General medical department doctor-patient information interaction method and system
By collecting and structuring multimodal information, a summary report for doctors is generated, which solves the problem of incomplete information in doctor-patient communication, improves diagnosis and treatment efficiency and patient compliance, and achieves continuous and dynamic optimization of information.
Patent Information
- Application Number
- CN202511311868.8
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-09-15
- Publication Date
- 2025-12-12
AI Technical Summary
In general practice, doctor-patient communication suffers from incomplete information presentation, logical confusion, and omission of key details, making it difficult for doctors to accurately obtain patients' health status. This affects the efficiency and accuracy of diagnosis and treatment decisions, especially in remote consultations or multiple follow-up visits, where information is scattered, reducing the efficiency of diagnosis and treatment and patient compliance.
By collecting multimodal information, semantic parsing, and structuring, a structured health information vector is generated. Combined with cross-modal mapping of image features and physiological indicators, the symptom fluctuation and behavior consistency index is calculated, and a summary report for doctors is generated and updated synchronously, achieving continuous and dynamic optimization of information.
It has achieved continuity and dynamic optimization of doctor-patient information interaction, improved diagnosis and treatment efficiency, information utilization rate and patient compliance, and enhanced the operability of doctor-patient communication.
Smart Images
Figure CN121122789A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of information interaction technology, and more specifically, to a method and system for doctor-patient information interaction in general practice. Background Technology
[0002] In the daily clinical practice of general medicine, information exchange between doctors and patients is a core component of the treatment process. Patients typically provide health information through verbal descriptions, text input, or uploaded images and voice messages. However, this information often suffers from incomplete descriptions, logical inconsistencies, omissions of key details, or inconsistent formatting. Faced with a large amount of unstructured information, doctors struggle to obtain a timely and accurate understanding of the patient's overall health status, thus affecting the efficiency and accuracy of treatment decisions. Furthermore, during remote consultations or multiple follow-up visits, the information provided by patients is often fragmented, preventing doctors from forming a continuous and systematic view of health management. This not only reduces treatment efficiency but also impacts patient understanding and adherence to treatment plans.
[0003] Therefore, there is an urgent need for a method to achieve intelligent doctor-patient information interaction in a networked environment. This method should be able to automatically identify, structure, and fuse patient-reported symptoms and auxiliary information in a multimodal manner, and generate real-time health information summaries and diagnostic feedback that doctors can use. Without these functions, patient health information will be difficult to understand and use efficiently and accurately, doctor-patient information interaction will be disjointed, diagnostic decision-making efficiency will be reduced, the development and implementation of health management plans will be limited, and patient experience and compliance will significantly decline. This technical problem has become a key challenge that urgently needs to be addressed in the information management and intelligent assisted diagnosis and treatment of general practitioners. Summary of the Invention
[0004] To overcome the aforementioned deficiencies in the prior art, this invention provides a method and system for doctor-patient information interaction in general practice. This method involves collecting multimodal information by time-encapsulation, generating structured health information vectors through semantic parsing, mapping these vectors across modalities with image features and physiological indicators, fusing historical and examination data to calculate symptom fluctuation and consistency indices and generate a comprehensive judgment coefficient, and outputting a doctor-side summary with synchronous updates. This addresses the problems mentioned in the background section.
[0005] To achieve the above objectives, the present invention provides the following technical solution: a method for doctor-patient information interaction in general practice, comprising: Step S1: The patient terminal collects multimodal health information, including text, voice, and images, and encapsulates the timestamp and patient identifier into a raw data packet; Step S2: Perform semantic parsing and entity extraction on the text and voice content in the original data packet, map symptoms, location, and duration to structured health information vectors, extract image features and associate them with structured health information vectors; Step S3: The structured health information vector is matched and fused with the patient's historical health data and previous examination results in multiple dimensions to form a comprehensive health information set. The patient's symptom fluctuation index and behavior consistency index are calculated. The symptom fluctuation index is calculated by the change in the frequency and time interval of symptom occurrence, and the behavior consistency index is calculated by the difference and score change between the patient's behavior and health advice. The results are then input into a predefined model to generate a comprehensive health judgment coefficient. Step S4: Based on the comprehensive health information set and comprehensive health judgment coefficient, generate a doctor's summary report and diagnosis and treatment reference tips, including potential abnormalities, missing information prompts, and health indicators that need attention; Step S5: Synchronize the doctor's operation feedback and modification results back to the patient's end, and update the health information vector and summary report so that subsequent information interaction can directly use the latest structured data for processing and analysis.
[0006] Preferably, in step S2, a timestamp and source identifier are uniformly added to the original data packet, and symptoms, location and duration are formed according to semantic segmentation, and a symptom evolution chain is constructed based on sentence logic and time sequence; Physiological indicators are standardized and mapped across modally with the symptom evolution chain to establish a correspondence between semantic nodes and vital sign parameters, and structured health information vectors are generated through vectorized encoding.
[0007] Preferably, the image feature processing includes: performing multi-scale edge detection and texture pattern recognition on the patient-uploaded image, extracting image feature parameters that characterize the morphology, color distribution, and spatial location of the lesion, registering the image feature parameters with the corresponding symptom location identifier according to the timestamp, and embedding the registered image feature parameters into the structured health information vector in a vectorized form, so that the image features form a one-to-one multimodal mapping relationship with the symptom, location, and duration information obtained from text and speech parsing, which is used to enhance the accuracy of subsequent information fusion and health status analysis.
[0008] Preferably, in the process of matching and fusing structured health information vectors with patients' historical health data and previous examination results, the patient information is segmented according to the chronological order, and a corresponding first weight parameter is generated based on the continuity of symptom records and the completeness of physiological indicator collection within each segment. This first weight parameter is used to dynamically adjust the contribution of information from different sources during information fusion, thereby ensuring that the fusion result can simultaneously reflect immediate symptom changes and long-term health trends.
[0009] Preferably, the patient symptom fluctuation index The method of obtaining it is: ; in, The weight of the i-th symptom reflects the importance of the symptom in health information. Its value ranges from [0, 1] and is assigned by the symptom type and severity. represents the magnitude of change in the semantic vector of the i-th symptom, and represents the difference in the description of the patient's self-reported symptoms in different time periods, with a quantization range of [0, 1]. The time interval for the onset of symptom i is in days (or hours), and the dimension is time. represents the time interval decay coefficient, which controls the influence of the symptom interval on the index. It is set to (0, 1) to ensure that the weight of long-term symptoms decays but does not disappear. N is the number of symptoms in the time period and is a positive integer. The Patient Symptom Fluctuation Index serves to reflect the dynamic fluctuation characteristics of symptoms by weighting the semantic changes of symptoms and combining them with nonlinear decay processing of the time interval between occurrences, thus providing a quantitative indicator for subsequent comprehensive health assessment.
[0010] Preferably, the behavioral consistency index Obtained through the following predefined model: ; in, This represents the magnitude of change of the j-th physiological indicator relative to the standardization; the smaller the value, the more the behavior conforms to the recommendations. This represents the time interval from the occurrence of the j-th physiological indicator to the current time, in days (or hours), with the dimension of time. The time decay adjustment coefficient is represented, with a value range of (0, 1], which is used to reduce the impact of historical behavior on current consistency; M represents the number of physiological index measurements during the observation period, which is a positive integer.
[0011] Preferably, the comprehensive health assessment coefficient (CHC) is obtained in the following way: ; in, The second weighting parameter [0, 1] is the output comprehensive health judgment coefficient CHC. The higher the CHC, the greater the symptom fluctuation and the unstable physiological indicators, which require special attention.
[0012] Preferably, step S3, in the process of generating the comprehensive health judgment coefficient, further includes: The symptom fluctuation index and the physiological indicator stability index are processed by time series segmentation, and a second weighting parameter is calculated based on the trend consistency and fluctuation amplitude difference of the index within each segment. The second weighting parameter is introduced into the calculation of the comprehensive health judgment coefficient, so that the comprehensive health judgment coefficient can simultaneously reflect the coupling relationship between short-term symptom changes and long-term health status, providing a differentiated reference for the generation of subsequent doctor-side summary reports.
[0013] Preferably, step S4, in the process of generating the doctor's summary report and treatment reference tips, further includes: Based on the comprehensive health judgment coefficient, the abnormal probability and information completeness of different health indicators of patients are hierarchically ranked, and the ranking results are dynamically associated with potential abnormalities and information missing prompts to automatically generate a graded attention list.
[0014] Preferably, in step S4, the comprehensive health judgment coefficient is mapped in multiple dimensions with the patient's historical health trend and structured health information vector. A visual summary is dynamically generated according to the degree of abnormality of health indicators, data completeness and rate of change. Potential abnormal indicators, missing information items and key health indicators that need attention are displayed in priority groups in the visual summary.
[0015] To achieve the above objectives, the present invention provides the following technical solution: a general practice medical doctor-patient information interaction system, comprising: Information acquisition module: Collects multimodal health information from patients, including text descriptions, voice recordings, and images (such as photos and medical images); outputs raw data packets uniformly encapsulated according to timestamps and patient identifiers; The information structuring module extracts structured health information vectors from the raw data packets and associates them with image features to form a multimodal vector. This includes semantic parsing and entity extraction of text and speech, mapping symptom, location, and duration information into vectorized representations, and extracting image features and associating them with the structured health information vectors. Health Integration Module: It summarizes structured health information vectors, patient historical health data and previous examination results into a comprehensive health information set, calculates symptom fluctuation index and behavior consistency index, inputs into a predefined model, and outputs comprehensive health judgment coefficient; Report generation module: Based on the comprehensive health information set and comprehensive health judgment coefficient, it generates a summary report for doctors and diagnosis and treatment reference tips, including potential abnormalities, missing information prompts, and health indicators that need attention; Feedback Update Module: Synchronizes the doctor's operation feedback and modification results back to the patient's end, and updates the health information vector and summary report.
[0016] The technical effects and advantages of this invention are as follows: This invention collects multimodal health information, including text, voice, and images, from the patient's end and encapsulates it into raw data packets according to timestamps and patient identifiers. This achieves complete information collection and temporal consistency, effectively solving the problems of fragmented doctor-patient information and difficulty in integrating modalities in traditional methods. Simultaneously, by performing semantic parsing and entity extraction on the text and voice, symptoms, locations, and durations are mapped to structured health information vectors. These vectors are then correlated with image features to form a unified multimodal encoding representation. This enables the automatic conversion of unstructured information into structured health information, providing a reliable foundation for subsequent health analysis.
[0017] This invention performs multi-dimensional matching and fusion of health information vectors with patients' historical data and previous examination results, summarizing them into a comprehensive health information set. It then calculates symptom fluctuation and behavioral consistency indices, inputting these into a predefined model to generate a comprehensive health judgment coefficient. This enables intelligent judgment of health status and identification of key concerns, effectively solving the problem of medical staff struggling to quickly obtain crucial information. Based on the comprehensive health information set and judgment coefficients, it generates summary reports and treatment reference prompts for doctors, while simultaneously updating patient-side information vectors. This ensures continuous and dynamically optimized doctor-patient information interaction, significantly improving treatment efficiency, information utilization, and patient compliance, while also enhancing the operability of doctor-patient communication. Attached Figure Description
[0018] Figure 1 This is a flowchart of the doctor-patient information interaction method in general practice according to the present invention. Detailed Implementation
[0019] Exemplary embodiments of the present disclosure will now be described in more detail with reference to the accompanying drawings. While exemplary embodiments of the present disclosure are shown in the drawings, it should be understood that the present disclosure may be implemented in various forms and should not be limited to the embodiments set forth herein. Rather, these embodiments are provided so that this disclosure will be thorough and complete, and will fully convey the scope of the disclosure to those skilled in the art.
[0020] At the same time, it should be understood that, for ease of description, the dimensions of the various parts shown in the accompanying drawings are not drawn according to actual scale.
[0021] The following description of at least one exemplary embodiment is merely illustrative and is in no way intended to limit the scope of this application and its application or use.
[0022] Techniques, methods, and equipment known to those skilled in the art may not be discussed in detail, but where appropriate, such techniques, methods, and equipment should be considered part of the specification.
[0023] Example 1, see Figure 1A flowchart illustrating a method for doctor-patient information interaction in general practice. This invention provides a method for doctor-patient information interaction in general practice, comprising: Step S1: The patient terminal collects multimodal health information, including text, voice, and images, and encapsulates the timestamp and patient identifier into a raw data packet; Step S2: Perform semantic parsing and entity extraction on the text and voice content in the original data packet, map symptoms, location, and duration to structured health information vectors, extract image features and associate them with structured health information vectors; Step S3: The structured health information vector is matched and fused with the patient's historical health data and previous examination results in multiple dimensions to form a comprehensive health information set. The patient's symptom fluctuation index and behavior consistency index are calculated and input into a predefined model to generate a comprehensive health judgment coefficient. The explanation is as follows: the symptom fluctuation index is calculated by the change in the frequency and time interval of symptom occurrence, the behavior consistency index is calculated by the difference between the patient's behavior and health advice and the change in the score, and the comprehensive health judgment coefficient is used to determine whether special attention is needed. Step S4: Based on the comprehensive health information set and comprehensive health judgment coefficient, generate a doctor's summary report and diagnosis and treatment reference tips, including potential abnormalities, missing information prompts, and health indicators that need attention; Step S5: Synchronize the doctor's operation feedback and modification results back to the patient's end, and update the health information vector and summary report so that subsequent information interaction can directly use the latest structured data for processing and analysis.
[0024] Furthermore, in step S2, a timestamp and source identifier are uniformly added to the original data packet, and symptoms, location, and duration are formed according to semantic segmentation. A symptom evolution chain is constructed based on sentence logic and time sequence. Physiological indicators are standardized and mapped across modally with the symptom evolution chain to establish a correspondence between semantic nodes and vital sign parameters. Structured health information vectors are generated through vectorization encoding and used as input for subsequent health information fusion and index calculation.
[0025] Furthermore, the image feature processing includes: performing multi-scale edge detection and texture pattern recognition on the patient-uploaded images (medical images or photos), extracting image feature parameters that characterize the morphology, color distribution, and spatial location of lesions, and registering the image feature parameters with the corresponding symptom location identifiers according to timestamps. The registered image feature parameters are then embedded into the structured health information vector in a vectorized form, so that the image features form a one-to-one multimodal mapping relationship with the symptom, location, and duration information obtained from text and speech parsing, which is used to enhance the accuracy of subsequent information fusion and health status analysis.
[0026] Furthermore, in the process of matching and fusing structured health information vectors with patients' historical health data and previous examination results, patient information is segmented according to chronological order. Based on the continuity of symptom records and the completeness of physiological indicator collection within each segment, a corresponding first weight parameter is generated to dynamically adjust the contribution of information from different sources during information fusion, thereby ensuring that the fusion result can simultaneously reflect immediate symptom changes and long-term health trends.
[0027] Furthermore, the patient symptom fluctuation index The method of obtaining it is: ; in, The weight of the i-th symptom reflects the importance of the symptom in health information. Its value ranges from [0, 1] and is assigned by the symptom type and severity. represents the magnitude of change in the semantic vector of the i-th symptom, and represents the difference in the description of the patient's self-reported symptoms in different time periods, with a quantization range of [0, 1]. The time interval for the onset of symptom i is in days (or hours), and the dimension is time. represents the time interval decay coefficient, which controls the influence of the symptom interval on the index. It is set to (0, 1) to ensure that the weight of long-term symptoms decays but does not disappear. N is the number of symptoms in the time period and is a positive integer. The Patient Symptom Fluctuation Index serves to reflect the dynamic fluctuation characteristics of symptoms by weighting the semantic changes of symptoms and combining them with nonlinear decay processing of the time interval between occurrences, thus providing a quantitative indicator for subsequent comprehensive health assessment.
[0028] Furthermore, the behavioral consistency index Obtained through the following predefined model: ; in, This represents the magnitude of change of the j-th physiological indicator relative to the standardization; the smaller the value, the more the behavior conforms to the recommendations. This represents the time interval from the occurrence of the j-th physiological indicator to the current time, in days (or hours), with the dimension of time. represents the time decay adjustment coefficient, with a value range of (0, 1], used to reduce the impact of historical behavior on current consistency; M represents the number of physiological index measurements during the observation period, which is a positive integer. The role of the behavioral consistency index: It is a weighted average of behavioral differences, taking into account the decay effect of the time of behavior occurrence, and reflects the stability and regularity of patients' behavioral performance.
[0029] Furthermore, the comprehensive health assessment coefficient (CHC) is obtained as follows: ; in, The second weighting parameter [0, 1] is the output comprehensive health judgment coefficient CHC. The higher the CHC, the greater the symptom fluctuation and the unstable physiological indicators, which require special attention.
[0030] Furthermore, a second weighting parameter is dynamically set by combining the contribution of patient symptoms and physiological indicators at different time scales. The specific method is as follows: First, in the initial stage, the contribution weights of the symptom fluctuation index and the physiological indicator stability index are set to be equal to ensure that information from different dimensions is incorporated into the comprehensive judgment in a balanced manner. Subsequently, during continuous monitoring, the weighting of symptoms and physiological indicators was dynamically adjusted based on the ratio of the number of recent visits to the frequency of physiological indicator collection. When the frequency of visits to the patient was high and the symptom updates were more frequent, the second weighting parameter shifted towards the direction of symptom fluctuations to highlight the immediate role of clinical complaints in assessing health status. When the patient's symptom records were sparse but the collection of physiological indicators was stable and continuous, the second weighting parameter shifted towards the direction of physiological indicators to enhance the reliability of objective measurement data in assessing health status. By dynamically switching between data time density and acquisition completeness, weights can be flexibly set in different scenarios, so that the comprehensive health judgment coefficient has both immediate responsiveness and long-term stability.
[0031] Furthermore, step S3, in the process of generating the comprehensive health judgment coefficient, also includes: The symptom fluctuation index and the physiological indicator stability index are processed by time series segmentation, and a second weighting parameter is calculated based on the trend consistency and fluctuation amplitude difference of the index within each segment. The second weighting parameter is introduced into the calculation of the comprehensive health judgment coefficient, so that the comprehensive health judgment coefficient can simultaneously reflect the coupling relationship between short-term symptom changes and long-term health status, providing a differentiated reference for the generation of subsequent doctor-side summary reports.
[0032] Furthermore, step S4, in the process of generating the doctor's summary report and treatment reference prompts, also includes: Based on the comprehensive health judgment coefficient, the abnormal probability and information completeness of different health indicators of patients are hierarchically ranked, and the ranking results are dynamically associated with potential abnormalities and information gaps to automatically generate a graded attention list. This allows doctors to intuitively identify high-risk indicators and data gaps when reading summary reports, thereby prioritizing the most critical health information in diagnosis and treatment decisions and achieving a synergistic effect of information aggregation and key point reminders.
[0033] Furthermore, in step S4, the comprehensive health judgment coefficient is mapped in multiple dimensions with the patient's historical health trends and structured health information vectors. A visual summary is dynamically generated according to the degree of abnormality of health indicators, data integrity, and rate of change. The visual summary is then grouped by priority to display potential abnormal indicators, missing information items, and key health indicators that need attention. This allows doctors to obtain information on immediate abnormalities, long-term trends, and data integrity in a single view, thereby improving the efficiency and accuracy of diagnosis and treatment decisions and achieving a synergistic effect of multi-dimensional information aggregation and key point reminders.
[0034] Example 2: This embodiment of the invention provides a general practice medical doctor-patient information interaction system, including: Information acquisition module: Collects multimodal health information from patients, including text descriptions, voice recordings, and images (such as photos and medical images); outputs raw data packets uniformly encapsulated according to timestamps and patient identifiers; The information structuring module extracts structured health information vectors from the raw data packets and associates them with image features to form a multimodal vector. This includes semantic parsing and entity extraction of text and speech, mapping symptom, location, and duration information into vectorized representations, and extracting image features and associating them with the structured health information vectors. Health Integration Module: It summarizes structured health information vectors, patient historical health data and previous examination results into a comprehensive health information set, calculates symptom fluctuation index and behavior consistency index, inputs into a predefined model, and outputs comprehensive health judgment coefficient; Report generation module: Based on the comprehensive health information set and comprehensive health judgment coefficient, it generates a summary report for doctors and diagnosis and treatment reference tips, including potential abnormalities, missing information prompts, and health indicators that need attention; Feedback Update Module: Synchronizes the doctor's operation feedback and modification results back to the patient's end, and updates the health information vector and summary report.
[0035] In conclusion, the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.
Claims
1. A method for doctor-patient information interaction in general practice, characterized in that, include: Step S1: The patient terminal collects multimodal health information, including text, voice, and images, and encapsulates it into a raw data packet according to timestamps and patient identifiers; Step S2: Perform semantic parsing and entity extraction on the text and voice content in the original data packet, map symptoms, location, and duration to structured health information vectors, extract image features and associate them with structured health information vectors; Step S3: The structured health information vector is matched and fused with the patient's historical health data and previous examination results in multiple dimensions to form a comprehensive health information set. The patient's symptom fluctuation index and behavior consistency index are calculated. The symptom fluctuation index is calculated by the change in the frequency and time interval of symptom occurrence, and the behavior consistency index is calculated by the difference and score change between the patient's behavior and health advice. The results are then input into a predefined model to generate a comprehensive health judgment coefficient. Step S4: Based on the comprehensive health information set and comprehensive health judgment coefficient, generate a doctor's summary report and diagnosis and treatment reference tips, including potential abnormalities, missing information prompts, and health indicators that need attention; Step S5: Synchronize the doctor's operation feedback and modification results back to the patient's end, and update the health information vector and summary report so that subsequent information interaction can directly use the latest structured data for processing and analysis.
2. The method for doctor-patient information interaction in general practice according to claim 1, characterized in that, In step S2, a timestamp and source identifier are uniformly added to the original data packet, and symptoms, location and duration are formed according to semantic segmentation. A symptom evolution chain is constructed based on sentence logic and time sequence. Physiological indicators are standardized and mapped across modally with the symptom evolution chain to establish a correspondence between semantic nodes and vital sign parameters, and structured health information vectors are generated through vectorized encoding.
3. The method for doctor-patient information interaction in general practice according to claim 2, characterized in that, The image feature processing includes: performing multi-scale edge detection and texture pattern recognition on the images uploaded by the patient, extracting image feature parameters that characterize the morphology, color distribution and spatial location of the lesions, registering the image feature parameters with the corresponding symptom location identifiers according to the timestamp, and embedding the registered image feature parameters into the structured health information vector in a vectorized form, so that the image features form a one-to-one multimodal mapping relationship with the symptom, location and duration information obtained from text and speech parsing.
4. The method for doctor-patient information interaction in general practice according to claim 3, characterized in that, In the process of matching and fusing structured health information vectors with patients' historical health data and previous examination results, patient information is segmented according to chronological order. Based on the continuity of symptom records and the completeness of physiological indicator collection within each segment, a corresponding first weight parameter is generated to dynamically adjust the contribution of information from different sources during information fusion, thereby ensuring that the fusion result can simultaneously reflect immediate symptom changes and long-term health trends.
5. The method for doctor-patient information interaction in general practice according to claim 1, characterized in that, The patient's symptom fluctuation index The method of obtaining it is: ; in, The weight of the i-th symptom reflects the importance of the symptom in health information. Its value ranges from [0, 1] and is assigned by the symptom type and severity. represents the magnitude of change in the semantic vector of the i-th symptom, and represents the difference in the description of the patient's self-reported symptoms in different time periods, with a quantization range of [0, 1]. The time interval for the onset of symptom i is measured in units of time. represents the time interval decay coefficient, which controls the influence of symptom interval on the index, and is taken as (0, 1); N is the number of symptoms within the time period, and is a positive integer.
6. The method for doctor-patient information interaction in general practice according to claim 5, characterized in that, The behavioral consistency index Obtained through the following predefined model: ; in, This represents the magnitude of change of the j-th physiological indicator relative to the standardization; the smaller the value, the more the behavior conforms to the recommendations. This represents the time interval from the occurrence of the j-th physiological indicator to the current time, with the dimension of time. The time decay adjustment coefficient is represented, with a value range of (0, 1], which is used to reduce the impact of historical behavior on current consistency; M represents the number of physiological index measurements during the observation period, which is a positive integer.
7. A method for doctor-patient information interaction in general practice according to claim 6, characterized in that, The method for obtaining the comprehensive health assessment coefficient (CHC) is as follows: ; in, The second weighting parameter [0, 1] is the output comprehensive health judgment coefficient CHC. The higher the CHC, the greater the symptom fluctuation and the unstable physiological indicators, which require special attention.
8. A method for doctor-patient information interaction in general practice according to claim 7, characterized in that, Step S3, in the process of generating the comprehensive health assessment coefficient, further includes: The symptom fluctuation index and the physiological indicator stability index are processed by time series segmentation, and a second weighting parameter is calculated based on the trend consistency and fluctuation amplitude difference of the index within each segment. The second weighting parameter is introduced into the calculation of the comprehensive health judgment coefficient, so that the comprehensive health judgment coefficient can simultaneously reflect the coupling relationship between short-term symptom changes and long-term health status, providing a differentiated reference for the generation of subsequent doctor-side summary reports.
9. A method for doctor-patient information interaction in general practice according to claim 8, characterized in that, In step S4, the comprehensive health judgment coefficient is mapped in multiple dimensions with the patient's historical health trend and structured health information vector. A visual summary is dynamically generated according to the degree of abnormality of health indicators, data completeness and rate of change. Potential abnormal indicators, missing information items and key health indicators that need attention are displayed in priority groups in the visual summary.
10. A general practice medical doctor-patient information interaction system, used to implement the method described in claim 1, characterized in that, include: Information collection module: Collects multimodal health information from patients, including text descriptions, voice recordings, and images; Output raw data packets uniformly encapsulated by timestamp and patient identifier; The information structuring module extracts structured health information vectors from the raw data packets and associates them with image features to form a multimodal vector. This includes semantic parsing and entity extraction of text and speech, mapping symptom, location, and duration information into vectorized representations, and extracting image features and associating them with the structured health information vectors. Health Integration Module: It summarizes structured health information vectors, patient historical health data and previous examination results into a comprehensive health information set, calculates symptom fluctuation index and behavior consistency index, inputs into a predefined model, and outputs comprehensive health judgment coefficient; Report generation module: Based on the comprehensive health information set and comprehensive health judgment coefficient, it generates a summary report for doctors and diagnosis and treatment reference tips, including potential abnormalities, missing information prompts, and health indicators that need attention; Feedback Update Module: Synchronizes the doctor's operation feedback and modification results back to the patient's end, and updates the health information vector and summary report.