; Step S103: Based on the respiratory chronic disease big model and the respiratory chronic disease knowledge graph, semantic analysis and multi-dimensional calculation are performed on the patient's symptom description and the evaluation scale answer to generate the patient risk stratification evaluation result; Step S104: Based on the patient risk stratification assessment results, the respiratory chronic disease big model is used to retrieve clinical guideline rules and combine the patient's individual characteristics to generate a personalized diagnosis and treatment plan and matching score; Step S105: Based on the personalized diagnosis and treatment plan and the matching score, a multi-level medical institution collaboration framework is constructed and dynamic efficacy monitoring is performed to generate efficacy indicators and treatment response indexes; Step S106: Customize patient health education content and self-management tools based on efficacy indicators and treatment response indexes, and collect system operation data for model updating and process optimization.

[0014] It is understandable that the execution subject of the present application can be a large model-based intelligent diagnosis and treatment management system for chronic respiratory diseases, or a terminal or a server, which is not limited here. The embodiment of the present application is described by taking the server as the execution subject as an example.

[0015] Specifically, starting with the multidimensional data collection of patients with chronic respiratory diseases, data is collected through three main channels: clinical entry in medical institutions, self-reporting by patients on mobile terminals, and automatic collection by smart wearable devices. These data include basic information of patients, medical history, medication status, symptom assessment, and physiological monitoring data. The collected raw data are heterogeneous and need to be preprocessed. The structured data are standardized and normalized, the text data are segmented and entities are standardized, and the time series data are denoised and outlier processed. Subsequently, a unique patient identifier is established, and all data are organized into a tree-like hierarchical structure. After integrity check and consistency verification, a structured patient data set is formed.

[0016] The structured patient dataset and the professional knowledge base of respiratory diseases (including authoritative clinical guidelines and evidence-based medical literature) are input into the dual encoding pre-training and contrastive learning fine-tuning framework. The framework uses the Transformer encoder structure to pre-train large-scale medical corpora through masked language models and next sentence prediction tasks, and then uses contrastive learning methods for fine-tuning to bring the representation vectors of similar cases closer and the representation vectors of different cases farther apart, thereby obtaining a large model of chronic respiratory diseases. At the same time, through entity extraction and relationship extraction, medical concepts such as diseases, symptoms, and drugs and their associations are identified to form knowledge triples, and a knowledge graph of chronic respiratory diseases is constructed through a graph embedding algorithm. Based on these two core components, the patient's symptom description and the evaluation scale answer are intelligently analyzed. The system obtains the patient's symptom description through a natural language interactive interface, extracts key symptom attributes and forms a symptom entity matrix; at the same time, it guides the patient to complete a structured symptom assessment questionnaire to obtain a questionnaire numerical vector. These two parts of data are combined with basic physiological monitoring data to construct a multidimensional feature space. Weight coefficients are assigned to core symptoms, and the symptom burden index is obtained through a weighted calculation method. A time series analysis is performed on symptom fluctuation trends, acute attack risks, etc. to generate a disease stability index. The comprehensive risk score is calculated based on the severity classification of the underlying disease and the presence of comorbidities, and patients are divided into three levels: low, medium, and high risk.

[0017] Subsequently, according to the risk stratification assessment results, the respiratory chronic disease large model was used to retrieve the corresponding clinical guidelines. For patients with chronic obstructive pulmonary disease, they were divided into four groups of A / B / C / D according to the degree of airflow limitation and symptom score; for asthma patients, they were marked based on the control level and severity. Then, different medication regimens were matched according to the disease classification results to form a basic treatment framework. This framework was compared and analyzed with the patient's previous treatment records and history of adverse drug reactions, and the priority of drug options was adjusted to form a personalized drug regimen, and detailed medication information and non-drug treatment recommendations were supplemented. Finally, by calculating the matching scores of the patient's previous medication reactions, living habits, compliance factors, economic affordability, etc., a percentage matching score was obtained.

[0018] According to the patient's risk level and matching score, a multi-level medical institution collaboration framework is constructed to assign low-risk patients to primary medical institutions, arrange primary management and specialist follow-up for medium-risk patients, and manage high-risk patients by specialist hospitals. Connect medical institutions of different levels through an intelligent collaborative platform to achieve data sharing. For patients treated at the primary level, the system provides large-scale model-assisted decision support. For patients who start treatment, multi-dimensional monitoring is carried out to record changes in symptom scores, changes in physiological indicators, medication compliance data, and quality of life assessment data. The improvement of each dimension is calculated through time series analysis to generate multi-dimensional efficacy indicators, and the comprehensive score of treatment response is calculated through a weighted aggregation algorithm to form a time series treatment response index. According to the efficacy indicators and treatment response index, personalized health education content and self-management tools are customized. According to patient characteristics and efficacy, disease knowledge, symptom identification, medication guidance and other contents are screened and converted into multimedia resources, and the presentation depth is dynamically adjusted according to the patient's understanding. Self-management tools such as electronic medicine box reminder system and symptom monitoring diary are developed and integrated into the multi-channel access platform. At the same time, data such as user behavior, feedback, and treatment effects during the use of the system are recorded and analyzed for updating the large model knowledge base, algorithm tuning, and system process optimization.

[0019] In the embodiment of the present application, a structured patient data set is formed by multi-dimensional data collection and processing, thereby achieving comprehensive acquisition and standardized processing of information related to respiratory chronic diseases; secondly, a dual encoding pre-training and contrastive learning fine-tuning framework is used to process structured patient data sets and respiratory disease professional knowledge bases, and the generated respiratory chronic disease large model and respiratory chronic disease knowledge graph have deep medical understanding and professional reasoning capabilities, which are significantly better than traditional machine learning models, and can understand complex medical contexts and extract key clinical features from them; thirdly, semantic analysis and multi-dimensional calculation of patient symptom descriptions and evaluation scale answers are performed based on the respiratory chronic disease large model and knowledge graph, and the generated patient risk stratification assessment results are highly accurate, can capture subtle differences in disease characteristics, and achieve accurate stratification, thereby optimizing the allocation of medical resources; fourthly, According to the results of patient risk stratification assessment, the respiratory chronic disease big model is used to retrieve clinical guideline rules and generate personalized diagnosis and treatment plans and matching scores based on the individual characteristics of patients, which fully considers the individual differences of patients, so that the treatment plan is both in line with the principles of evidence-based medicine and highly personalized, thereby improving the accuracy and compliance of treatment; Fifth, the multi-level medical institution collaboration framework and dynamic efficacy monitoring system constructed based on personalized diagnosis and treatment plans and matching scores have broken the medical information island, achieved the rational allocation of high-quality medical resources and continuous medical services, and reduced repeated examinations and treatment errors; patient health education content and self-management tools customized according to efficacy indicators and treatment response indexes, combined with the continuous collection of system operation data, model updates and process optimization, have formed a closed-loop quality improvement system, enabling the system to continuously evolve and improve. It is particularly worth emphasizing that the large model algorithm used in this solution has made great contributions to the field of diagnosis, treatment and management of chronic respiratory diseases. Its contextual understanding ability enables the model to extract key clinical information from unstructured medical texts, and the knowledge-enhanced reasoning mechanism realizes automated support for complex medical decision-making. The multimodal data fusion algorithm can integrate patient information from different sources to form a comprehensive view, and the adaptive learning characteristics enable the system to continuously optimize diagnosis and treatment strategies based on new data, greatly improving the efficiency and accuracy of chronic respiratory disease management.

[0020] In a specific embodiment, the process of executing step S101 may specifically include the following steps: Through clinical input from medical institutions, self-reporting from patients’ mobile terminals, and smart wearable devices, basic patient information, medical history, medication status, symptom assessment, and physiological monitoring data are collected to form a multi-source heterogeneous data set; Standardize and normalize structured data in multi-source heterogeneous data sets, perform word segmentation, stop word removal and entity standardization operations on text data, implement noise reduction, missing value interpolation and outlier processing on time series data, and generate preprocessed data sets; The preprocessed data set is annotated according to the respiratory system medication classification method, a unique patient identifier is established, and an initial annotated data set is generated; The initial annotated data set is organized into a tree-like hierarchical structure, and the top-level patient basic information, the second-level disease classification, and the bottom-level symptom monitoring indicators are associated with the corresponding patient files to obtain a hierarchical associated data set; Perform integrity check, consistency verification and outlier marking on hierarchical association datasets to generate quality-controlled datasets; The quality-controlled data sets are stored in a secure database to form a structured patient data set.

[0021] Specifically, the basic data of patients is collected through clinical entry in medical institutions. During the diagnosis and treatment process, medical staff enter the patient's demographic information such as gender, age, occupation, and living environment into the system. At the same time, they collect medical history information such as the date of disease diagnosis, the number of previous acute attacks, and the history of hospitalization, as well as medication information such as the type of drug currently used, dosage, and compliance records. Patients' mobile self-reports collect patients' daily symptom changes and medication status through specially designed mobile applications. Patients fill out assessment forms such as the degree of dyspnea, cough frequency, sputum characteristics, and degree of activity limitation regularly or when symptoms change. Smart wearable devices automatically collect patients' physiological monitoring data such as lung function measurement values, blood oxygen saturation, and respiratory rate. These devices synchronize data with the system in real time or regularly to form continuous physiological parameter monitoring records. These data from different channels form a multi-source heterogeneous data set. Due to the diverse data sources, differences in format and content, data preprocessing is required. Standardize structured data to convert numerical data of different dimensions into a unified standard, for example, unify the lung function test values ​​of different laboratories into percentages of expected values; normalize the data range to between 0 and 1 to make the indicators comparable. Perform word segmentation on text data to decompose the symptom experience described by the patient into basic semantic units; remove stop words, and delete function words and conjunctions that have no substantive meaning for medical judgment; entity standardization maps synonymous expressions to medical terms. The processing of time series data includes noise reduction, eliminating short-term fluctuation interference through methods such as sliding window averaging; missing value interpolation, making reasonable inferences based on the previous and next values ​​of the time series or similar patterns of similar patients; outlier processing, identifying and correcting data points that deviate significantly from the normal range.

[0022] The preprocessed dataset needs to be annotated according to the respiratory medication classification. The respiratory medication classification is a classification system specifically for drugs for respiratory diseases, including bronchodilators, anti-inflammatory drugs, mucus regulators and other categories. The data is marked accordingly according to the patient's medication history and current treatment plan. At the same time, a unique patient identifier is established. By synthesizing multiple personal information and encrypting them, the uniqueness of each patient in the system is ensured to avoid data confusion and form an initial annotated dataset. The initial annotated dataset is organized in a tree-like hierarchical structure. The top layer stores basic patient information, including demographic characteristics and unique identifiers; the second layer is disease classification, which classifies patients into specific respiratory disease types, such as chronic obstructive pulmonary disease, asthma, interstitial lung disease, etc.; the bottom layer is symptom monitoring indicators, including all disease-related symptom scores, physiological indicators and medication records. The three layers of data are associated with the corresponding patient files through unique identifiers to form a patient view, and a clear reference relationship is established between the data at each level.

[0023] Quality control is performed on hierarchical association data sets, and integrity checks are performed to confirm whether the necessary information of each patient is complete, including basic demographic information, disease diagnosis, core symptom assessment, and key physiological indicators; consistency verification checks whether there are contradictions within the data, such as whether different records at the same time point conflict, and whether symptom descriptions match objective indicators; outlier marking sets reasonable ranges for each indicator and marks data that exceeds the range, such as extreme lung function values ​​or symptom scores that do not conform to common sense. Data sets that have passed quality control are stored in a secure database, using encrypted storage and access control mechanisms to ensure patient data security and form a structured patient data set.

[0024] Take the data processing of a patient with chronic obstructive pulmonary disease as an example: the original data was collected from three channels. The basic information entered by the medical institution showed that the patient was a 65-year-old male and a retired teacher; the medical history record showed that COPD was diagnosed 5 years ago and there were 2 acute attacks in the past year; the current medication included salmeterol / fluticasone inhaler twice a day. The symptom data reported by the mobile terminal included daily dyspnea scores (mMRC scale) fluctuating between 2-3 points in the past week, coughing frequency of 3-5 times a day, and white viscous sputum. The average nighttime blood oxygen saturation recorded by the smart wearable device was 93%, with a brief drop to 88%. After preprocessing of these heterogeneous data, the dyspnea score was standardized to moderate with an average of 2.5 points; the blood oxygen data was processed by noise reduction to eliminate an obvious false reading of device detachment; the patient-reported "chest tightness and shortness of breath" was standardized to the medical term "dyspnea" through text processing. The data were annotated according to the combination of LABA+ICS and organized into a tree structure after establishing a unique identifier. The top layer was personal basic information, the second layer was COPD diagnosis (GOLD grade B), and the bottom layer was symptoms and monitoring indicator details. Quality control found that there was a 30-minute gap in the night blood oxygen data, which was interpolated by the average value of the previous and next time periods; at the same time, a record of a sudden increase in self-reported dyspnea without a corresponding change in physiological indicators was marked, which required further verification. The structured data set formed contains the patient's chronic respiratory disease portrait.

[0025] In a specific embodiment, the process of executing step S102 may specifically include the following steps: Integrate clinical guidelines for respiratory diseases, evidence-based medicine literature, expert consensus documents, typical case collections, and drug interaction databases to build a professional knowledge base for respiratory diseases; The structured patient dataset and the medical text corpus in the respiratory disease professional knowledge base were simultaneously input into the Transformer encoder structure, and multiple rounds of iterative training were performed using the masked language model and the next sentence prediction task to obtain the initial large model parameters. A comparative learning framework was constructed based on the initial large model parameters to bring similar case representation vectors closer together and different case representation vectors farther apart. The model parameters were optimized by minimizing the comparative loss function to generate a large model of chronic respiratory diseases. Entity extraction was performed on the structured patient dataset and the professional knowledge base of respiratory diseases to identify diseases, symptoms, drugs, examinations, and risk factors as medical concepts and obtain a set of medical entities; Extract relations from the medical entity set, determine disease-symptom, disease-drug, drug-drug, symptom-severity, examination-reference value as the associations between entities, and form a set of relation triples; The medical entity set and the relationship triple set are used for knowledge representation learning through graph embedding algorithm to construct a knowledge graph of respiratory chronic diseases. It is then associated with the large model of respiratory chronic diseases through entity linking technology to support knowledge-enhanced reasoning.

[0026] Specifically, the authoritative clinical guidelines for respiratory diseases are integrated, including the COPD guidelines issued by the Global Initiative for Chronic Obstructive Lung Disease (GOLD), the asthma guidelines issued by the Global Initiative for Asthma (GINA), and the diagnosis and treatment guidelines for interstitial lung disease. These clinical guideline data are organized in the form of structured XML files, extracting core content such as disease definitions, diagnostic criteria, grading and classification methods, treatment recommendations, and management strategies. The integration of evidence-based medicine literature covers high-quality research literature on respiratory diseases. Through professional databases, relevant randomized controlled trials, systematic reviews, and meta-analyses published in the past five years are retrieved to extract research findings, evidence-based conclusions, and evidence levels. The integration of expert consensus documents includes the disease treatment consensus issued by respiratory professional societies in various countries, and extracts expert experience and clinical recommendations. The typical case collection collects representative cases of various chronic respiratory diseases, including disease manifestations, diagnostic processes, treatment plans, and prognosis results. The integration of drug interaction databases summarizes the interaction information between commonly used respiratory system drugs and other system drugs to form a drug interaction network. These multi-source data are constructed into a structured professional knowledge base for respiratory diseases through text extraction, topic modeling, and content indexing. The structured patient dataset and the medical text corpus in the respiratory disease expertise database were input into the Transformer encoder structure for model training. The Transformer encoder is a neural network structure based on the self-attention mechanism that can capture long-distance dependencies in the text. The training process uses two key tasks: masked language model and next sentence prediction task. In the masked language model task, 15% of the tokens in the input text are randomly replaced with special tags, and then the model is trained to predict these masked words based on the context. For example, the word "dyspnea" in "patients show severe marks and sputum" is masked to let the model learn to infer reasonable medical terms based on the context. The next sentence prediction task provides the model with two sentences and trains it to determine whether the second sentence is a natural continuation of the first sentence, thereby learning the coherence and logical relationship of the text. Through multiple rounds of iterative training of these two tasks, the model gradually learns the representation of medical language and the expertise of respiratory diseases, forming the initial large model parameters.

[0027] Based on the initial large model parameters, a contrastive learning framework is constructed to further optimize the model. Contrastive learning is a method of learning representations by comparing similarities and differences between samples. In this scheme, clinically similar respiratory chronic disease cases are used as positive sample pairs, and cases of different types or severity are used as negative sample pairs. Different expression variants of cases are generated through data augmentation technology, keeping the core medical content unchanged but diversifying the expression form. The large model is used to extract features for each case to obtain a high-dimensional representation vector. Then the cosine similarity between vectors is calculated, and a contrastive loss function is constructed to reduce the distance between the representation vectors of similar cases and increase the distance between the representation vectors of different cases. The contrastive loss function is minimized by the gradient descent algorithm, and the model parameters are iteratively optimized to generate a dedicated large model sensitive to respiratory chronic diseases. Entity extraction is performed on the structured patient data set and the professional knowledge base to identify key medical concepts. Named entity recognition technology is used to identify medical entities such as diseases, symptoms, drugs, examinations, and risk factors in the text through biomedical-specific dictionaries and contextual semantic analysis. Each identified entity is assigned a unique identifier and category label to form a medical entity set.

[0028] Then, the medical entity set is subjected to relationship extraction to establish semantic connections between entities. The relationship between entities is extracted from the text through dependency syntactic analysis and semantic role labeling technology. For example, the two disease-symptom relationships "asthma-manifestation-wheezing" and "asthma-manifestation-dyspnea" are extracted from the sentence "asthma often manifests as wheezing and dyspnea"; the drug-drug relationship "albuterol-synergistic effect-ipratropium bromide" is extracted from "salbutamol and ipratropium bromide combined can enhance the bronchial dilation effect". In addition, the association between symptoms and severity and the association between examinations and reference values ​​are also extracted. Each extracted relationship forms a subject-relationship-object triple structure, which is summarized to form a set of relationship triples. The medical entity set and the relationship triple set are converted into a knowledge graph through a graph embedding algorithm. The graph embedding algorithm maps each medical entity and relationship to a low-dimensional vector space, so that concepts with similar semantics are close in the vector space. Commonly used graph embedding algorithms include TransE, ComplEx, and RotatE. In this solution, the improved TransE algorithm is used, which learns the vector representation of entities and relationships by optimizing the objective function of the head entity vector plus the relationship vector in the triple to approach the tail entity vector. The generated vector representation retains the semantic relationship between entities and constitutes a knowledge graph of chronic respiratory diseases. Through entity linking technology, a mapping relationship is established between the medical concepts mentioned in the text and the entity nodes in the knowledge graph, thereby realizing the fusion of the knowledge graph and the large model and supporting knowledge-enhanced medical reasoning.

[0029] Take the processing of knowledge related to chronic obstructive pulmonary disease (COPD) as an example: the definition "COPD is a common, preventable and treatable disease characterized by persistent respiratory symptoms and airflow limitation" is extracted from the GOLD guidelines. Through entity extraction, "COPD" is identified as a disease entity, and "persistent respiratory symptoms" and "airflow limitation" are identified as symptom entities. In the relationship extraction stage, two triplets "COPD-characteristic is-persistent respiratory symptoms" and "COPD-characteristic is-airflow limitation" are identified. At the same time, "smoking is the main risk factor for COPD" is extracted from the evidence-based literature, and "smoking" is identified as a risk factor entity to form the triple "smoking-is-risk factor for COPD". COPD grading standards are extracted from clinical guidelines, such as "FEV1 / FVC<0.7 and FEV1<30% predicted value is severe COPD", forming triplets such as "FEV1 / FVC<0.7-diagnostic criteria-COPD" and "FEV1<30% predicted value-severity assessment-severe COPD". The drug treatment part extracts "Long-acting β2 receptor agonists (LABA) and long-acting anticholinergics (LAMA) are used in combination for COPD patients with severe symptoms", forming relationships such as "LABA-combination medication-LAMA" and "LABA+LAMA-treatment for-COPD with severe symptoms". These entities and relationships are mapped to the vector space through the graph embedding algorithm. For example, "COPD" is mapped to a 128-dimensional vector, and clustered with related entities such as symptoms and drugs in the vector space to construct a local graph structure of COPD-related knowledge. When new patient data is entered into the system, the patient's symptom "severe dyspnea" is associated with the corresponding node in the knowledge graph through entity linking. The large model infers possible diagnoses and suitable treatment plans based on this, realizing knowledge-enhanced intelligent diagnosis and treatment decisions.

[0030] In a specific embodiment, the process of executing step S103 may specifically include the following steps: Through the natural language interactive interface, patients are guided to describe their current symptom experience, and the patient's natural language input is input into the respiratory chronic disease large model. The symptom duration, severity, inducement, and relief factors are extracted as key symptom attributes to generate a symptom entity matrix. Guide the patients to complete the dyspnea score, cough frequency score, sputum characteristics score, activity limitation score, sleep impact score, and medication compliance score as a structured symptom assessment questionnaire to obtain the questionnaire numerical vector; The symptom entity matrix and the questionnaire numerical vector are fused, and the lung function index and blood oxygen saturation values ​​in the basic physiological monitoring data are combined to construct a multidimensional feature space to form a comprehensive symptom representation of the patient; The dyspnea, cough and sputum in the comprehensive symptom representation of the patient are assigned weight coefficients as core symptoms. The severity value of each symptom is multiplied by the corresponding weight coefficient through the weighted calculation method and the sum is obtained to obtain the symptom burden index. Conduct time series analysis on recent symptom fluctuation trends, acute attack risks, medication compliance values, and environmental factors, generate a disease stability index through a stability calculation model, and calculate a comprehensive risk score based on the severity of the underlying disease, comorbidities, and social support scores; The threshold interval is set according to the comprehensive risk score, and patients with a risk score lower than the first threshold are classified as a low-risk level, patients with a risk score between the first and second thresholds are classified as a medium-risk level, and patients with a risk score higher than the second threshold are classified as a high-risk level, thereby generating patient risk stratification assessment results.

[0031] Specifically, the natural language interactive interface is used to guide patients to describe their current symptom experience. The interactive interface adopts a semi-structured inquiry method, asking open questions such as "Please describe your recent dyspnea", supplemented by guiding questions such as "How long has this situation lasted?" and "What factors will aggravate or relieve symptoms?" to guide patients to make a comprehensive statement. The natural language description entered by the patient is passed into the respiratory chronic disease model, and the key symptom attributes are identified through named entity recognition and relationship extraction technology. The symptom duration attribute extracts time expressions from the text, such as "started three days ago" is converted into a specific time length value; the severity attribute identifies words that describe intensity, such as "mild", "obvious", and "severe" are converted into a three-level scale of light, moderate, and severe; the inducement attribute extracts symptom triggering factors, such as "after exercise", "contact with pollen", "inhalation of cold air", etc.; the mitigation factor attribute captures the conditions for improving symptoms, such as "after rest" and "after using an inhaler". These extracted attribute values ​​are organized into a structured symptom entity matrix, where each row represents a symptom, each column represents an attribute dimension, and the elements in the matrix are the corresponding attribute values. At the same time, the patients were guided to complete a structured symptom assessment questionnaire, including a rating scale for multiple dimensions. The dyspnea score was scored using the modified Medical Research Council (mMRC) scale, from 0 points, indicating that dyspnea was felt only during strenuous activities to dyspnea so severe that it was impossible to leave home or put on clothes; the cough frequency score ranged from 0 to 3 points, indicating that it ranged from no cough to continuous cough affecting daily activities; the sputum characteristics score ranged from 0 to 3 points, indicating that it ranged from no sputum to a large amount of purulent sputum; the activity limitation score ranged from 0 to 5 points, indicating that it ranged from no restrictions to complete bed rest; the sleep impact score ranged from 0 to 5 points, indicating that it ranged from normal sleep to being completely unable to fall asleep due to respiratory symptoms; the medication compliance score ranged from 0 to 2 points, indicating that it ranged from completely following the doctor's orders to basically not taking the medication according to the doctor's orders. Patients completed the score by selecting the option that best suited their own situation. These scores formed a questionnaire numerical vector, and each element of the vector corresponded to a scoring dimension.

[0032] The symptom entity matrix and the questionnaire numerical vector are feature fused using feature splicing and cross-coding methods. Feature splicing directly combines the two parts of data into a larger feature vector; cross-coding calculates the interaction between the two parts of the features to generate interactive features. For example, when the duration of dyspnea in the symptom entity matrix is ​​"long-term" and the dyspnea score in the questionnaire is 4 points, the interactive feature will strengthen the representation of this serious condition. Combined with basic physiological monitoring data, in an embodiment of the present application, the basic physiological monitoring data includes lung function indicators and blood oxygen saturation values, which further expand the feature space and obtain a multi-dimensional comprehensive symptom representation of patients.

[0033] The core symptoms in the comprehensive symptom representation of the patient are weighted to obtain the symptom burden index. Dyspnea, as the indicator that best reflects the patient's quality of life and disease severity, is given the highest weight; cough, as the most common symptom, is closely related to social function and sleep quality, and is given the second highest weight; sputum changes reflect airway inflammation and infection risk, and are also given appropriate weights. In the weighted calculation process, each symptom score is first standardized to the 0-1 range, then multiplied by the corresponding weight coefficient, and finally summed to obtain the symptom burden index. For example, the standardized dyspnea score of 0.75 multiplied by the weight of 0.5, the cough score of 0.6 multiplied by the weight of 0.3, and the sputum score of 0.4 multiplied by the weight of 0.2, the sum of the three 0.75×0.5+0.6×0.3+0.4×0.2=0.62 is used as the symptom burden index.

[0034] Multiple time series data of patients were analyzed to evaluate disease stability. Recent symptom fluctuation trends were calculated through the coefficient of variation and trend slope of continuous monitoring data. A large coefficient of variation or a positive slope indicated unstable symptoms. The risk of acute attacks was quantified as a risk value between 0 and 1 based on historical attack frequency, seasonal factors, and current symptom change predictions. The medication compliance value was obtained from electronic medicine box records and patient self-reported data and converted into a compliance percentage. Environmental factors considered meteorological conditions, air quality, and personal sensitivity factors. These time series indicators were calculated to generate a disease stability index. The higher the value, the more stable the condition. At the same time, the overall risk score was calculated based on the disease severity classification of the medical guidelines, the number and severity of the patient's comorbidities, and the social support score.

[0035] The threshold setting of risk score is based on large-scale clinical data analysis and expert consensus. By retrospectively analyzing the historical data of patients with chronic respiratory diseases, statistically analyzing the correlation between different risk scores and adverse events, two key dividing points were determined: the first threshold was set at 40 points, and the historical data of patients below this value showed that the incidence of adverse events was less than 5%; the second threshold was set at 65 points, and the incidence of adverse events in patients above this value exceeded 20%. Based on these two thresholds, patients are divided into three risk levels: low risk level (score <40), suitable for management by primary medical institutions, only basic intervention is required; medium risk level (score 40-65), requiring enhanced monitoring and more active treatment strategies; high risk level (score>65), requiring intervention and close monitoring by specialized hospitals.

[0036] In a specific embodiment, the process of executing step S104 may specifically include the following steps: The patient risk stratification assessment results are input into the respiratory chronic disease model to identify the patient's disease classification. When the patient has chronic obstructive pulmonary disease, the patient is divided into four groups A / B / C / D according to the degree of airflow limitation and symptom score. When the patient has asthma, the patient is marked based on the control level and severity to generate disease classification results; The disease classification results are matched with clinical guideline rules. When the patient belongs to group A, a mild medication regimen is retrieved; when the patient belongs to group B, a moderate medication regimen is retrieved; when the patient belongs to group C or group D, a severe medication regimen is retrieved to obtain a basic treatment framework; Compare and analyze the basic treatment framework with the patient's previous treatment records and history of adverse drug reactions to generate a preliminary personalized drug plan; Based on the preliminary personalized drug plan, formulate drug administration information, including preferred drug class combination, specific drug selection, administration route, dosage, treatment course design, and generate a drug treatment plan; In combination with drug treatment plans, non-drug treatment recommendations are created, including respiratory rehabilitation training programs, lifestyle adjustment recommendations, mental health interventions, and disease self-management skills training. When the patient's risk level is high, an acute exacerbation emergency plan is developed to form a personalized diagnosis and treatment plan; The personalized diagnosis and treatment plan is matched and evaluated with the patient's characteristics, and the patient's previous medication response matching score, lifestyle habit matching score, compliance factor matching score, economic affordability matching score, and medical insurance coverage matching score are calculated to obtain a percentage matching score.

[0037] Specifically, the patient risk stratification assessment results are input into the respiratory chronic disease big model for disease classification. The big model analyzes the patient's basic clinical data, symptom manifestations and physiological indicators to identify specific disease types. For patients with chronic obstructive pulmonary disease (COPD), according to the Global Initiative for Chronic Obstructive Lung Disease (GOLD) guidelines, the degree of airflow limitation (expressed as a percentage of the predicted FEV1 value) and the symptom score (usually using the CAT score or mMRC dyspnea score) are used as two key dimensions. When FEV1 ≥ 80% of the predicted value is mildly limited, 50% ≤ FEV1 < 80% of the predicted value is moderately limited, 30% ≤ FEV1 < 50% of the predicted value is severely limited, and FEV1 < 30% of the predicted value is extremely severely limited. In terms of symptom burden, when CAT < 10 points or mMRC < 2 points, the symptoms are mild, and CAT ≥ 10 points or mMRC ≥ 2 points are severe symptoms. Combined with the history of acute exacerbations in the past year (0, 1 non-hospitalized exacerbation, ≥2 non-hospitalized exacerbations, or ≥1 hospitalized exacerbation), COPD patients were divided into four groups: A / B / C / D: Group A (low symptoms, low risk), Group B (high symptoms, low risk), Group C (low symptoms, high risk), and Group D (high symptoms, high risk). When patients have asthma, they are classified based on the level of symptom control and acute exacerbation risk assessment according to the Global Initiative for Asthma (GINA) guidelines. The level of symptom control is divided into good control (no daytime symptoms or ≤2 times per week, no nighttime symptoms, no activity limitation, no need for rescue medication or ≤2 times per week), partial control (1-2 poor control indicators), and uncontrolled (≥3 poor control indicators). At the same time, the severity of asthma is considered and divided into mild, moderate, severe, and very severe to form a standardized disease classification result.

[0038] According to the disease classification results, the large model retrieves and matches the corresponding clinical guideline rules to generate a basic treatment framework. For COPD patients, when the patient belongs to group A, the mild medication regimen is retrieved, mainly including the use of short-acting bronchodilators on demand; when the patient belongs to group B, the moderate medication regimen is retrieved, and the long-term regular use of long-acting bronchodilators is recommended; when the patient belongs to group C or group D, the severe medication regimen is retrieved, including the combined use of LABA and LAMA, and the addition of inhaled corticosteroids (ICS) or phosphodiesterase-4 inhibitors (PDE4i) when necessary. For asthma patients, the treatment plan is matched according to the principle of step-by-step treatment, starting with low-dose ICS, and gradually adjusted to ICS / LABA combination therapy according to the control situation. Severe patients consider adding leukotriene receptor antagonists or biological agents. The relevant drug indications, contraindications and recommended doses are retrieved through the knowledge graph to form a basic treatment framework for specific disease classifications.

[0039] The basic treatment framework was compared and analyzed with the patient's previous treatment records and history of adverse drug reactions. The drug use, dosage adjustment history, and treatment response records in the patient's previous treatment records were structured and extracted to analyze the effectiveness of the drugs. At the same time, the adverse reaction history was checked to confirm whether the patient had allergic reactions, intolerance, or other adverse reactions to specific drugs or drug components. By setting the priority rules for drug selection: when the patient has a good response to a certain class of drugs in the past, the priority of this class of drugs is increased; when the patient has a history of adverse reactions to a certain class of drugs, the drug is deleted from the recommended list or the priority is reduced; when the patient responds to multiple drugs, the combination with the least adverse reactions and the best efficacy is selected. Based on the preliminary personalized drug plan, detailed dosing information is developed. The preferred drug category combination is determined, such as COPD patients may be a LAMA+LABA dual bronchodilator combination, and asthma patients may be ICS+LABA combination medication. Then select specific drugs from each category, considering drug characteristics (such as onset time, duration of action), type of drug delivery device (such as metered dose inhaler, dry powder inhaler, nebulizer), and patient use ability. Inhalation is the preferred route of administration to ensure that the drug acts directly on the respiratory tract and reduce systemic adverse reactions. The dosage and usage are adjusted individually according to the standard dosage in the drug instructions combined with the patient's weight, age, liver and kidney function, and the precise time and frequency of administration are determined. The course of treatment is designed to take into account the different needs of short-term intensive treatment in the acute phase and long-term maintenance treatment, including the initial dose stage, stabilization period and adjustment period, and to clarify the staged evaluation points and dose adjustment conditions.

[0040] On the basis of the drug treatment plan, non-drug treatment recommendations are created. The respiratory rehabilitation training program includes breathing control techniques, limb exercise prescriptions and endurance training programs, which are personalized according to the patient's basic physical fitness, activity endurance and comorbidities. Lifestyle adjustment recommendations include smoking cessation guidance, environmental control measures, dietary guidance and occupational protection measures. Mental health intervention measures mainly target common anxiety and depression problems in patients with chronic respiratory diseases, providing cognitive behavioral intervention strategies and coping skills. Disease self-management skills training includes symptom identification, medication techniques, correct use of inhalation devices, daily monitoring points, etc. For patients with high risk levels, additional emergency plans for acute exacerbations are formulated to clarify warning symptoms (such as significantly worsened dyspnea, increased sputum volume, and changes in sputum color), self-treatment measures (such as adjusting drug dosages and postural drainage) and criteria for judging the timing of medical treatment to ensure that patients can respond to changes in their condition in a timely manner. These contents are integrated into a personalized diagnosis and treatment plan, covering all-round guidance on drug and non-drug interventions.

[0041] The matching degree of personalized diagnosis and treatment plan and patient characteristics is evaluated. The matching degree evaluation is divided into multiple dimensions: the matching score of previous medication response evaluates the consistency between the recommended drugs in the plan and the patient's historical medication response; the matching score of lifestyle habits evaluates the compatibility of treatment recommendations with the patient's daily activity pattern and life routine; the matching score of compliance factors evaluates the patient's ability and willingness to implement complex treatment plans; the matching score of economic affordability evaluates the degree of matching between treatment costs and the patient's economic situation; the matching score of medical insurance coverage evaluates the medical insurance reimbursement of recommended drugs and treatment items. Different weights are assigned to each dimension, and the weighted total score is calculated to obtain a matching degree score in percentage. When the matching degree score is lower than the preset threshold (usually 60 points), the diagnosis and treatment plan is readjusted until a satisfactory matching degree is achieved to ensure that the plan is both in line with medical guidelines and feasible.

[0042] In a specific embodiment, the process of executing step S105 may specifically include the following steps: According to the patient risk stratification assessment results and matching scores, low-risk patients are assigned to primary medical institutions, medium-risk patients are arranged to be managed by primary medical institutions and regularly followed up by specialist hospitals, and high-risk patients are managed by specialist hospitals and followed up by primary medical institutions, so as to build a hierarchical diagnosis and treatment pathway map; Connect medical institutions of different levels through an intelligent collaborative platform to share medical records, examination results, and treatment opinions, forming a medical collaborative network; For patients with chronic respiratory diseases seen by primary medical institutions, the patient data and the doctor's preliminary judgment are input into the chronic respiratory disease big model to analyze the symptom characteristics, disease progression risk and treatment difficulty, and generate diagnosis and treatment decision recommendations; Conduct multi-dimensional symptom monitoring on patients who start treatment, record changes in core symptom scores, changes in physiological indicators, medication compliance data, and quality of life assessment data, and build a full-dimensional monitoring data set; Conduct time series analysis on the full-dimensional monitoring data set, calculate symptom improvement rate, physiological index recovery rate, medication compliance index, and quality of life improvement, and generate multidimensional efficacy indicators; Based on multidimensional efficacy indicators, the comprehensive score of treatment response is calculated through a weighted aggregation algorithm, and the trend of efficacy is judged according to the trend of score changes. When the score drops for more than three consecutive monitoring cycles, the treatment plan adjustment process is triggered to form a time series treatment response index.

[0043] Specifically, a hierarchical diagnosis and treatment pathway is used to achieve reasonable patient diversion. According to the patient risk stratification assessment results and the matching score of the diagnosis and treatment plan, a resource allocation decision matrix is ​​established. Low-risk patients (risk score <40) are assigned to primary medical institutions, and routine management is performed by community health service centers or township health centers; medium-risk patients (risk score 40-65) are arranged to primary medical institutions for daily management, and follow-up visits are conducted in specialist hospitals every 3 months to achieve two-way management; high-risk patients (risk score >65) are managed by the respiratory department of specialist hospitals, and primary medical institutions cooperate to conduct monthly follow-up visits to ensure timely detection of changes in their condition. For patients with a matching score below 70 points, regardless of the risk level, the first visit is arranged in a specialist hospital, and the specialist doctor formulates an initial plan before transferring to the corresponding level of medical institutions. The hierarchical diagnosis and treatment pathway is constructed in the form of a flow chart, which clarifies the responsible parties, referral standards and information flow rules of each node, and transforms the abstract hierarchical management principles into executable specific processes. Multi-level medical institutions achieve data interoperability and collaboration through an intelligent collaborative platform. The platform adopts a federated learning architecture, and each medical institution retains local data storage while sharing necessary information through a secure channel. Medical record information sharing includes standardization of disease diagnosis, structuring of medication history, and quantification of treatment response, making medical records readable across institutions; examination results are shared in a unified standard format, such as converting lung function test values ​​into percentages of expected values, to achieve data comparability between different laboratories; and diagnosis and treatment opinion sharing records the judgments and suggestions of specialists and primary care doctors according to standardized templates. During data sharing, differential privacy technology is used to protect patient sensitive information, and blockchain technology is used to record data access and modification logs to ensure traceability of the entire process. Through shared data and interconnection mechanisms, a collaborative network connecting medical institutions at different levels is built to achieve resource integration and diagnosis and treatment collaboration.

[0044] For patients with chronic respiratory diseases seen by primary medical institutions, primary doctors enter the patient's basic data and preliminary judgment, including the main symptoms, physical examination findings, simple pulmonary function test results and preliminary diagnosis opinions. After standardized processing, these data are input into the large model of chronic respiratory diseases. The large model is based on knowledge graph-assisted analysis, calculates the similarity of symptom patterns, identifies typical or atypical manifestations, evaluates the difficulty of treatment, and generates diagnosis and treatment decision recommendations. Decision recommendations include: diagnosis confirmation or revision, examination recommendations, treatment adjustment recommendations and referral indication judgment. When the large model detects complex situations that are difficult for primary medical institutions to handle, such as serious complications, atypical symptom combinations or poor treatment response, it automatically generates a referral application, including a summary of key information and reasons for referral, and pushes it to the corresponding specialist hospital workstation.

[0045] Multi-dimensional symptom monitoring is carried out on patients with chronic respiratory diseases who have started treatment, and a dynamic tracking system is established. Changes in core symptom scores are collected through regular questionnaires, including dyspnea mMRC scores, cough scores, sputum scores, etc. Patients fill out the questionnaire once a week to record the trend of changes; changes in physiological indicators are collected through portable monitoring equipment, such as pulmonary function meters to measure FEV1 values, and blood oxygen saturation monitors to record nighttime blood oxygen fluctuations, and the data are automatically uploaded to the management platform; medication compliance data are recorded through smart medicine boxes to record the actual medication time and frequency, combined with the patient's self-reported situation, quantified into compliance percentages; quality of life assessment data uses respiratory disease-specific scales such as CAT or AQLQ to regularly assess changes in patients' quality of life. After the multi-source data is unified in format and missing values ​​are processed, it is stored in the patient's electronic health record to form a time series data set, which constitutes a full-dimensional monitoring data set.

[0046] Time series analysis was performed on the full-dimensional monitoring data set to calculate multiple efficacy indicators. The symptom improvement rate was calculated based on the change in the core symptom score. The difference in scores before and after treatment was divided by the initial score to obtain a standardized improvement percentage; the physiological indicator recovery rate calculated the degree of recovery of key physiological parameters such as FEV1 relative to the expected normal value and tracked its change curve; the medication compliance index was calculated by calculating the ratio of the actual number of medications to the number of times prescribed by the doctor, combined with the medication time accuracy score to calculate the comprehensive compliance index; the improvement in quality of life was calculated by the difference before and after the standardized scale score. Thresholds were set for each indicator, such as a symptom improvement rate of >30% for significant improvement, 15-30% for moderate improvement, and <15% for slight improvement. These calculation results form structured multidimensional efficacy indicators that intuitively reflect all aspects of the treatment effect.

[0047] Based on multidimensional efficacy indicators, the comprehensive score of treatment response is calculated by weighted aggregation algorithm. The weight of each indicator is determined. For COPD patients, lung function indicators and dyspnea scores have higher weights; for asthma patients, symptom control rate and acute attack frequency have higher weights. Each indicator is standardized to a range of 0-100 points, and the weighted sum is calculated according to the weight to obtain the comprehensive score of treatment response. The score is calculated once per cycle (usually 2 weeks) to form time series data. The changing trend of treatment response is evaluated by time series analysis methods, including moving average, trend analysis and seasonal decomposition. Set monitoring and early warning rules: when the score rises continuously, it indicates that the treatment effect is good; when the score is stable, it indicates that the condition is stable; when the score fluctuates significantly, it indicates that the control is unstable; when the score drops continuously for more than three monitoring cycles, the treatment plan adjustment process is triggered. The comprehensive score and its change rate calculated each time are stored in the patient time series database to form a time series treatment response index.

[0048] In a specific embodiment, the process of executing step S106 may specifically include the following steps: Based on the patient's disease type, severity, cognitive level, technology acceptance and efficacy indicators, we screen the basic knowledge of the disease, symptom identification methods, medication guidance content, lifestyle adjustment strategies, and acute exacerbation prevention measures to generate a personalized health education content library; Convert the personalized health education content library into multimedia resources in the form of text, pictures, videos, and interactive questions and answers, dynamically adjust the presentation depth according to the patient's learning progress and content understanding, and build a graded education resource package; Based on personalized diagnosis and treatment plans and treatment response indexes, develop an electronic pill box reminder system, a symptom self-monitoring diary, a breathing training guidance system, and an environmental risk warning system as functional modules to form a patient self-management tool set; Integrate health education content and self-management tool sets into mobile applications, web pages, and smart wearable devices to generate patient-side interactive interfaces based on patients’ technological acceptance capabilities; Record and analyze the operation behaviors, feedback, diagnosis and treatment effect evaluation, and large model performance indicators of patients and medical staff during the use of the system to form a system optimization data set; Based on the system optimization data set, the knowledge base of the respiratory chronic disease model is updated and the algorithm is tuned. The efficiency of the system process nodes is analyzed and the resource allocation is optimized. The user interface and interaction logic are optimized for human-computer interaction to complete system iteration and process optimization.

[0049] Specifically, personalized health education content is screened through patient characteristic parameter analysis, where the patient's disease type is divided into chronic obstructive pulmonary disease, asthma, bronchiectasis and other respiratory chronic diseases; the severity is classified according to clinical grading standards, such as the GOLD grading of COPD or the control level grading of asthma; the cognitive level is determined as elementary, intermediate or advanced through a simple knowledge assessment; the technical acceptance is assessed as low, medium and high through a questionnaire on technology usage habits; and the previously calculated efficacy index data is integrated. Based on this five-dimensional feature vector, relevant content is extracted from the professional knowledge base of respiratory diseases, and the etiology mechanism, natural course, prognostic factors and other content related to the patient's specific disease type are selected for the basic knowledge of the disease; the warning symptom identification points and self-assessment tools suitable for the severity of the patient are extracted for the symptom identification method; the medication guidance content is directly related to the patient's personalized diagnosis and treatment plan, including the mechanism of action, correct use method, precautions, etc. of the drugs used; the lifestyle adjustment strategy selects appropriate exercise programs, dietary recommendations and environmental control measures based on the individual situation of the patient; and the acute exacerbation prevention measures are customized according to the patient's previous attack pattern and risk factors. The content modules are matched with the patient's feature vectors through a similarity matching algorithm to select the content combination that best suits the patient and form a personalized health education content library. The personalized health education content library is transformed into a variety of multimedia resources to achieve the acceptability and interactivity of the content. The text materials use graded reading difficulty control, set vocabulary complexity and sentence structure according to the patient's cognitive level, and are accompanied by professional terminology explanations; picture resources include medical illustrations, anatomical diagrams, and operation diagrams, which intuitively present abstract concepts in a visual way; video resources focus on dynamic skill demonstrations, such as how to use an inhaler, breathing exercises, and postural drainage techniques; interactive questions and answers are designed with knowledge test questions of stepped difficulty to assess the patient's understanding of what they have learned. All resources are divided into three levels: basic, advanced and professional according to the complexity of the content. The basic level content is initially presented, and the system records the patient's learning behavior data, including reading time, video completion rate, question and answer accuracy and other indicators. The learning effect is analyzed through a decision tree algorithm. When the question and answer accuracy of the basic level content exceeds 80%, the advanced level content is automatically unlocked. When the advanced content test is passed, the professional level content will be opened to achieve dynamic adjustment of the education depth and build a graded education resource package.

[0050] Develop a set of self-management tools with complementary functions based on personalized diagnosis and treatment plans and treatment response indexes. The electronic medicine box reminder system sets precise medication time points and dosage reminders according to the patient's medication plan, records actual medication use, and calculates medication compliance indicators; when three consecutive missed or wrong medications are taken, the intelligent reminder upgrade mechanism is triggered to ensure medication regularity through stronger reminder sounds, contact with relatives, etc. The symptom self-monitoring diary is designed with a simple daily recording interface to guide patients to regularly evaluate changes in core symptoms, automatically generate symptom trend charts, and identify abnormal fluctuations. The respiratory training guidance system provides personalized respiratory muscle training plans and whole-body aerobic exercise recommendations based on the patient's lung function status and physical fitness level, combines sensor feedback to evaluate training effects, and dynamically adjusts the difficulty. The environmental risk warning system integrates real-time meteorological data, air quality index, and patient personal sensitivity factors to predict high-risk exposures and issue warnings 12-24 hours in advance, while providing corresponding protective measures.

[0051] According to the patient's technology acceptance, health education content and self-management tools are integrated into multiple technology platforms to achieve seamless access. Technology acceptance is determined by questionnaire assessment combined with age characteristics and actual equipment usage habits, and is divided into three levels: low, medium, and high. For patients with high technology acceptance, functional mobile applications are provided, which include all educational resources and management tools, support complex interactions and data visualization; for patients with medium technology acceptance, simplified mobile applications or web version entrances are provided to highlight core functions and simplify operation procedures; for patients with low technology acceptance, basic monitoring is mainly achieved through smart wearable devices, and auxiliary management is carried out with paper materials and telephone follow-ups. Data is shared synchronously between platforms to ensure that patients can obtain consistent health management services at different entrances. The interactive interface design follows the principle of intuitiveness, using large fonts, high-contrast colors, and icons that conform to cognitive habits to lower the threshold for use and improve patient acceptance.

[0052] Comprehensive data collection is carried out on the actual operation of the system to form the basis for iterative optimization. Patient operation behavior data, including frequency of use, function access path, duration of stay and depth of interaction, are recorded in real time through point-of-use technology; medical staff operation behavior data focuses on the use and adjustment frequency of clinical decision support functions. Feedback is collected through structured questionnaires and free text forms, covering the evaluation of system usability, content practicality and functional integrity. The diagnosis and treatment effect evaluation data integrates changes in disease control indicators, changes in acute exacerbation frequency and changes in quality of life scores to quantify the effect of system intervention. The large model performance indicators record the accuracy of diagnostic recommendations, the rationality of treatment plans and the satisfaction of patients with answers to questions, and evaluate the performance of the core components of artificial intelligence. These multidimensional data are cleaned, standardized and correlated to form a structured system optimization data set, which provides a basis for the next iteration.

[0053] Multi-level iterative updates are performed based on the system optimization data set. The knowledge base of the respiratory chronic disease model is updated through incremental learning, adding new clinical guidelines, research findings and typical cases to the training corpus to expand the model's knowledge. Algorithm tuning uses error analysis to identify case types with large model prediction deviations, and specifically strengthens the weight of such samples in the training set to improve accuracy in specific scenarios. System process node optimization is based on user behavior data to identify links with long operation times or high error rates, and simplifies operations through process reengineering or auxiliary prompts; resource allocation optimization adjusts the computing resource allocation ratio according to the frequency and importance of each functional module to improve system response speed. The user interface and interaction logic optimization adopts the A / B testing method to compare the user acceptance and operation efficiency of different design solutions and select the optimal solution. The entire iterative process is carried out in a closed loop of "collecting data-analyzing problems-implementing optimization-effect evaluation" to ensure that the system continues to evolve and adapt to changing clinical needs and technical environments.

[0054] The above describes the intelligent diagnosis and treatment management method for chronic respiratory diseases based on a large model in the embodiment of the present application. The following describes the intelligent diagnosis and treatment management system for chronic respiratory diseases based on a large model in the embodiment of the present application. Please refer to Figure 2 In the embodiment of the present application, an embodiment of the intelligent diagnosis and treatment management system for chronic respiratory diseases based on a large model includes: The acquisition module 201 is used to collect and process multi-dimensional data of patients with chronic respiratory diseases to obtain a structured patient data set; An input module 202 is used to input the structured patient data set and the respiratory disease professional knowledge base into a dual encoding pre-training and contrastive learning fine-tuning framework to obtain a large respiratory chronic disease model and a respiratory chronic disease knowledge graph; The calculation module 203 is used to perform semantic analysis and multi-dimensional calculation on the patient's symptom description and the evaluation scale answer based on the respiratory chronic disease big model and the respiratory chronic disease knowledge graph to generate a patient risk stratification evaluation result; A generation module 204 is used to generate a personalized diagnosis and treatment plan and a matching score based on the patient risk stratification assessment result, using the respiratory chronic disease big model to retrieve clinical guideline rules and combining individual characteristics of the patient; A construction module 205 is used to construct a multi-level medical institution collaboration framework and perform dynamic efficacy monitoring based on the personalized diagnosis and treatment plan and the matching score, and generate efficacy indicators and treatment response indexes; The customization module 206 is used to customize patient health education content and self-management tools according to the efficacy indicators and the treatment response index, and collect system operation data for model update and process optimization.

[0055] Through the collaborative cooperation of the above components, a structured patient data set is formed through multi-dimensional data collection and processing, which realizes the comprehensive acquisition and standardized processing of information related to respiratory chronic diseases; secondly, the dual encoding pre-training and contrastive learning fine-tuning framework is used to process the structured patient data set and the professional knowledge base of respiratory diseases. The generated respiratory chronic disease model and respiratory chronic disease knowledge graph have deep medical understanding and professional reasoning capabilities, which are significantly better than traditional machine learning models. They can understand complex medical contexts and extract key clinical features from them; thirdly, semantic analysis and multi-dimensional calculation of patient symptom descriptions and evaluation scale answers are performed based on the respiratory chronic disease model and knowledge graph. The generated patient risk stratification assessment results are highly accurate, can capture subtle differences in disease characteristics, and achieve accurate stratification, thereby optimizing the allocation of medical resources; Fourth, based on the results of patient risk stratification assessment, the respiratory chronic disease big model is used to retrieve clinical guideline rules and generate personalized diagnosis and treatment plans and matching scores based on the individual characteristics of patients, which fully considers the individual differences of patients, so that the treatment plan is both in line with the principles of evidence-based medicine and highly personalized, thereby improving the accuracy and compliance of treatment; Fifth, the multi-level medical institution collaboration framework and dynamic efficacy monitoring system constructed based on personalized diagnosis and treatment plans and matching scores have broken the medical information island, achieved the rational allocation of high-quality medical resources and continuous medical services, and reduced repeated examinations and treatment errors; patient health education content and self-management tools customized according to efficacy indicators and treatment response indexes, combined with the continuous collection of system operation data, model updates and process optimization, have formed a closed-loop quality improvement system, enabling the system to continuously evolve and improve. It is particularly worth emphasizing that the large model algorithm used in this solution has made great contributions to the field of diagnosis, treatment and management of chronic respiratory diseases. Its contextual understanding ability enables the model to extract key clinical information from unstructured medical texts, and the knowledge-enhanced reasoning mechanism realizes the automated support of complex medical decision-making. The multimodal data fusion algorithm can integrate patient information from different sources to form a comprehensive view, and the adaptive learning characteristics enable the system to continuously optimize diagnosis and treatment strategies based on new data. These specific algorithm features jointly build a highly specialized intelligent decision-making support system for respiratory diseases, which greatly improves the efficiency and accuracy of chronic respiratory disease management.

[0056] above Figure 2 The intelligent diagnosis and treatment management system for chronic respiratory diseases based on a big model in the embodiment of the present invention is described in detail from the perspective of modular functional entities. The intelligent diagnosis and treatment management device for chronic respiratory diseases based on a big model in the embodiment of the present invention is described in detail from the perspective of hardware processing.

[0057] Figure 33 is a structural diagram of a large-scale model-based intelligent diagnosis and treatment management device for respiratory chronic diseases provided by an embodiment of the present invention. The large-scale model-based intelligent diagnosis and treatment management device 300 for respiratory chronic diseases may have relatively large differences due to different configurations or performances, and may include one or more processors (central processing units, CPU) 310 (for example, one or more processors) and a memory 320, and one or more storage media 330 (for example, one or more mass storage device terminals) storing application programs 333 or data 332. Among them, the memory 320 and the storage medium 330 may be short-term storage or permanent storage. The program stored in the storage medium 330 may include one or more modules (not shown in the figure), and each module may include a series of instruction operations in the large-scale model-based intelligent diagnosis and treatment management device 300 for respiratory chronic diseases. Furthermore, the processor 310 may be configured to communicate with the storage medium 330, and execute a series of instruction operations in the storage medium 330 on the large-scale model-based intelligent diagnosis and treatment management device 300 for respiratory chronic diseases, so as to implement the steps of the large-scale model-based intelligent diagnosis and treatment management method for respiratory chronic diseases.

[0058] The large model-based intelligent diagnosis and treatment management device for chronic respiratory diseases 300 may also include one or more power supplies 340, one or more wired or wireless network interfaces 350, one or more input and output interfaces 360, and / or one or more operating systems 331, such as Windows Serve, Mac OS X, Unix, Linux, FreeBSD, etc. Those skilled in the art will appreciate that Figure 3 The structure of the large-model-based intelligent diagnosis, treatment and management device for chronic respiratory diseases shown does not constitute a limitation on the large-model-based intelligent diagnosis, treatment and management device for chronic respiratory diseases provided by the present invention, and may include more or fewer components than shown in the figure, or a combination of certain components, or a different arrangement of components.

[0059] The present invention also provides a computer-readable storage medium, which may be a non-volatile computer-readable storage medium or a volatile computer-readable storage medium. Instructions are stored in the computer-readable storage medium. When the instructions are executed on a computer, the computer executes the steps of the large model-based intelligent diagnosis and treatment management method for chronic respiratory diseases.

[0060] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the above-described systems, systems and units can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0061] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention is essentially or partly contributed to the prior art or all or part of the technical solution can be embodied in the form of a software product, and the computer software product is stored in a storage medium, including several instructions for enabling a large model-based intelligent diagnosis and treatment management device for chronic respiratory diseases (which can be a personal computer, server, or network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM), random access memory (RAM), disk or optical disk, etc., and other media that can store program codes.

[0062] The above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit the same. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that the technical solutions described in the aforementioned embodiments may still be modified, or some of the technical features may be replaced by equivalents. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for intelligent diagnosis and treatment of chronic respiratory diseases based on a large model, characterized in that: The method comprises: Collect and process multi-dimensional data of patients with chronic respiratory diseases to obtain structured patient data sets; Input the structured patient dataset and the professional knowledge base of respiratory diseases into the dual encoding pre-training and contrastive learning fine-tuning framework to obtain a large model of chronic respiratory diseases and a knowledge graph of chronic respiratory diseases; Based on the respiratory chronic disease big model and the respiratory chronic disease knowledge graph, semantic analysis and multi-dimensional calculation are performed on the patient's symptom description and the evaluation scale answer to generate the patient risk stratification evaluation result; According to the patient risk stratification assessment results, the respiratory chronic disease big model is used to retrieve clinical guideline rules and combine the individual characteristics of the patient to generate a personalized diagnosis and treatment plan and matching score; Based on the personalized diagnosis and treatment plan and the matching score, a multi-level medical institution collaboration framework is constructed and dynamic efficacy monitoring is performed to generate efficacy indicators and treatment response indexes; According to the efficacy indicators and the treatment response index, patient health education content and self-management tools are customized, and system operation data is collected for model updating and process optimization.

2. The large model-based intelligent diagnosis and treatment management method for chronic respiratory diseases according to claim 1 is characterized in that: The multi-dimensional data of patients with chronic respiratory diseases are collected and processed to obtain a structured patient data set, including: Through clinical input from medical institutions, self-reporting from patients’ mobile terminals, and smart wearable devices, basic patient information, medical history, medication status, symptom assessment, and physiological monitoring data are collected to form a multi-source heterogeneous data set; Standardize and normalize the structured data in the multi-source heterogeneous data set, perform word segmentation, stop word removal and entity standardization operations on the text data, perform noise reduction, missing value interpolation and outlier processing on the time series data, and generate a preprocessed data set; Annotate the preprocessed data set according to the respiratory system medication classification method, establish a unique patient identifier, and generate an initial annotated data set; Organizing the initial annotated data set into a tree-like hierarchical structure, associating the top-level patient basic information, the second-level disease classification, and the bottom-level symptom monitoring indicators with the corresponding patient files, to obtain a hierarchically associated data set; Performing integrity check, consistency verification and outlier marking on the hierarchical association data set to generate a quality-controlled data set; The quality-controlled dataset is stored in a secure database to form a structured patient dataset.

3. The large model-based intelligent diagnosis and treatment management method for chronic respiratory diseases according to claim 1 is characterized in that: The structured patient dataset and the respiratory disease professional knowledge base are input into the dual encoding pre-training and contrastive learning fine-tuning framework to obtain a large respiratory chronic disease model and a respiratory chronic disease knowledge graph, including: Integrate clinical guidelines for respiratory diseases, evidence-based medicine literature, expert consensus documents, typical case collections, and drug interaction databases to build a professional knowledge base for respiratory diseases; Input the structured patient dataset and the medical text corpus in the respiratory disease professional knowledge base into the Transformer encoder structure at the same time, perform multiple rounds of iterative training using a masked language model and a next sentence prediction task, and obtain initial large model parameters; Based on the initial large model parameters, a comparative learning framework is constructed to bring similar case representation vectors closer together and different case representation vectors farther apart, and the model parameters are optimized by minimizing the comparative loss function to generate a large model of chronic respiratory diseases; Performing entity extraction on the structured patient data set and the respiratory disease professional knowledge base, identifying diseases, symptoms, drugs, examinations, and risk factors as medical concepts, and obtaining a medical entity set; Extracting relationships from the medical entity set, determining disease-symptom, disease-drug, drug-drug, symptom-severity, and examination-reference value as entity associations, and forming a relationship triplet set; The medical entity set and the relationship triple set are subjected to knowledge representation learning through a graph embedding algorithm to construct a knowledge graph of chronic respiratory diseases, and are associated with the chronic respiratory disease large model through entity linking technology to support knowledge-enhanced reasoning.

4. The large model-based intelligent diagnosis and treatment management method for chronic respiratory diseases according to claim 1 is characterized in that: Based on the respiratory chronic disease big model and the respiratory chronic disease knowledge graph, semantic analysis and multi-dimensional calculation are performed on the patient's symptom description and the evaluation scale answer to generate the patient risk stratification evaluation result, including: Guide the patient to describe the current symptom experience through a natural language interactive interface, input the patient's natural language input into the respiratory chronic disease large model, extract the symptom duration, severity, inducement, and relief factors as key symptom attributes, and generate a symptom entity matrix; Guide the patients to complete the dyspnea score, cough frequency score, sputum characteristics score, activity limitation score, sleep impact score, and medication compliance score as a structured symptom assessment questionnaire to obtain the questionnaire numerical vector; The symptom entity matrix and the questionnaire numerical vector are subjected to feature fusion, and the lung function index and blood oxygen saturation value in the basic physiological monitoring data are combined to construct a multidimensional feature space to form a comprehensive symptom representation of the patient; The dyspnea, cough and sputum in the comprehensive symptom representation of the patient are assigned weight coefficients as core symptoms, and the severity value of each symptom is multiplied by the corresponding weight coefficient by a weighted calculation method and the sum is calculated to obtain the symptom burden index; Conduct time series analysis on recent symptom fluctuation trends, acute attack risks, medication compliance values, and environmental factors, generate a disease stability index through a stability calculation model, and calculate a comprehensive risk score based on the severity of the underlying disease, comorbidities, and social support scores; A threshold interval is set according to the comprehensive risk score, patients with a risk score lower than a first threshold are classified as a low risk level, patients with a risk score between the first threshold and the second threshold are classified as a medium risk level, and patients with a risk score higher than the second threshold are classified as a high risk level, thereby generating a patient risk stratification assessment result.

5. The large model-based intelligent diagnosis and treatment management method for chronic respiratory diseases according to claim 1 is characterized in that: According to the patient risk stratification assessment results, the respiratory chronic disease large model is used to retrieve clinical guideline rules and combined with the patient's individual characteristics, Generate personalized treatment plans and matching scores, including: The patient risk stratification assessment results are input into the respiratory chronic disease model to identify the patient's disease classification. When the patient has chronic obstructive pulmonary disease, the patient is divided into four groups of A / B / C / D according to the degree of airflow limitation and symptom score. When the patient has asthma, the patient is marked based on the control level and severity to generate a disease classification result; Match the disease classification results with clinical guideline rules, retrieve a mild medication regimen when the patient belongs to group A, retrieve a moderate medication regimen when the patient belongs to group B, and retrieve a severe medication regimen when the patient belongs to group C or group D, to obtain a basic treatment framework; Compare and analyze the basic treatment framework with the patient's previous treatment records and adverse drug reaction history to generate a preliminary personalized drug plan; Based on the preliminary personalized drug regimen, formulate drug administration information, including preferred drug category combination, specific drug selection, administration route, usage and dosage, and treatment course design, and generate a drug treatment regimen; In combination with the drug treatment plan, non-drug treatment recommendations are created, including respiratory rehabilitation training programs, lifestyle adjustment recommendations, mental health interventions, and disease self-management skills training. When the patient's risk level is high, an acute exacerbation emergency plan is developed to form a personalized diagnosis and treatment plan; The personalized diagnosis and treatment plan is matched and evaluated with the patient's characteristics, and the patient's previous medication response matching score, lifestyle matching score, compliance factor matching score, economic affordability matching score, and medical insurance coverage matching score are calculated to obtain a percentage matching score.

6. The large model-based intelligent diagnosis and treatment management method for chronic respiratory diseases according to claim 1 is characterized in that: According to the personalized diagnosis and treatment plan and the matching score, a multi-level medical institution collaboration framework is constructed and dynamic efficacy monitoring is performed to generate efficacy indicators and treatment response indexes, including: According to the patient risk stratification assessment results and the matching score, low-risk patients are assigned to primary medical institutions, medium-risk patients are arranged to be managed by primary medical institutions and regularly followed up by specialist hospitals, and high-risk patients are managed by specialist hospitals and followed up by primary medical institutions, so as to construct a hierarchical diagnosis and treatment pathway map; Connect medical institutions of different levels through an intelligent collaborative platform to share medical records, examination results, and treatment opinions, forming a medical collaborative network; For patients with chronic respiratory diseases seen by primary medical institutions, the patient data and the doctor's preliminary judgment are input into the chronic respiratory disease big model to analyze the symptom characteristics, disease progression risk and treatment difficulty, and generate diagnosis and treatment decision recommendations; Conduct multi-dimensional symptom monitoring on patients who start treatment, record changes in core symptom scores, changes in physiological indicators, medication compliance data, and quality of life assessment data, and build a full-dimensional monitoring data set; Performing time series analysis on the full-dimensional monitoring data set, calculating the symptom improvement rate, physiological index recovery rate, medication compliance index, and quality of life improvement, and generating multidimensional efficacy indicators; Based on the multidimensional efficacy indicators, a comprehensive score of the treatment response is calculated through a weighted aggregation algorithm, and the direction of the efficacy is determined according to the trend of the score change. When the score drops continuously for more than three monitoring cycles, the treatment plan adjustment process is triggered to form a time series treatment response index.

7. The large model-based intelligent diagnosis and treatment management method for chronic respiratory diseases according to claim 1 is characterized in that: The method of customizing patient health education content and self-management tools based on the efficacy indicators and the treatment response index, and collecting system operation data for model updating and process optimization, includes: Based on the patient's disease type, severity, cognitive level, technology acceptance and efficacy indicators, basic disease knowledge, symptom identification methods, medication guidance content, lifestyle adjustment strategies, and acute exacerbation prevention measures are screened to generate a personalized health education content library; Convert the personalized health education content library into multimedia resources in the form of text, pictures, videos, and interactive questions and answers, dynamically adjust the presentation depth according to the patient's learning progress and content understanding, and construct a graded education resource package; Based on the personalized diagnosis and treatment plan and the treatment response index, an electronic medicine box reminder system, a symptom self-monitoring diary, a breathing training guidance system, and an environmental risk warning system are developed as functional modules to form a patient self-management tool set; Integrate the health education content and the self-management tool set into mobile applications, web pages, and smart wearable devices to generate a patient-side interactive interface based on the patient's technical acceptance ability; Record and analyze the operation behaviors, feedback, diagnosis and treatment effect evaluation, and large model performance indicators of patients and medical staff during the use of the system to form a system optimization data set; Based on the system optimization data set, the knowledge base of the respiratory chronic disease model is updated and the algorithm is tuned, the efficiency analysis and resource allocation optimization of the system process nodes are performed, the user interface and interaction logic are optimized for human-computer interaction, and the system iteration and process optimization are completed.

8. An intelligent diagnosis and treatment management system for chronic respiratory diseases based on a large model, characterized in that: Used to implement the large model-based intelligent diagnosis and treatment management method for chronic respiratory diseases according to any one of claims 1 to 7, the large model-based intelligent diagnosis and treatment management system for chronic respiratory diseases includes: The acquisition module is used to collect and process multi-dimensional data of patients with chronic respiratory diseases to obtain a structured patient data set; An input module, used to input the structured patient dataset and the professional knowledge base of respiratory diseases into a dual encoding pre-training and contrastive learning fine-tuning framework to obtain a large respiratory chronic disease model and a respiratory chronic disease knowledge graph; A calculation module, for performing semantic analysis and multi-dimensional calculation on the patient's symptom description and the evaluation scale answer based on the respiratory chronic disease big model and the respiratory chronic disease knowledge graph, to generate a patient risk stratification evaluation result; A generation module is used to generate a personalized diagnosis and treatment plan and a matching score based on the patient risk stratification assessment results, using the respiratory chronic disease big model to retrieve clinical guideline rules and combining the patient's individual characteristics; A construction module is used to construct a multi-level medical institution collaboration framework and perform dynamic efficacy monitoring based on the personalized diagnosis and treatment plan and the matching score, and generate efficacy indicators and treatment response indexes; The customization module is used to customize patient health education content and self-management tools according to the efficacy indicators and the treatment response index, and collect system operation data for model updating and process optimization.

9. An intelligent diagnosis and treatment management device for chronic respiratory diseases based on a large model, characterized in that: It includes a memory and a processor, the memory stores a computer program that can be run on the processor, and the processor implements the large model-based intelligent diagnosis and treatment management method for chronic respiratory diseases as described in any one of claims 1 to 7 when executing the computer program.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the processor executes the large model-based intelligent diagnosis and treatment management method for chronic respiratory diseases as described in any one of claims 1 to 7.

Images