Emergency internal medicine clinical diagnosis and treatment system and method

Abstract

The invention discloses an emergency internal medicine clinical diagnosis and treatment system and method, and relates to the technical field of medical information processing, and the method comprises the steps: collecting and preprocessing emergency initial multi-modal diagnosis data, and generating a structured symptom label set; in combination with basic information of the patient, identifying risk priorities of symptoms and diagnosis and treatment nodes, and generating a personalized diagnosis and treatment path; checking according to the personalized diagnosis and treatment path, and collecting diagnosis and treatment feedback data of each diagnosis and treatment node; the diagnosis and treatment node task completion condition is judged according to the diagnosis and treatment feedback data, the next diagnosis and treatment node task in the personalized diagnosis and treatment path is dynamically executed, the diagnosis and treatment feedback data are collected until the personalized diagnosis and treatment path is completed, and stage diagnosis and treatment feedback data are generated; the diagnosis and treatment effect score of each stage is analyzed, and the personalized diagnosis and treatment path is dynamically optimized. Dynamic path adjustment is achieved through feedback driving, an intelligent diagnosis and treatment system with the patient state as the core and closed-loop feedback regulation is constructed, and path execution flexibility and individual adaptability are improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of medical information processing, in particular to an emergency internal medicine clinical diagnosis and treatment system and method. BACKGROUND

[0002] In the hospital emergency medical scene, the internal medicine patient's condition is complex and changes rapidly, and diagnosis and treatment needs to rely on doctors to integrate multi-modal information in a short time to make a quick judgment. The existing emergency process usually includes inquiry, image examination, inspection and detection, and diagnosis and treatment path promotion, relying on experienced human path combination. Doctors usually collect information and make decisions through a combination of film reading, auscultation, physical examination and data retrieval, supplemented by electronic medical records to realize digital management of diagnosis and treatment information. With the development of medical informatization, the use of voice recognition, image assisted diagnosis and process recording tools improves data processing efficiency and clinical decision-making ability.

[0003] In the previous method, the diagnosis and treatment nodes are mostly set based on static template matching, which is difficult to reflect the influence of individual patient characteristics on path decision-making, limiting the implementation of personalized diagnosis and treatment. At the same time, there is a lack of dynamic adjustment mechanism based on real-time feedback, and the completion or not of the diagnosis and treatment node state cannot drive path rearrangement and update, affecting the flexible response and continuous promotion of the diagnosis and treatment process, which is particularly insufficient in the high-efficiency and high-variability clinical scene of emergency. SUMMARY

[0004] In view of the above existing problems, the present application is proposed.

[0005] Therefore, the present application provides an emergency internal medicine clinical diagnosis and treatment method to solve the problems of difficulty in automatic generation of individualized path and lack of real-time feedback driving mechanism in diagnosis and treatment process.

[0006] To solve the above technical problems, the present application provides the following technical solutions:

[0007] In a first aspect, the present application provides an emergency internal medicine clinical diagnosis and treatment method, which comprises collecting emergency initial multi-modal diagnosis data and preprocessing to generate a structured symptom label set; combining the structured symptom label set and patient basic information, identifying the correlation and risk priority between the structured symptom label and the diagnosis and treatment node through a path reasoning engine, obtaining the optimal diagnosis and treatment process node combination and defining the execution order, generating an individualized diagnosis and treatment path; performing examination and treatment according to the individualized diagnosis and treatment path, collecting diagnosis and treatment feedback data of each node in the individualized diagnosis and treatment path; judging the completion of the current node task according to the diagnosis and treatment feedback data, dynamically executing the next node task in the individualized diagnosis and treatment path, and collecting the diagnosis and treatment feedback data of the next node, until the individualized diagnosis and treatment path is executed, to generate stage diagnosis and treatment feedback data; analyzing the diagnosis and treatment effect score of each stage according to the stage diagnosis and treatment feedback data, and dynamically optimizing the individualized diagnosis and treatment path.

[0008] As a preferred scheme of the emergency internal medicine clinical diagnosis and treatment method, the emergency initial multi-modal diagnosis data comprises patient condition description voice obtained by a voice acquisition device and medical image information collected from a patient related part.

[0009] The preprocessing comprises format standardization, noise elimination and semantic annotation processing.

[0010] By natural language recognition combined with image feature extraction, keyword extraction and semantic analysis are performed on the preprocessed emergency initial multi-modal diagnosis data, and a structured symptom label set is generated based on a text-image fusion mechanism.

[0011] As a preferred scheme of the emergency internal medicine clinical diagnosis and treatment method, the emergency initial multi-modal diagnosis data comprises patient condition description voice obtained by a voice acquisition device and medical image information collected from a patient related part.

[0012] The symptom name, part description, duration and manifestation in the structured symptom label set are extracted and compared with the age, gender, medical history, allergy record and basic disease information of the patient to obtain a symptom-patient adaptability matching feature set.

[0013] Based on the symptom-patient adaptability matching feature set, the path reasoning engine identifies the semantic correspondence between the structured symptom label set and the diagnosis and treatment nodes by matching the symptom-node mapping relationship, and calculates the matching degree of each diagnosis and treatment node with the structured symptom label set.

[0014] According to the matching degree, the adaptive node set of the current patient in all candidate diagnosis and treatment paths is defined, and the diagnosis and treatment nodes are sorted in combination with the patient's emergency degree, resource availability state and risk level.

[0015] The top N diagnosis and treatment nodes are taken as the optimal diagnosis and treatment nodes of the current patient, and the execution order of the diagnosis and treatment nodes is set according to the task priority to generate a personalized diagnosis and treatment path.

[0016] As a preferred scheme of the emergency internal medicine clinical diagnosis and treatment method, the emergency initial multi-modal diagnosis data comprises patient condition description voice obtained by a voice acquisition device and medical image information collected from a patient related part.

[0017] According to the sorting result, the diagnosis and treatment equipment corresponding to the tasks of the personalized diagnosis and treatment path are activated in sequence, and equipment operation parameters are pushed according to the examination type and treatment mode of each optimal diagnosis and treatment node.

[0018] In the case where the first task node is an image examination, original image data is obtained by image scanning.

[0019] Edge recognition, region segmentation, and tissue density analysis are performed on the raw image data to identify lesion areas and tissue structure abnormalities, and the data is transformed into structured diagnostic and treatment feedback data through symptom feature label matching rules.

[0020] If the current node is a laboratory testing node, the execution flow control logic controls the sampling process and calls the rapid testing equipment to complete the reading and format conversion of blood routine, biochemical indicators and infection indicators;

[0021] The execution status, response time, device feedback, and result data of the diagnosis and treatment nodes are recorded synchronously and packaged and uploaded to the path judgment process for the next step to determine whether to proceed to the next diagnosis and treatment node.

[0022] If execution occurs at any node, path progression is paused, a task exception signal is issued, and an exception handling sub-process is initiated, including rescheduling, path redirection, and manual confirmation.

[0023] As a preferred embodiment of the emergency internal medicine clinical diagnostic and treatment method described in this invention, the steps for determining the completion status of the current node task and dynamically executing the next node task in the personalized treatment path are as follows:

[0024] Identify the completion status of diagnosis and treatment tasks based on the criteria for determining the completion of diagnosis and treatment tasks;

[0025] If the current treatment node task is determined to be completed, the path determination process will activate the next treatment node to perform the treatment operation according to the execution order of the treatment nodes in the personalized treatment path.

[0026] If the current treatment node task is determined to be incomplete, the path dynamic adjustment logic is triggered, that is, the current personalized treatment path is stopped, and the treatment feedback data of the completed treatment nodes and the newly generated treatment feedback data of the current treatment node are obtained. The treatment nodes are reordered, an updated personalized treatment path is generated, and the current treatment node continues to execute the updated personalized treatment path until all treatment node tasks are completed.

[0027] As a preferred embodiment of the emergency internal medicine clinical diagnosis and treatment method described in this invention, the generation of phased diagnosis and treatment feedback data refers to the process of performing personalized diagnosis and treatment node tasks, in which the diagnosis and treatment feedback data is structurally classified according to the node time sequence, and the content of the diagnosis and treatment feedback data is converted into a unified format standard to form a phased diagnosis and treatment feedback dataset with a unified field structure and traceable association relationship.

[0028] As a preferred embodiment of the emergency internal medicine clinical diagnostic and treatment method described in this invention, the steps for analyzing the treatment effectiveness scores at each stage and dynamically optimizing the personalized treatment pathway are as follows:

[0029] Based on the phased diagnosis and treatment feedback data, the diagnosis and treatment scoring engine is invoked to score the diagnosis and treatment effect of each diagnosis and treatment node that has been executed in the personalized diagnosis and treatment path;

[0030] The scoring data of each node are archived in the order of the nodes to form a structured treatment effect scoring table;

[0031] If the score is lower than the diagnostic confidence threshold, the path optimization process is automatically triggered. The path optimization process inputs the score and treatment feedback data into the path reasoning engine to recalculate the matching degree between structured symptom labels and treatment nodes, and optimizes the personalized treatment path.

[0032] If the treatment outcome score is higher than the diagnostic confidence threshold, the current personalized treatment pathway will be retained and the treatment outcome score sheet will be archived.

[0033] Secondly, the present invention provides an emergency internal medicine clinical diagnosis and treatment system, including a data acquisition module, a path generation module, an execution control module, a path judgment module, and a path optimization module;

[0034] The data acquisition module is used to collect initial multimodal diagnostic data from the emergency department and preprocess it to generate a set of structured symptom labels.

[0035] The path generation module combines a set of structured symptom tags and basic patient information, and through the path reasoning engine, identifies the correlation and risk priority between structured symptom tags and treatment nodes, obtains the optimal combination of treatment process nodes, defines the execution order, and generates a personalized treatment path.

[0036] The execution control module is used to perform examinations and treatments according to the personalized treatment pathway and to collect treatment feedback data at each node in the personalized treatment pathway.

[0037] The path determination module is used to determine the completion status of the current node task based on the diagnosis and treatment feedback data, dynamically execute the next node task in the personalized diagnosis and treatment path, and collect the diagnosis and treatment feedback data of the next node until the personalized diagnosis and treatment path is completed and the stage diagnosis and treatment feedback data is generated.

[0038] The path optimization module is used to analyze the treatment effect scores at each stage based on the phased treatment feedback data, and dynamically optimize the personalized treatment path.

[0039] Thirdly, the present invention provides a computer device including a memory and a processor, wherein the memory stores a computer program, wherein when the computer program is executed by the processor, it implements any step of the emergency internal medicine clinical diagnostic and treatment method as described in the first aspect of the present invention.

[0040] Fourthly, the present invention provides a computer-readable storage medium having a computer program stored thereon, wherein: when the computer program is executed by a processor, it implements any step of the emergency internal medicine clinical diagnostic and treatment method as described in the first aspect of the present invention.

[0041] The beneficial effects of this invention are as follows: By combining a structured symptom tag set with basic patient information, it achieves accurate identification of the correlation and risk priority between symptoms and treatment nodes, constructs personalized treatment pathways, and improves the matching of pathway formulation. By judging the completion status of the current node task based on treatment feedback data, it dynamically executes the next node in the personalized pathway, realizing real-time adjustment and continuous advancement of the pathway. The two work together to construct a treatment process driven by patient status and controlled by a feedback loop, enhancing the flexibility and adaptability of the pathway. Attached Figure Description

[0042] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the following description of the embodiments will be briefly introduced. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0043] Fig. 1 This is a flowchart of clinical diagnostic and treatment methods in the emergency internal medicine department.

[0044] Fig. 2 This is a schematic diagram of the emergency internal medicine clinical diagnosis and treatment system.

[0045] Fig. 3 A flowchart for generating personalized treatment pathways.

[0046] Fig. 4 This is a flowchart for dynamic execution and optimization. Detailed Implementation

[0047] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings.

[0048] Many specific details are set forth in the following description in order to provide a full understanding of the invention. However, the invention may also be practiced in other ways different from those described herein, and those skilled in the art can make similar extensions without departing from the spirit of the invention. Therefore, the invention is not limited to the specific embodiments disclosed below.

[0049] Secondly, the term "one embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The phrase "in one embodiment" appearing in different places in this specification does not necessarily refer to the same embodiment, nor is it a single or selective embodiment that is mutually exclusive with other embodiments.

[0050] Reference Figs. 1-4 This is one embodiment of the present invention, which provides an emergency internal medicine clinical diagnostic and treatment method, including the following steps:

[0051] S1. Collect initial multimodal diagnostic data from the emergency department and preprocess it to generate a set of structured symptom labels;

[0052] Initial multimodal diagnostic data in the emergency department includes the patient's voice description of their condition acquired by a voice acquisition device, and medical image information of relevant parts of the patient.

[0053] Furthermore, upon the patient's arrival at the emergency reception area, the initial multimodal diagnostic data collection and preprocessing process is initiated. Two key types of data are collected: one is the patient's subjective description of their condition, recorded via voice acquisition equipment and transcribed into text using medical semantic recognition methods; the other is medical image information of the patient's affected areas, acquired by image acquisition equipment. The image acquisition equipment refers to imaging devices used in the emergency treatment process to acquire medical images of relevant patient areas, encompassing digital X-ray, CT, MRI, and ultrasound equipment. These devices can provide image representations of different structures and lesions, offering fundamental image data support for subsequent diagnostic analysis.

[0054] It should be noted that medical semantic recognition methods refer to using a medical-specific thesaurus to extract keywords and perform semantic analysis on text information, identifying the names, durations, manifestations, and locations related to symptoms, and then performing structured encoding.

[0055] Preprocessing includes format normalization, noise removal, and semantic annotation.

[0056] Furthermore, format standardization includes uniformly converting the collected text information and medical image information into the correct format, such as unifying character encoding, image resolution, and data structure.

[0057] Noise removal includes removing background noise and invalid words (e.g., interjections, non-medical content) from text information, as well as artifacts, motion artifacts, and irrelevant regions from medical image information;

[0058] Semantic annotation processing includes annotating symptom-related information fields for text information using named entity recognition and semantic role annotation methods;

[0059] For medical image information, boundary contours, tissue density, and abnormal signal regions are identified through image segmentation and feature extraction.

[0060] By combining natural language recognition with image feature extraction, keyword extraction and semantic analysis are performed on the preprocessed initial multimodal diagnostic data of the emergency department, and a set of structured symptom labels is generated based on the image-text fusion mechanism.

[0061] Furthermore, semantic association analysis and feature fusion processing are performed on text information and medical image information. By constructing cross-modal feature mapping relationships, the correspondence between key symptoms in the text information and corresponding anatomical regions in the image is identified, achieving structural alignment and semantic complementarity.

[0062] Based on the symptom label mapping rules, the fused semantic information is normalized and integrated into a multi-source consistent structured symptom label set.

[0063] It should be noted that the symptom label mapping rules were obtained through data mining and feature summarization of a large number of historical emergency cases. The rules extract the correspondence between symptom descriptions and actual treatment steps taken in the cases, constructing a semantic association rule set covering common symptom features and standard treatment steps. This set is used to achieve accurate matching between structured symptom labels and treatment steps. The actual treatment steps taken refer to all diagnostic, examination, testing, and treatment procedures actually performed by clinicians in the actual treatment process of historical cases, based on the patient's complaints and structured symptom labels.

[0064] S2. Combining the structured symptom tag set and patient basic information, the path reasoning engine identifies the correlation and risk priority between structured symptom tags and treatment nodes, obtains the optimal combination of treatment process nodes and defines the execution order, and generates a personalized treatment path.

[0065] Extract symptom names, location descriptions, durations, and manifestations from the structured symptom tag set, and compare them with the patient's age, gender, past medical history, allergy records, and underlying disease information to obtain a symptom-patient adaptive matching feature set;

[0066] Furthermore, specific diagnostic dimension information corresponding to each structured symptom label is extracted from the set of structured symptom labels, including symptom name, location of onset, duration, and specific manifestation.

[0067] The system retrieves basic patient information, including age, gender, past medical history, allergy records, and underlying disease information. All of the above information is obtained with the user's consent and used for legitimate purposes. The diagnostic dimension information and the patient's basic information are linked one by one and aligned with features. Through matching rules, such as the frequency of cross-occurrence of symptoms and past medical history, the degree of clinical relevance between symptoms and underlying diseases, and the impact of duration on acute and chronic tendencies, a symptom-patient adaptive matching feature set is formed.

[0068] It should be noted that the patient's underlying medical information includes age, gender, past medical history, allergy history, chronic diseases, organ dysfunction, immunodeficiency, specific constitution, history of infection, and history of surgery and trauma.

[0069] Based on the symptom-patient adaptive matching feature set, the path reasoning engine identifies the semantic correspondence between the structured symptom label set and the diagnosis and treatment nodes by matching the symptom-node mapping relationship, and calculates the matching degree between each diagnosis and treatment node and the structured symptom label set.

[0070] Furthermore, symptom-node mapping rules are defined. These rules refer to a system of rules used to establish semantic relationships between structured symptom labels and corresponding diagnostic nodes. The construction process includes two core steps: data-driven analysis and medical knowledge summarization.

[0071] Data-driven analysis includes extracting symptom descriptions recorded during the initial patient visit and subsequent diagnostic and treatment procedures based on a large amount of historical real-life emergency internal medicine case data. Through statistical analysis techniques, it aims to uncover high-frequency co-occurrence relationships, sequential patterns, and path dependencies between symptom tags (such as "chest pain," "dyspnea," and "persistent fever") and diagnostic and treatment nodes (such as "electrocardiogram," "chest CT," and "inflammatory marker detection"), and construct preliminary symptom-node matching rules.

[0072] Medical knowledge summarization refers to the process of verifying the medical rationality and supplementing the semantics based on the initial symptom-node relationship, and further adding conditions such as diagnostic indications, contraindications, and the urgency of the condition to form a multi-dimensional symptom-node matching rule.

[0073] The path reasoning engine is built on a diagnostic and treatment knowledge graph, which consists of multiple heterogeneous entity nodes and multi-dimensional relational edges. Among them, the entity nodes include structured symptom labels, diagnostic and treatment nodes, and patient symptoms, and the edges represent semantic associations between different entities, such as "symptom-pointing-node", "node-depending-node", and "patient symptoms-adapting-node".

[0074] By using a diagnostic knowledge graph, based on the static matching of "symptom-diagnosis node", a complete path reasoning network can be constructed by integrating the dynamic adaptation of "patient symptoms-diagnosis node" and the task dependency relationship of "node-node".

[0075] The reasoning process employs a weighted path search algorithm (such as heuristic graph traversal based on matching score and priority). Starting from structured symptom labels, it jointly evaluates the set of matching diagnosis and treatment nodes, scores the importance and feasibility of diagnosis and treatment nodes based on patient symptoms, and dynamically constructs the optimal personalized path that meets the current clinical goals.

[0076] By combining multidimensional parameters (such as patient age, gender, underlying diseases, and past medical history) in the symptom-patient adaptive matching feature set, semantic adaptability assessment is performed on the diagnosis and treatment nodes to measure the clinical applicability and priority of each diagnosis and treatment node in the current patient state.

[0077] The path reasoning engine calculates the matching score between structured symptom labels and each diagnosis and treatment node based on semantic relevance and adaptability. The matching score function is as follows:

[0078]

[0079] Where M(j) represents the matching score between treatment node j and the structured symptom label set, used to evaluate the suitability of the node, j represents the j-th treatment node, and t i Let sim(t) represent the i-th structured symptom label, and n represent the total number of labels in the current set of structured symptom labels. i ,j) represents the structured symptom label t i The semantic similarity score between the diagnosis node j and the treatment node j.

[0080] Semantic similarity scores are obtained by calculating the degree of semantic matching between structured symptom labels and diagnostic nodes.

[0081] Specifically, the first step is to structure symptom labels t i The keywords contained in the data (such as symptom name, duration, location of onset, and manifestation) are vectorized, and the medical operation description corresponding to the diagnosis node j is also converted into word vectors.

[0082] By using a cosine similarity calculation method based on semantic embedding, the closeness between two semantic representations is evaluated, thus obtaining a similarity score between 0 and 1.

[0083] Define the set of suitable nodes for the current patient in all candidate treatment paths based on the degree of matching, and sort the treatment nodes by combining the patient's urgency, resource availability and risk level;

[0084] Furthermore, after obtaining the matching degree score of the treatment nodes, the candidate nodes are ranked by multiple factors, taking into account the urgency of the current patient, the resource availability of each node, and the risk level, to obtain the treatment priority score. The treatment priority score formula is as follows:

[0085] S(j)=W1·M(j)+W2·R(j)+W3·E+W4·(1-L(j))

[0086] Where S(j) represents the treatment priority score of treatment node j; R(j) represents the current resource availability score of treatment node j; E represents the urgency of the patient; L(j) represents the potential treatment risk level of treatment node j; W1 represents the weight coefficient of the matching degree score in the treatment priority score; W2 represents the weight coefficient of the resource availability score in the treatment priority score; W3 represents the weight coefficient of the patient urgency in the treatment priority score; W4 represents the weight coefficient of the potential treatment risk suppression term 1-L(j) in the treatment priority score; W1+W2+W3+W4=1, and the value range of each weight is in [0,1].

[0087] It should be noted that the resource availability score is calculated using the following formula:

[0088]

[0089] Where n represents the number of resource types involved in the j-th node (e.g., equipment, doctors, consumables); k represents the k-th resource type; ω jk This represents the dependency weight of the j-th treatment node on the k-th type of resource (the value ranges from 0 to 1, and ∑...). j ω ij =1); ηk represents the penalty coefficient for the current tension of the k-th type of resource, defined as: η k =1-ρ k , where ρ k Resource load factor (usage / total capacity); A jk This represents the current availability of the k-th type of resource, with a value ranging from 0 to 1, and can be calculated using the following formula:

[0090]

[0091] in, This indicates the current number of available resources (such as the number of idle devices or the number of doctors that can be scheduled).

[0092] This represents the amount of resources required to execute the diagnosis node j; ∈ represents a small constant (such as 1e-6) to prevent the denominator from being zero.

[0093] Resource availability scores are used to assess the availability of resources required by each diagnostic and treatment node in the current healthcare system.

[0094] The potential diagnosis and treatment risk level refers to the probability of adverse events that may occur at each diagnosis and treatment node in the current patient situation. It is obtained by comprehensively judging the intensity of medical intervention at the diagnosis and treatment node itself, the complexity of the clinical operation involved, and the individual health status of the patient. The calculation formula is as follows:

[0095]

[0096] G represents the risk event category (such as "complications", "equipment failure", "rapid deterioration of condition"), which can be listed according to historical data and specifications; ω j,g P represents the weight of the g-th type of risk event at node j; j,g S represents the probability of the g-th type of risk event occurring at node j; j,g This indicates the severity of the g-th type of risk event at node j.

[0097] It should be noted that the risk categories were determined based on historical data analysis and clinical guidelines, such as: adverse reactions / complications; equipment or operational malfunctions; sudden changes / severity of the condition.

[0098] Risk event weights: uniformly assigned ω j,k =1 / K;

[0099] The probability of occurrence is calculated by combining the frequency statistics of adverse reactions, complications and emergency referrals at each treatment node in historical data, and assigning each treatment node a grade score representing potential clinical risk.

[0100] The calculation method affecting the score is to extract key features such as symptom severity, duration and comorbidities, and assign values ​​based on the clinical operation complexity and intervention intensity of each diagnosis and treatment node.

[0101] The patient's urgency score is derived from a comprehensive assessment of their chief complaint, physiological indicators, and past medical history upon admission to the emergency department. Keywords such as "severe," "sudden onset," and "loss of consciousness" are extracted from structured symptom tags to determine symptom severity. Simultaneously, real-time vital sign data (such as heart rate, blood pressure, respiratory rate, and blood oxygen saturation) are read to assess the presence of life-threatening conditions. This is further combined with the patient's high-risk past medical history (such as myocardial infarction, stroke, and asthma) to quantify the urgency level. Finally, a weighted score is used to obtain a numerical value representing the priority of the patient's emergency treatment.

[0102] The top N treatment nodes are selected as the optimal treatment nodes for the current patient, and the execution order of the treatment nodes is set according to the sorting to generate a personalized treatment path.

[0103] Furthermore, the top N treatment nodes ranked by treatment priority score are selected as the optimal node set for the current patient, and the execution order is set according to the ranking to generate a personalized treatment path that is highly adaptable, risk-controllable, and resource-coordinated.

[0104] S3. Conduct examinations and treatments according to the personalized treatment pathway, and collect treatment feedback data at each node in the personalized treatment pathway;

[0105] Execute personalized treatment pathways, activate the corresponding diagnostic and treatment devices according to the sorting results, and push device operation parameters based on the examination type and treatment method of each optimal diagnostic and treatment node;

[0106] Furthermore, based on the personalized treatment pathway, various examinations and treatments are scheduled sequentially according to the determined treatment nodes. Upon reaching each treatment node, the matching treatment equipment is automatically activated, and standardized equipment operation parameters are pushed according to the examination type corresponding to the treatment node, ensuring that the operation process accurately corresponds to the individualized task.

[0107] It should be noted that the types of examinations include: medical imaging examinations (including X-ray, CT, MRI, and ultrasound) used to identify structural abnormalities, lesion boundaries, and density distribution of tissues and organs; laboratory tests (including complete blood count, complete biochemistry panel, infection markers, coagulation function, and liver and kidney function tests), focusing on obtaining quantitative analysis results of body fluid components to identify inflammatory responses, organ damage, or metabolic abnormalities; physiological function monitoring (including electrocardiogram, blood oxygen saturation, blood pressure, and respiratory rate) to assess changes in the patient's vital signs and key physiological parameters and determine the urgency of the condition; special etiological examinations (including respiratory pathogen screening, enterovirus detection, and blood culture) used for rapid identification of acute infectious diseases; and bedside rapid diagnostic examinations (including bedside ultrasound, fingertip blood glucose testing, and arterial blood gas analysis), providing rapid and immediate response and suitable for critical care scenarios.

[0108] In a scenario where the first task node is image inspection, raw image data is obtained through image scanning;

[0109] Image processing algorithms are used to perform edge recognition, region segmentation, and tissue density analysis on the raw image data to identify lesion areas and tissue structure abnormalities, and then transform them into structured diagnostic and treatment feedback data through symptom feature label matching rules.

[0110] Furthermore, if the first node is an image inspection node, an image acquisition operation is performed to obtain raw image data through an imaging device. The raw image data then undergoes a processing flow of edge recognition, tissue region segmentation, density distribution calculation, and texture feature analysis to identify potential lesions and abnormal structures.

[0111] Image analysis features (including edge morphology, tissue region shape, density distribution, and texture characteristics) are encoded into multidimensional feature vectors. By comparing the similarity between these multidimensional feature vectors and the encoded semantic features in a structured symptom label set, highly matching labels are identified. Corresponding diagnostic keywords are then extracted from the matching labels as the symptom representation of the image content. Based on symptom feature label matching rules, the image features are transformed into structured diagnostic feedback data.

[0112] If the current node is a laboratory testing node, the execution flow control logic controls the sampling process and calls the rapid testing equipment to complete the reading and format conversion of blood routine, biochemical indicators and infection indicators;

[0113] The execution status, response time, device feedback, and result data of the diagnosis and treatment nodes are recorded synchronously and packaged and uploaded to the path judgment process for the next step to determine whether to proceed to the next diagnosis and treatment node.

[0114] If execution occurs at any node, path progression is paused, a task exception signal is issued, and an exception handling sub-process is initiated, including rescheduling, path redirection, and manual confirmation.

[0115] Furthermore, if the first node in the personalized treatment pathway is an image examination node, then the image data acquisition and analysis process begins.

[0116] First, an image acquisition process is performed, using medical imaging equipment (such as CT, MRI, or ultrasound) to obtain raw image data of the patient's affected area. The raw image data will then undergo the following processing steps:

[0117] Edge recognition processing, for example: using the Sobel operator to perform edge detection on the image and extract the shape features of tissue boundaries and lesion contours;

[0118] Tissue region segmentation, for example: based on the U-Net semantic segmentation network, pixel-level annotation and region division are performed on different tissue structures (such as organs, lesions, cavities) in the image;

[0119] Density distribution calculation, for example: statistical analysis of pixel grayscale values ​​within a segmented region to generate a distribution histogram or statistical features of mean and variance that reflect tissue density characteristics;

[0120] Texture feature analysis, for example: extracting texture features based on the gray-level co-occurrence matrix method to characterize the granularity, directionality and repeatability of the organizational structure.

[0121] Through the above steps, a set of image feature vectors is finally obtained, which consists of four types of sub-vectors: edge morphology features, region shape features, density statistics features, and texture structure features. Subsequently, the image feature vectors are matched with the semantic feature vectors corresponding to each label in the structured symptom label set (e.g., by calculating cosine similarity), and the symptom labels with higher matching degrees are selected.

[0122] The corresponding diagnostic keywords (such as "pulmonary nodules", "increased parenchymal density", "unclear boundaries") are extracted from the selected symptom labels and used as the symptom expression results of the image content. Finally, based on the symptom feature label matching rules, the symptom expression results are converted into standardized fields and encoded to form structured diagnosis and treatment feedback data, which are used for subsequent diagnosis and treatment node status judgment and path advancement process.

[0123] When the diagnostic pathway enters the laboratory testing node, the current node type is identified, and pathway progression is paused to ensure that the node has exclusive access to resources. The process control logic initiates the sampling process, automatically identifying the required sample type (e.g., venous blood, serum, finger prick blood) based on the task configuration, and controlling the sampling equipment to perform the sampling operation. After sampling is completed, rapid testing equipment is invoked to perform blood routine, biochemical, and infectious disease tests. During the testing process, the raw equipment results are automatically read and formatted, standardized to a uniform structure (e.g., HL7, FHIR).

[0124] It should be noted that rapid testing equipment refers to in vitro diagnostic equipment that can automatically complete blood, biochemical, and infection tests in a short time and support standard structured data output, including blood analyzers, biochemical analyzers, immunoassay analyzers, and portable POCT devices.

[0125] During the execution of the diagnosis and treatment nodes, key data, including node status, response time, equipment feedback information, and test results, are recorded in real time. The key data is packaged and uploaded synchronously to the path judgment process to determine whether the current diagnosis and treatment path meets the conditions for entering the next node.

[0126] If any anomaly occurs during node execution (such as sampling failure, detection interruption, data loss, or equipment failure), path advancement will be immediately suspended, and a task anomaly signal will be issued. Simultaneously, an anomaly handling sub-process will be automatically initiated, performing rescheduling, path redirection, and submission for manual confirmation. All information regarding the anomaly handling process will be fully recorded and supported for auditing and traceability.

[0127] It should also be noted that the path judgment process refers to the discrimination mechanism used to dynamically evaluate the completion status of the treatment node tasks and determine whether it is possible to proceed to the next treatment node during the execution of the personalized treatment path.

[0128] Specifically, this includes parsing the structured diagnostic feedback data (such as test results, response time, and execution status) uploaded at the current diagnostic node;

[0129] The system determines whether the diagnostic task has been completed based on the criteria for completion. If the criteria are met, a "node completed" signal is issued and the path is advanced. Otherwise, an exception handling mechanism is triggered.

[0130] The exception handling sub-process refers to the set of fault tolerance and repair logic that is automatically activated when a task fails or is interrupted during the execution of a diagnosis and treatment node.

[0131] Specifically, this includes, but is not limited to, the following processing methods: Rescheduling: rearranging the execution of the same diagnosis and treatment nodes (such as resampling or re-examination); Path jumping: recalculating path priority based on feedback from completed nodes and skipping abnormal nodes; Manual confirmation: submitting the current status to medical staff for intervention or judgment; Abnormal recording: all abnormal states and processing actions will be fully recorded for subsequent auditing and tracing and model correction training.

[0132] It should be noted that the symptom feature label matching rules are defined based on structured diagnosis and treatment data from a large number of historical emergency cases, using a method that integrates data mining and clinical knowledge.

[0133] Specifically, firstly, based on structured data from a large number of historical emergency room cases, statistical analysis was conducted on the relationship between symptom feature labels and actual treatment nodes. The statistical analysis focused on the co-occurrence frequency, temporal order, and combination patterns of structured symptom labels and treatment nodes in the treatment pathway, identifying representative high-frequency co-occurrence structures and typical treatment pathways;

[0134] Based on the standardized information in clinical guidelines, recommended examinations, tests, and treatments for various symptoms are mapped to supplement and standardize the label-node pairs in data-driven results. By integrating data patterns with clinical knowledge, the semantic relationships between structured symptom labels and diagnostic / treatment nodes are structurally summarized, clarifying the association strength and fit boundaries.

[0135] Ultimately, a set of standardized matching rules is formed to represent the priority association and semantic fit between each type of structured symptom label and the diagnosis and treatment node under specific patient characteristics, serving as the basis for personalized path construction.

[0136] Structured diagnostic and treatment feedback data includes: node execution status information, including whether each diagnostic and treatment node is completed, completion time, execution duration, and abnormal interruption; equipment operation and feedback information, including operating parameters of examination and treatment equipment, execution logs, images, and processing summaries of raw test results; examination or test result data, including numerical data and text results of image recognition output (such as edge structure, density distribution, lesion area) and laboratory test indicators (such as blood routine, biochemistry, and infection indicators) converted to standard formats; clinical response information, including the degree of symptom relief after treatment, complication response, and changes in re-examination indicators; path execution association information, including the logical association between this node and its predecessor and successor nodes, the status of execution condition satisfaction, and scheduling signal records; and doctor input and remarks, including records of manual intervention suggestions, correction opinions, and explanations of special circumstances.

[0137] S4. Based on the diagnosis and treatment feedback data, determine the completion status of the current node task, dynamically execute the next node task in the personalized diagnosis and treatment path, and collect the diagnosis and treatment feedback data of the next node until the personalized diagnosis and treatment path is completed, and generate phased diagnosis and treatment feedback data.

[0138] Based on the diagnosis and treatment feedback data, the status of the diagnosis and treatment task completion is identified according to the judgment rules for diagnosis and treatment task completion.

[0139] Furthermore, during the execution of personalized treatment pathways, treatment feedback data at each treatment node is continuously collected. Based on the treatment task completion judgment rules, the treatment feedback data is analyzed and judged, including: whether the predetermined examination and treatment tasks have been completed, whether the feedback data meets the completion conditions set for the treatment node, whether there are any abnormal interruptions, and whether key results are missing.

[0140] It should be noted that the criteria for determining whether a diagnostic or treatment task has been completed refer to a set of standard conditions used to judge whether a particular diagnostic or treatment task has been completed. These criteria are comprehensively formulated based on medical clinical standards and equipment operating procedures, and mainly cover the achievement status of task objectives, the validity of key data, and the completeness of the execution process at each node.

[0141] Specifically, the criteria for determining whether a treatment task has been completed include the following categories:

[0142] Data completeness determination, for example: the image inspection node must generate a complete and readable image and pass quality control (image clarity, resolution) verification; the laboratory testing node must return the test results of all indicators and meet the data format specifications.

[0143] Key indicator determination: Define diagnostic and treatment goals based on the type of diagnosis and treatment node, such as: whether biochemical indicators have been obtained, whether specific treatments have been completed, and whether symptoms have been confirmed through image recognition.

[0144] The execution process status is determined, such as whether the equipment responds normally, whether the task is interrupted, whether an error code is generated, and whether the completion is manually confirmed, to ensure that the execution process of the diagnosis and treatment node tasks is complete and without abnormalities.

[0145] It should also be noted that the completion conditions set based on the type of diagnosis and treatment node refer to the criteria used to determine whether the task of the diagnosis and treatment node has been achieved, based on the expected goals to be achieved for different types of diagnosis and treatment nodes (such as image examination, laboratory testing, and clinical treatment).

[0146] For example: the completion condition for the image inspection node is to obtain a clear and complete image of the target area, and to complete the image structure segmentation and lesion annotation; the completion condition for the laboratory testing node is that the sample collection is completed, the corresponding test index values ​​are obtained, and all key indicators are in a state of "pending supplementation" and "abnormal termination"; the completion condition for the clinical treatment node is that the diagnostic and treatment equipment completes the set action sequence, the execution parameters are consistent with the issued instructions, and the manually entered treatment record is complete and the status is "completed".

[0147] If the current treatment node task is determined to be completed, the path determination process sends a scheduling signal to the execution process control logic according to the execution order of the treatment nodes in the personalized treatment path, and activates the next treatment node to perform the treatment operation.

[0148] If the current treatment node task is determined to be incomplete, the path dynamic adjustment logic is triggered, that is, the current personalized treatment path is stopped, and the treatment feedback data of the completed treatment nodes and the new treatment feedback data generated by the current treatment node are obtained. The treatment nodes are reordered, an updated personalized treatment path is generated, and the current treatment node continues to execute the updated personalized treatment path until all treatment node tasks are completed.

[0149] During the execution of personalized diagnosis and treatment node tasks, the diagnosis and treatment feedback data is structured and categorized according to the node time sequence, and the content of the diagnosis and treatment feedback data is converted into a unified format standard to form a phased diagnosis and treatment feedback dataset with a unified field structure and traceable correlation.

[0150] Furthermore, throughout the execution of the personalized treatment pathway, the treatment feedback data collected at different time points is categorized sequentially according to the timeline of the treatment nodes and undergoes standardized format conversion to ensure that all treatment feedback data has a unified field structure and data format. The content includes diagnostic results, examination parameters, equipment status, response time, and operation records. Simultaneously, to ensure data traceability, different treatment feedback data are linked through node identifiers, timestamps, and task numbers to obtain phased treatment feedback data, providing a structured and complete data foundation for pathway optimization and efficacy evaluation.

[0151] S5. Analyze the treatment effectiveness scores at each stage based on the phased treatment feedback data, and dynamically optimize the personalized treatment pathway.

[0152] After the personalized treatment pathway is completed, the effectiveness of the examinations, treatment results, and response indicators corresponding to each treatment node is evaluated based on the phased treatment feedback data.

[0153] The effectiveness evaluation process is completed by the diagnosis and treatment scoring engine, which evaluates the accuracy of diagnosis, treatment response rate, complication handling time, and degree of improvement of key indicators based on the diagnosis and treatment scoring rules.

[0154] For the original score X j,i The following normalization method is adopted:

[0155]

[0156] X j,i x represents the original score / original index value of node j in the i-th item; i,max This represents the maximum value in the history of the original rating metric; x i,min This refers to the lowest value in the history of the original scoring indicator.

[0157] An effectiveness score is calculated for each treatment node. The effectiveness score function is as follows:

[0158]

[0159] Among them, E j X represents the effectiveness score of treatment node j, with a value ranging from 0 to 1, where a value closer to 1 indicates a better effect; n represents the number of effectiveness evaluation indicators involved in the scoring (such as diagnostic accuracy, treatment response, improvement of key indicators, and node time), which is determined by the actual number of normalized indicators used in the calculation; j,i This represents the normalized score of node j on the i-th performance evaluation indicator, with a value range of 0 to 1; each indicator is obtained by normalizing the node feedback data (such as accuracy, responsiveness, and improvement magnitude).

[0160] For example: X j,1 For diagnostic accuracy, X j,2 For treatment response.

[0161] The effectiveness scores of each treatment node are aggregated in sequence to form a structured treatment effectiveness score table, which is then analyzed across the entire treatment path. When the effectiveness score of any treatment node in the score table falls below the diagnostic confidence threshold, a path optimization process is immediately triggered. This automatically feeds the effectiveness score and corresponding stage treatment feedback data into the path inference engine, recalculates the matching degree between structured symptom labels and treatment nodes, and, combined with the patient's status, resets the priority and execution order of treatment nodes to generate an optimized personalized treatment path.

[0162] It should be noted that the diagnostic confidence threshold is determined by clustering and classifying the effectiveness scores of different types of patients at various treatment nodes in historical cases, identifying stable score ranges corresponding to personalized treatment pathways with excellent effectiveness scores. The diagnostic confidence threshold is set based on the fluctuation range of effectiveness scores at each treatment node.

[0163] Specifically, the effectiveness scores of each treatment node in a large number of high-quality treatment pathways are normalized, and a clustering algorithm is used to divide the effectiveness score levels. Then, the confidence interval is determined based on the mean and standard deviation of the high-quality score clusters. The lower limit of the confidence interval is used as the diagnostic confidence threshold to determine whether the current path has reached a credible treatment level.

[0164] It should be noted that a high-quality treatment pathway refers to a complete treatment process that demonstrates accurate diagnosis, effective treatment, timely management of complications, and high overall efficacy in past cases.

[0165] If the effectiveness scores of all diagnostic and treatment nodes are higher than the diagnostic confidence threshold, the current pathway is confirmed to have good adaptability and treatment effect in this stage. The existing pathway plan is retained and the scoring results are archived for subsequent pathway reuse and knowledge accumulation.

[0166] This embodiment also provides an emergency internal medicine clinical diagnosis and treatment system, including: a data acquisition module, a path generation module, an execution control module, a path judgment module, and a path optimization module;

[0167] The data acquisition module is used to collect initial multimodal diagnostic data from the emergency department and preprocess it to generate a set of structured symptom labels.

[0168] The path generation module combines a set of structured symptom tags and basic patient information, and through the path reasoning engine, identifies the correlation and risk priority between structured symptom tags and treatment nodes, obtains the optimal combination of treatment process nodes, defines the execution order, and generates a personalized treatment path.

[0169] The execution control module is used to perform examinations and treatments according to the personalized treatment pathway and to collect treatment feedback data at each node in the personalized treatment pathway.

[0170] The path determination module is used to determine the completion status of the current node task based on the diagnosis and treatment feedback data, dynamically execute the next node task in the personalized diagnosis and treatment path, and collect the diagnosis and treatment feedback data of the next node until the personalized diagnosis and treatment path is completed and the stage diagnosis and treatment feedback data is generated.

[0171] The path optimization module is used to analyze the treatment effect scores at each stage based on the phased treatment feedback data, and dynamically optimize the personalized treatment path.

[0172] This embodiment also provides a computer device applicable to emergency internal medicine clinical diagnosis and treatment methods, including: a memory and a processor; the memory is used to store computer-executable instructions, and the processor is used to execute the computer-executable instructions to implement the emergency internal medicine clinical diagnosis and treatment methods proposed in the above embodiments.

[0173] The computer device can be a terminal, comprising a processor, memory, communication interface, display screen, and input devices connected via a system bus. The processor provides computing and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system and computer programs. The internal memory provides an environment for the operation of the operating system and computer programs stored in the non-volatile storage media. The communication interface is used for wired or wireless communication with external terminals; wireless communication can be achieved through Wi-Fi, carrier networks, NFC (Near Field Communication), or other technologies. The display screen can be an LCD screen or an e-ink screen. The input devices can be a touch layer covering the display screen, buttons, a trackball, or a touchpad on the computer device's casing, or an external keyboard, touchpad, or mouse.

[0174] This embodiment also provides a storage medium storing a computer program that, when executed by a processor, implements the emergency internal medicine clinical diagnostic and treatment methods proposed in the above embodiments. The storage medium can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as Static Random Access Memory (SRAM), Electrically Erasable Programmable Read-Only Memory (EEPROM), Erasable Programmable Read Only Memory (EPROM), Programmable Red-Only Memory (PROM), Read-Only Memory (ROM), magnetic storage, flash memory, magnetic disk, or optical disk.

[0175] In summary, this invention achieves precise identification of the correlation and risk priority between symptoms and treatment nodes by combining a structured symptom tag set and basic patient information, thereby constructing personalized treatment pathways and improving the matching accuracy of pathway development. By judging the completion status of the current node task based on treatment feedback data, the next node in the personalized pathway is dynamically executed, enabling real-time adjustment and continuous advancement of the pathway. These two aspects work together to construct a patient-state-driven, feedback-closed-loop controlled treatment process, enhancing the flexibility and adaptability of the pathway.

[0176] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.

Claims

1. An emergency internal medicine clinical diagnostic and treatment method, characterized in that: include, Collect initial multimodal diagnostic data from the emergency department and preprocess it to generate a set of structured symptom labels; By combining a set of structured symptom tags and basic patient information, and using a path reasoning engine, the correlation and risk priority between structured symptom tags and treatment nodes are identified, the optimal combination of treatment process nodes is obtained and the execution order is defined, and a personalized treatment path is generated. Examinations and treatments are conducted according to personalized treatment pathways, and treatment feedback data at each node of the personalized treatment pathways are collected. Based on the diagnosis and treatment feedback data, determine the completion status of the current node task, dynamically execute the next node task in the personalized diagnosis and treatment path, and collect the diagnosis and treatment feedback data of the next node until the personalized diagnosis and treatment path is completed, and generate phased diagnosis and treatment feedback data. Based on the analysis of the phased treatment feedback data, the treatment effect scores at each stage are calculated, and the personalized treatment pathway is dynamically optimized.

2. The emergency internal medicine clinical diagnostic and treatment method as described in claim 1, characterized in that: The initial multimodal diagnostic data for the emergency department includes the patient's description of their condition obtained by the voice acquisition device, medical image information collected from relevant parts of the patient, and the patient's basic information. The preprocessing includes format standardization, noise removal, and semantic annotation. By combining natural language recognition with image feature extraction, keyword extraction and semantic analysis are performed on the preprocessed initial multimodal diagnostic data of the emergency department, and a set of structured symptom labels is generated based on the image-text fusion mechanism.

3. The emergency internal medicine clinical diagnostic and treatment method as described in claim 2, characterized in that: The process of combining structured symptom tag sets and patient basic information, and using a path reasoning engine to identify the correlation and risk priority between structured symptom tags and treatment nodes, involves the following steps: Extract symptom names, descriptions of affected areas, durations, and manifestations from the structured symptom label set, and compare them with the patient's age, gender, medical history, allergy records, and underlying disease information to obtain a symptom-patient adaptive matching feature set; Based on the symptom-patient adaptive matching feature set, the path reasoning engine identifies the semantic correspondence between the structured symptom label set and each diagnosis and treatment node by matching the symptom-node mapping relationship, and calculates the matching degree between each diagnosis and treatment node and the structured symptom label set. Define the set of suitable nodes for the current patient across all treatment pathways based on the degree of matching, and sort the treatment nodes by combining the patient's urgency, resource availability, and risk level; The top N treatment nodes are selected as the optimal treatment nodes for the current patient, and the execution order of the treatment nodes is set according to the task priority to generate a personalized treatment path.

4. The emergency internal medicine clinical diagnostic and treatment method as described in claim 3, characterized in that: The steps for conducting examinations and treatments based on personalized treatment pathways, and collecting treatment feedback data at each node of the personalized treatment pathway, are as follows: The diagnostic and treatment devices corresponding to the personalized treatment path are activated sequentially according to the sorting results, and the device operation parameters are pushed according to the examination type and treatment method of each optimal diagnostic and treatment node. In a scenario where the first task node is image inspection, raw image data is obtained through image scanning; Edge recognition, region segmentation, and tissue density analysis are performed on the raw image data to identify lesion areas and tissue structure abnormalities, and the data is transformed into structured diagnostic and treatment feedback data through symptom feature label matching rules. If the current diagnosis and treatment node is a laboratory testing node, the execution flow control logic is invoked to start the sampling operation, and the rapid testing equipment is invoked to complete the reading and format conversion of blood routine, biochemical indicators and infection indicators; Synchronously record the diagnosis and treatment feedback data of the diagnosis and treatment nodes, and package and upload it to the path judgment process.

5. The emergency internal medicine clinical diagnostic and treatment method as described in claim 4, characterized in that: The steps for determining the completion status of the current node task and dynamically executing the next node task in the personalized treatment path are as follows: Identify the completion status of diagnosis and treatment tasks based on the criteria for determining the completion of diagnosis and treatment tasks; If the current treatment node task is determined to be completed, the path determination process will activate the next treatment node to perform the treatment operation according to the execution order of the treatment nodes in the personalized treatment path. If the current treatment node task is determined to be incomplete, the path dynamic adjustment logic is triggered, and the treatment feedback data of the completed treatment nodes and the newly generated treatment feedback data of the current treatment node are obtained. The treatment nodes are then reordered to generate an updated personalized treatment path, which is then executed by the current treatment node until all treatment node tasks are completed.

6. The emergency internal medicine clinical diagnostic and treatment method as described in claim 5, characterized in that: The generation of phased diagnosis and treatment feedback data refers to the process of categorizing diagnosis and treatment feedback data in a structured manner according to the time sequence of the nodes during the execution of personalized diagnosis and treatment node tasks, and performing a unified format standardization conversion on the content of the diagnosis and treatment feedback data to form phased diagnosis and treatment feedback data with a unified field structure and traceable correlation.

7. The emergency internal medicine clinical diagnostic and treatment method as described in claim 6, characterized in that: The analysis of treatment effectiveness scores at each stage and the dynamic optimization of personalized treatment pathways follow these steps: Based on the phased diagnosis and treatment feedback data, the diagnosis and treatment scoring engine is invoked to score the diagnosis and treatment effect of each diagnosis and treatment node that has been executed in the updated personalized diagnosis and treatment path; The treatment effectiveness scores at each node are archived in the order of the nodes to form a structured treatment effectiveness score table; If the treatment outcome score is lower than the diagnostic confidence threshold, the path optimization process is automatically triggered. The path optimization process inputs the treatment outcome score and treatment feedback data into the path inference engine to recalculate the matching degree between structured symptom labels and treatment nodes, and optimizes the personalized treatment path. If the treatment outcome score is higher than the diagnostic confidence threshold, the current personalized treatment pathway will be retained and the treatment outcome score sheet will be archived.

8. An emergency internal medicine clinical diagnostic and treatment system, based on the emergency internal medicine clinical diagnostic and treatment method according to any one of claims 1 to 7, characterized in that: include, The data acquisition module is used to collect initial multimodal diagnostic data from the emergency department and preprocess it to generate a set of structured symptom labels. The path generation module combines a set of structured symptom tags and basic patient information, and through the path reasoning engine, identifies the correlation and risk priority between structured symptom tags and treatment nodes, obtains the optimal combination of treatment process nodes, defines the execution order, and generates a personalized treatment path. The execution control module is used to perform examinations and treatments according to the personalized treatment pathway and to collect treatment feedback data at each node in the personalized treatment pathway. The path determination module is used to determine the completion status of the current node task based on the diagnosis and treatment feedback data, dynamically execute the next node task in the personalized diagnosis and treatment path, and collect the diagnosis and treatment feedback data of the next node until the personalized diagnosis and treatment path is completed and the stage diagnosis and treatment feedback data is generated. The path optimization module is used to analyze the treatment effect scores at each stage based on the phased treatment feedback data, and dynamically optimize the personalized treatment path.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that: When the processor executes the computer program, it implements the steps of the emergency internal medicine clinical diagnostic and treatment methods according to any one of claims 1 to 7.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by the processor, it implements the steps of the emergency internal medicine clinical diagnostic and treatment methods according to any one of claims 1 to 7.

Images