Follow-up visit consultation system for chronic pain patients

Abstract

The invention provides a follow-up visit consultation system for chronic pain patients, and relates to the technical field of intelligent medical treatment, and the system comprises a data collection unit which is used for obtaining subjective pain description data, real-time physiological monitoring data and historical diagnosis and treatment data of a patient; the prediction analysis unit is used for performing pain deterioration risk prediction based on the multi-source data and an artificial intelligence model, and outputting a prediction result containing a risk level and individualized pathological features; and the dynamic intervention unit is used for generating dynamic treatment suggestions according to the prediction result, pushing the dynamic treatment suggestions to the doctor and the patient, and receiving execution feedback data of the patient so as to perform closed-loop optimization on the treatment strategy. According to the technical scheme, the subjective pain description, the real-time physiological monitoring data and the historical diagnosis and treatment records of the patient are obtained at the same time, the integrity of the data is improved through the comprehensive collection mode of multi-source data, the accuracy and timeliness of information are ensured, and the problems of subjective deviation and inaccuracy caused by the fact that follow-up visit depends on memory of the patient are solved.

Description

Technical Field

[0001] The present application relates to the field of smart medical technology, and specifically to a follow-up consultation system for patients with chronic pain. Background Art

[0002] Modern medicine has classified pain as the fifth most important vital sign, following respiration, pulse, blood pressure, and body temperature. Chronic pain, a complex health issue, currently affects a significant number of people worldwide. It is defined as pain that persists beyond the normal tissue healing period (generally defined as three months). Post-treatment follow-up is an essential component of chronic pain treatment. This process allows physicians to more accurately assess a patient's recovery and adjust treatment plans as needed to ensure optimal efficacy and quality of life.

[0003] In practice, follow-up is mainly implemented through outpatient visits. Doctors find it difficult to obtain patients' daily pain data in a timely manner, and relevant information is obtained by relying on patients' recollections during follow-up visits. Such descriptions are highly subjective and prone to memory bias over time, which hinders doctors' accurate understanding of the actual situation after treatment, thereby affecting the evaluation of treatment effects and the adjustment of subsequent treatment plans. Therefore, in response to this deficiency in existing follow-up implementations, how to propose a technical solution to improve it has become a technical problem that needs to be solved urgently. Summary of the Invention

[0004] In order to at least overcome the problems existing in the related art to a certain extent, the embodiment of the present application provides a follow-up consultation system for chronic pain patients, which, through specific system configuration, is conducive to achieving better evaluation of treatment effects and adjustment of subsequent treatment plans.

[0005] In some embodiments of the present application, a follow-up consultation system for chronic pain patients is provided, which includes: Data acquisition unit, used to obtain patients' subjective pain description data, real-time physiological monitoring data, and historical diagnosis and treatment data; A prediction analysis unit is used to predict the risk of pain exacerbation based on the multi-source data obtained by the data acquisition unit and an artificial intelligence model, and output a prediction result including the risk level and individualized pathological characteristics; The dynamic intervention unit is used to generate dynamic treatment recommendations based on the prediction results and push them to both doctors and patients, as well as receive patients' execution feedback data to close the loop and optimize the treatment strategy.

[0006] In one possible implementation, the data collection unit includes: an artificial intelligence question-and-answer engine deployed on the patient's mobile terminal, which is used to structuredly collect pain location, intensity, and duration data; a wearable physiological monitoring device, which is used to obtain the patient's physiological monitoring data in real time, and the physiological monitoring data includes blood pressure, pulse, and blood oxygen; and a HIS system interface, which is used to batch import historical diagnostic records and medication plans.

[0007] In one possible implementation, the wearable physiological monitoring device is further used to obtain monitoring data for quantifying physical activity capacity based on a motion sensor, and to obtain monitoring data for evaluating autonomic nervous system stress state based on a skin electrical response sensor.

[0008] In one possible implementation, the process of predicting the risk of pain exacerbation through an artificial intelligence model includes: matching association rules between pain locations and high-risk diseases based on a knowledge graph, and analyzing dynamic change trends of multi-source data based on a time series machine learning model.

[0009] In one possible implementation, the time series machine learning model adopts an integrated architecture, which includes: an LSTM neural network for processing physiological indicator time series data; a random forest classifier for fusing static features to output a risk score, wherein the static features include age, medical history, and medication records.

[0010] In one possible implementation, the process of generating dynamic treatment recommendations based on prediction results includes: matching a step-by-step intervention knowledge base based on risk scores, and optimizing recommendation priorities in combination with the patient's historical medication effectiveness data.

[0011] In one possible implementation, in the stepped intervention knowledge base, health education and non-drug intervention plans are pushed to low-risk levels, and high-risk levels trigger manual review by doctors and generate drug adjustment recommendations.

[0012] In one possible implementation, the follow-up consultation system also includes an emergency response unit, which is used to automatically send a high-risk alert to the doctor when the predicted risk value exceeds a threshold; and simultaneously generate referral recommendations and emergency plans and push them to the patient's mobile terminal.

[0013] In one possible implementation, the dynamic intervention unit is deployed with a feedback learning mechanism for incrementally training the artificial intelligence model of the predictive analysis unit based on collected data on the patient's compliance with treatment recommendations.

[0014] In one possible implementation, the follow-up consultation system further includes a management analysis unit that visually displays patient data trend changes and interpretable attribution analysis of prediction results in the form of a decision support dashboard.

[0015] In the follow-up consultation system for chronic pain patients provided in the embodiment of the present application, the patient's subjective pain description, real-time physiological monitoring data, and historical diagnosis and treatment records are obtained simultaneously through an integrated data acquisition unit. This comprehensive collection method of multi-source data not only improves the integrity of the data, but also ensures the accuracy and timeliness of the information, and overcomes the subjective bias and inaccuracy caused by the traditional outpatient follow-up relying on the patient's memory: the artificial intelligence model of the predictive analysis unit is used to analyze the collected multi-source data. The system can provide each patient with a personalized pain exacerbation risk assessment, and then formulate a more targeted treatment plan based on these prediction results; and the dynamic intervention unit in the system automatically generates and pushes personalized treatment recommendations to both doctors and patients based on the results of the predictive analysis, and can receive the patient's execution feedback data to form a complete closed-loop management system. This method not only enhances the flexibility and adaptability of the treatment process, but also continuously optimizes the treatment strategy to achieve better treatment efficacy.

[0016] Other advantages, objectives, and features of the present application will be described in part in the following description and, in part, will be apparent to those skilled in the art based on an examination of the following or may be taught from the practice of the invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] The accompanying drawings are used to provide a further understanding of the technical solution of this application or the prior art and constitute a part of the specification. Among them, the drawings that express the embodiments of this application are used together with the embodiments of this application to explain the technical solution of this application, but do not constitute a limitation of the technical solution of this application.

[0018] Figure 1 A schematic block diagram of a follow-up consultation system for chronic pain patients provided in one embodiment of the present application; Figure 2 This is a schematic diagram illustrating the process of implementing structured collection of pain data based on an artificial intelligence question-answering engine in one embodiment of the present application; Figure 3 This is a block diagram of a follow-up consultation system for chronic pain patients in one embodiment of the present application. DETAILED DESCRIPTION

[0019] To make the purpose, technical solutions, and advantages of this application more clear, the technical solutions of this application will be described in detail below. Obviously, the embodiments described are only some of the embodiments of this application, rather than all of them. Based on the embodiments in this application, all other implementation methods obtained by ordinary technicians in this field without making any creative work are within the scope of protection of this application.

[0020] As mentioned in the background section, pain has been classified by modern medicine as the fifth most important vital sign, following respiration, pulse, blood pressure, and body temperature. Chronic pain, a complex health issue, currently affects a significant number of people worldwide. It is defined as a pain state that persists beyond the normal tissue healing period (generally defined as three months in medicine). Post-treatment follow-up is an essential component of chronic pain treatment. This process allows doctors to more accurately assess patients' recovery and adjust treatment plans as needed to ensure optimal efficacy and quality of life. In practice, follow-up is primarily conducted through outpatient visits, making it difficult for doctors to obtain timely access to patients' daily pain data. Relevant information relies on patients' recollections during follow-up visits. These descriptions are highly subjective and prone to memory errors over time. This hinders doctors' ability to accurately understand the actual post-treatment situation, which in turn affects their ability to assess treatment effectiveness and adjust subsequent treatment plans.

[0021] Based on this, the embodiment of the present application provides a follow-up consultation system for chronic pain patients, which, through specific system configuration, is conducive to achieving better evaluation of treatment effects and adjustment of subsequent treatment plans.

[0022] like Figure 1 FIG. 1 shows an embodiment of a follow-up consultation system for chronic pain patients proposed in this application. The follow-up consultation system includes: The data acquisition unit 100 is used to obtain the patient's subjective pain description data, real-time physiological monitoring data and historical diagnosis and treatment data. The data acquisition unit in this application is mainly used to integrate multi-source heterogeneous data to provide data basis for subsequent related processing.

[0023] Specifically, in some embodiments, the data collection unit includes: an artificial intelligence question-answering engine deployed on the patient's mobile terminal, which is used to structuredly collect pain location, intensity, and duration data. It is easy to understand that the patient's mobile terminal here can be a smartphone held by the patient, on which the artificial intelligence question-answering engine is deployed to obtain subjective pain description data. For example, Figure 2 As shown, based on the specific program configuration, the front end of the artificial intelligence question answering engine is implemented in the form of a pain diary, similar to Figure 2 The implementation process shown is based on the interaction with the patient to obtain structured data. For example, the structured data obtained may be {location: back, intensity: 7, duration: 3h, trigger factor: bending over}, etc.

[0024] This method of collecting pain-related data in a structured manner through an artificial intelligence question-answering engine can significantly improve the accuracy and consistency of the data, avoiding the ambiguity and errors that may be caused by traditional free-text input. In addition, the specific setting system based on AI technology can dynamically adjust questions according to the patient's answers to provide more precise options or further inquire about details to improve the quality of the data. At the same time, such standardized recording of patients' pain conditions will form a huge database, which is of great significance for the study of chronic pain. For example, big data-related technical means can be used based on a large number of actual cases to conduct further in-depth research; on the other hand, this method makes the recording process more intuitive and friendly, increasing user participation and the possibility of continued use.

[0025] In these embodiments, the data acquisition unit also includes a wearable physiological monitoring device, which is used to obtain the patient's physiological monitoring data in real time, including blood pressure, pulse, blood oxygen, etc. It is easy to understand that the wearable physiological monitoring device here is to achieve the acquisition and synchronous transmission of physiological monitoring data to facilitate subsequent related correlation analysis; for example, the wearable physiological monitoring device can be implemented based on a smart bracelet and a patch sensor, such as acquiring blood pressure, pulse, and blood oxygen indicators at a sampling frequency of 5 minutes per time.

[0026] In some embodiments, the wearable physiological monitoring device is further configured to obtain monitoring data for quantifying physical activity based on a motion sensor, and to obtain monitoring data for assessing autonomic nervous system stress based on a galvanic skin response sensor. Specifically, the motion sensor may be an integrated module of a three-axis accelerometer and a gyroscope, which is worn around the patient's waist at the L3-L4 vertebrae or wrist to obtain monitoring data for quantifying physical activity. The galvanic skin response sensor is used to capture the patient's autonomic nervous system stress value, filter it, and calculate the standard deviation as a fluctuation threshold. This is then combined with relevant physiological indicator data for subsequent predictive analysis.

[0027] For example, the built-in accelerometer in the bracelet can be used to detect the wearer's movements and combined with relevant algorithms to analyze the patient's body movements (for example, at night, the body will be more still in deep sleep, and there will be more movements in light sleep or when awake). Different sleep stages and durations can be distinguished, etc., which can be used as characteristic data information to evaluate the impact of pain on patients.

[0028] In these embodiments, the data collection unit further includes a HIS system interface for batch importing historical diagnostic records and medication regimens. Specifically, for example, to ensure system compatibility, the HIS system interface can be implemented based on the standard HL7 protocol to smoothly obtain the required data from the patient's electronic medical record.

[0029] like Figure 1As shown, the follow-up consultation system includes a prediction analysis unit 200, which is used to predict the risk of pain exacerbation based on the multi-source data obtained by the data acquisition unit 100 and the artificial intelligence model, and output a prediction result including the risk level and individualized pathological characteristics.

[0030] Specifically, in some embodiments, the above-mentioned process of predicting the risk of pain exacerbation through an artificial intelligence model includes: matching association rules between pain locations and high-risk diseases based on a knowledge graph, and analyzing dynamic change trends of multi-source data based on a time series machine learning model. In other words, this prediction and analysis unit adopts a dual-logic prediction architecture.

[0031] Specifically, for example, in the process of implementing the knowledge graph engine, the rules for constructing the pain-disease association rule base can be shown in Table 1 below: Table 1 Regarding the analysis of dynamic changing trends of multi-source data based on time series machine learning models, the time series machine learning model here can adopt an integrated architecture, which includes: LSTM neural network for processing physiological indicator time series data; random forest classifier, which integrates static features to output risk scores. Here, static features include age, medical history, medication records, etc.

[0032] For example, the input layer of the above-mentioned artificial intelligence model is used to input 72-hour sliding window data, including physiological indicators and subjective scoring data; in the processing layer: the LSTM network has 128 hidden units to learn the temporal dependency of physiological indicators, and the random forest classifier has 100 decision trees; the output layer outputs a risk level of 0-10 points.

[0033] Furthermore, in model implementation, the time window can be dynamically adjusted to adapt to different clinical needs. For example, when a longer time range is needed to analyze the development trend of the disease, the time window can be extended to capture more long-term trend information. In the LSTM network, multi-level structures can also be configured, such as stacked LSTM) to enhance the learning ability of complex time series data. For the random forest classifier, the maximum depth, minimum number of split samples, etc. can also be optimized to achieve the best performance.

[0034] This application adopts the above-mentioned predictive analysis method, based on the acquired multi-source data, combined with artificial intelligence models, starting from the two dimensions of knowledge graph reasoning and time series machine learning, to build a model with a dual logic prediction architecture, so as to improve the accuracy of prediction and clinical interpretability. In terms of knowledge graph reasoning, by establishing an association rule base between pain sites and high-risk diseases, and setting trigger conditions based on real-time physiological parameters, it can achieve rapid identification and matching of potential high-risk events, and enhance the system's pathological feature recognition capabilities; in terms of time series machine learning models, an integrated architecture is adopted, using LSTM neural networks to process physiological indicator time series data within a sliding window, effectively capturing time dependencies through hidden units, and combining random forest classifiers to integrate patients' static features (such as age, medical history, and medication records) for comprehensive evaluation, and ultimately outputting a risk level score, thereby achieving real-time, dynamic, and individualized pain exacerbation risk prediction, providing a scientific basis and technical support for clinical intervention.

[0035] like Figure 1 As shown, the follow-up consultation system also includes a dynamic intervention unit 300, which is used to generate dynamic treatment suggestions based on the prediction results and push them to both the doctor and the patient, and receive the patient's execution feedback data to close the loop and optimize the treatment strategy.

[0036] Specifically, the process of generating dynamic treatment recommendations based on prediction results involves matching risk scores with a stepped intervention knowledge base and prioritizing recommendations based on the patient's historical medication effectiveness data. As a specific implementation, the stepped intervention knowledge base pushes health education and non-drug intervention plans to low-risk tiers, while triggering manual review by doctors and generating medication adjustment recommendations for high-risk tiers.

[0037] For example, in some implementation scenarios, the stepwise intervention strategy can be as shown in Table 2 below: Table 2 Regarding receiving patient feedback data for closed-loop optimization of treatment strategies, a feedback learning mechanism can be deployed in the dynamic intervention unit through specific program settings. This feedback learning mechanism is used to incrementally train the AI ​​model of the predictive analysis unit based on the collected data on patient compliance with treatment recommendations. For example, based on specific program configuration, a relevant interactive interface can be provided on the patient's mobile device to receive patient feedback, such as "dizziness after taking pregabalin," to enable incremental model updates, such as adjusting the weights of drug side effect features in the random forest.

[0038] This application adopts the above-mentioned method to generate and push treatment recommendations based on the dynamic prediction and analysis results, and forms a closed-loop optimization mechanism by receiving patient execution feedback data, which is conducive to improving the intelligence level and clinical application value of the system. Among them, the dynamic intervention unit automatically matches the corresponding intervention strategy in the step-by-step intervention knowledge base based on the risk score output by the prediction and analysis unit and the patient's historical medication effectiveness data to ensure the scientific nature and individual adaptability of the intervention measures; and by deploying an interactive feedback interface on the patient side, timely collecting patient compliance with treatment recommendations and adverse reaction information, such as drug side effect feedback, and using these feedback data for incremental training of the artificial intelligence model in the prediction and analysis unit, it is possible to continuously optimize the risk prediction accuracy and the rationality of treatment recommendations, and achieve continuous optimization and improvement of the system.

[0039] In other embodiments, Figure 3 As shown, the follow-up consultation system for chronic pain patients proposed in this application also includes: The data acquisition unit 100 is used to obtain the patient's subjective pain description data, real-time physiological monitoring data, and historical diagnosis and treatment data; Similar data collection units include: an artificial intelligence question-and-answer engine deployed on the patient's mobile terminal, which is used to collect structured data on pain location, intensity and duration; wearable physiological monitoring equipment, which is used to obtain the patient's physiological monitoring data in real time, including blood pressure, pulse, blood oxygen, etc.; the data collection unit also includes a HIS system interface for batch importing historical diagnostic records and medication plans.

[0040] The prediction and analysis unit 200 is used to predict the risk of pain exacerbation based on the multi-source data obtained by the data acquisition unit 100 and an artificial intelligence model, and output a prediction result including the risk level and individualized pathological characteristics; In this embodiment, the above-mentioned process of predicting the risk of pain exacerbation through an artificial intelligence model includes: matching association rules between pain locations and high-risk diseases based on a knowledge graph, and analyzing dynamic change trends of multi-source data based on a time series machine learning model; regarding the analysis of dynamic change trends of multi-source data based on a time series machine learning model, the time series machine learning model here can adopt an integrated architecture, which includes: an LSTM neural network for processing physiological indicator time series data; a random forest classifier, which integrates static features to output a risk score, where static features include age, medical history, medication records, etc.

[0041] The dynamic intervention unit 300 is used to generate dynamic treatment recommendations based on the prediction results and push them to both the doctor and the patient, as well as receive the patient's execution feedback data to optimize the treatment strategy in a closed loop. Specifically, in the process of generating dynamic treatment recommendations based on the prediction results, it includes matching the step-by-step intervention knowledge base according to the risk score, and optimizing the recommendation priority in combination with the patient's historical medication effectiveness data. As a specific implementation method, in the step-by-step intervention knowledge base, health education and non-drug intervention plans are pushed to low-risk levels, and high-risk levels trigger manual review by doctors and generate drug adjustment recommendations.

[0042] Furthermore, in this embodiment, Figure 3 As shown, the follow-up consultation system for chronic pain patients also includes an emergency response unit 400, which is used to automatically send a high-risk alarm to the doctor when the predicted risk value exceeds a preset threshold; and simultaneously generate referral suggestions and first aid plans and push them to the patient's mobile terminal.

[0043] Specifically, for example, when the predicted risk value exceeds 8 points and meets relevant preset conditions, an emergency response is triggered. The preset conditions here may be systolic blood pressure continuously exceeding 180 mmHg or blood oxygen saturation less than 90% for more than 10 minutes. The corresponding emergency response trigger action may push a red alert containing the patient's location information to the doctor and send the patient emergency plan information, such as "Please stop the activity immediately and call 120! The nearest emergency center: XX Hospital (1.2 km away)." This approach of setting up an emergency response unit can enhance the follow-up consultation system's real-time response and handling capabilities in high-risk pain events, providing a critical guarantee for patient safety.

[0044] Furthermore, in this embodiment, Figure 3 As shown, the follow-up consultation system also includes a management analysis unit 500, which uses a decision support dashboard to visually display changes in patient data trends and interpretable attribution analysis of prediction results. Specifically, in actual implementation, the decision dashboard here may include a multi-dimensional data panel area to display pain score trend charts, physiological indicator heat maps, etc.; it also includes an interpretable analysis area for displaying interpretable attribution analysis information, such as the use of SHAP value visualization to display feature contribution. In addition, based on specific scenario requirements, historical intervention comparison display analysis can be performed to display the effectiveness statistics of different plans for patients with the same risk level.

[0045] Using this approach, interpretable attribution analysis of patient data trends and prediction results is visualized through a decision-support dashboard. Healthcare professionals can leverage this functionality of the follow-up consultation system to improve data analysis efficiency and, combined with their own medical knowledge, further enhance their decision-making capabilities. For example, this allows them to quickly identify trends in key health indicators, reducing interpretation time. Interpretable analysis provides a more robust data basis for medical decision-making, helping to understand the logic behind predictive models. This allows them to uncover implicit medical principles within the data, facilitating the development of more personalized and appropriate treatment plans.

[0046] In addition, it should be noted that the technical solution of this application can adopt a layered distributed architecture from the hardware architecture, such as consisting of a patient-side wearable layer, an edge computing layer, a central server layer, and a doctor terminal layer. Each layer is connected through a secure communication channel; Specifically, short-range wireless transmission between the patient's wearable layer and the edge computing layer uses protocols such as Bluetooth 5.2 to ensure real-time synchronization of physiological data. From the edge computing layer to the central server layer, encrypted uplink transmission is achieved through the HTTPS protocol overlaid with MQTT message queues. A high-bandwidth, low-latency connection between the central server layer and the doctor's terminal layer relies on a dedicated medical network to support high-load services such as video transmission.

[0047] The hardware components of the patient-side wearable layer include: a motion sensing unit, such as an integrated 9-axis inertial measurement unit (including a three-axis accelerometer, gyroscope, and magnetometer), which continuously captures body motion data at a 100Hz sampling rate; a neurostimulation unit, equipped with a dry electrode galvanic skin response sensor and a photoplethysmography module to simultaneously monitor heart rate variability; and a mobile terminal device, such as a smartphone or tablet, running an AI question-and-answer engine to structure the collection of patients' subjective pain descriptions. The edge computing layer primarily refers to the deployed edge intelligent gateway, which performs real-time filtering and feature extraction on raw physiological data.

[0048] The hardware architecture of the central server layer includes an AI analysis cluster, deploying integrated learning models such as LSTM neural networks and random forest classifiers to handle in-depth analysis and processing of multi-source fused data; a medical data storage system for data storage and disaster recovery. The physician terminal layer hardware includes a decision support workstation equipped with a large-screen display for display output to the physician end; and an emergency response terminal, carried by the managing physician for responding to crises.

[0049] Based on the above embodiments, the follow-up consultation system for chronic pain patients proposed in this application has the following technical effects: By integrating a mobile AI question-and-answer engine and wearable physiological monitoring devices, real-time and objective collection of patients' daily pain conditions and physiological indicator data is achieved, greatly improving the accuracy and timeliness of information and overcoming the limitations of traditional outpatient follow-up that relies on subjective recollection. Based on multi-source data, an AI model is used to predict the risk of pain exacerbation. Combined with knowledge graphs and time-series machine learning algorithms, it can accurately identify individualized pathological characteristics and dynamic change trends, providing a scientific basis for relevant intervention measures. Dynamic treatment recommendations are automatically generated based on the prediction results and pushed to both doctors and patients. At the same time, the patient's execution feedback data is received for closed-loop optimization of treatment strategies. This mechanism can promote timely adjustments to treatment plans, enhance patient participation and compliance, and help improve treatment outcomes.

[0050] An emergency response unit is set up to automatically trigger an alarm when high-risk is detected and push referral recommendations and emergency plans to relevant parties, ensuring that patients can quickly obtain appropriate medical assistance in emergencies, thereby improving the system's security and emergency response capabilities: a management analysis unit is set up to provide patient data trend change information and explainable attribution analysis information of prediction results in the form of a decision support dashboard, helping doctors to more intuitively understand the development of the disease and the reasons behind it, thereby facilitating more reasonable clinical decision-making.

[0051] The above description is merely a preferred embodiment of the present application, but the scope of protection of the present application is not limited thereto. Any changes or substitutions that can be easily conceived by anyone familiar with the technology within the technical scope disclosed in this application should be included in the scope of protection of the present application. Therefore, the scope of protection of the present application should be based on the scope of protection of the claims.

[0052] It can be understood that the same or similar parts of the above embodiments can be referenced to each other, and the contents not described in detail in some embodiments can refer to the same or similar contents in other embodiments.

[0053] It should be noted that, in the description of this application, the terms "first", "second", etc. are used for descriptive purposes only and should not be understood as indicating or implying relative importance. In addition, in the description of this application, unless otherwise specified, the meaning of "plurality" refers to at least two.

[0054] Throughout this specification, reference to terms such as "one embodiment," "some embodiments," "examples," "specific examples," or "some examples" means that a specific feature, structure, material, or characteristic described in conjunction with that embodiment or example is included in at least one embodiment or example of the present application. In this specification, schematic representations of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in any one or more embodiments or examples.

[0055] Although the embodiments of the present application have been shown and described above, it can be understood that the above embodiments are exemplary and cannot be understood as limitations on the present application. Ordinary technicians in this field can change, modify, replace and modify the above embodiments within the scope of the present application.

Claims

1. A follow-up consultation system for patients with chronic pain, characterized by: include: Data acquisition unit, used to obtain patients' subjective pain description data, real-time physiological monitoring data, and historical diagnosis and treatment data; A prediction analysis unit is used to predict the risk of pain exacerbation based on the multi-source data obtained by the data acquisition unit and an artificial intelligence model, and output a prediction result including the risk level and individualized pathological characteristics; The dynamic intervention unit is used to generate dynamic treatment recommendations based on the prediction results and push them to both doctors and patients, as well as receive patients' execution feedback data to close the loop and optimize the treatment strategy.

2. The follow-up consultation system for chronic pain patients according to claim 1, wherein: The data acquisition unit includes: The artificial intelligence question-and-answer engine deployed on the patient's mobile terminal is used to collect structured data on pain location, intensity and duration; the wearable physiological monitoring device is used to obtain the patient's physiological monitoring data in real time, including blood pressure, pulse and blood oxygen; the HIS system interface is used to batch import historical diagnostic records and medication plans.

3. The follow-up consultation system for chronic pain patients according to claim 2, wherein: The wearable physiological monitoring device is also used to obtain monitoring data for quantifying physical activity capacity based on a motion sensor, and to obtain monitoring data for evaluating the autonomic nervous system stress state based on a skin electrical response sensor.

4. The follow-up consultation system for chronic pain patients according to claim 1, wherein: The process of predicting the risk of pain exacerbation through an artificial intelligence model includes: matching association rules between pain locations and high-risk diseases based on a knowledge graph, and analyzing dynamic change trends of multi-source data based on a time series machine learning model.

5. The follow-up consultation system for chronic pain patients according to claim 4, wherein: The time series machine learning model adopts an integrated architecture, which includes: an LSTM neural network for processing physiological indicator time series data; a random forest classifier for fusing static features to output a risk score, wherein the static features include age, medical history, and medication records.

6. The follow-up consultation system for chronic pain patients according to claim 5, wherein: The process of generating dynamic treatment recommendations based on the prediction results includes: matching the step-by-step intervention knowledge base according to the risk score, and optimizing the recommendation priority in combination with the patient's historical medication effectiveness data.

7. The follow-up consultation system for chronic pain patients according to claim 6, wherein: In the stepped intervention knowledge base, health education and non-drug intervention plans are pushed to low-risk levels, and manual review by doctors and generation of drug adjustment suggestions are triggered for high-risk levels.

8. The follow-up consultation system for chronic pain patients according to claim 5, wherein: The follow-up consultation system also includes an emergency response unit for automatically sending a high-risk alert to the doctor when the predicted risk value exceeds a threshold; And simultaneously generate referral recommendations and emergency plans and push them to the patient's mobile terminal.

9. The follow-up consultation system for chronic pain patients according to claim 1, wherein: The dynamic intervention unit is deployed with a feedback learning mechanism for incrementally training the artificial intelligence model of the predictive analysis unit based on the collected patient compliance data on treatment recommendations.

10. The follow-up consultation system for chronic pain patients according to claim 1, wherein: The follow-up consultation system also includes a management analysis unit that visually displays patient data trend changes and interpretable attribution analysis of prediction results in the form of a decision support dashboard.

Images