Team construction method and system for individual case management in whole course of disease

Abstract

The invention relates to the technical field of medical informatization, and discloses a team construction method and system for whole course case management, and the method comprises the steps: integrating medical data and non-medical Internet of Things data through a cross-domain data fusion engine to generate a patient behavior map, carrying out the reverse deduction of wearable equipment data as an inducement based on a reverse reasoning algorithm, and triggering hierarchical intervention, detachable combination and dynamic pricing of points are realized by using a block chain smart contract; according to the system, the data island limitation of traditional medical monitoring is avoided, the risk inducement can be reversely locked before the physiological indexes are abnormal through the implicit association between the smart home power utilization frequency and the medicine use record, accurate early warning is achieved, and the safety of the system is improved through multi-modal data collaborative analysis and deep coupling of sandbox rehearsal decision and the smart contract. And the technical jump from passive treatment to active health management is realized.

Description

Technical Field

[0001] The present invention relates to a method and system for constructing a team for whole-course case management, belonging to the technical field of medical informatization. Background Art

[0002] In scenarios such as chronic disease management and postoperative rehabilitation, the medical team needs to integrate multi-source data such as patients' medical records, physiological monitoring, and living habits, and dynamically coordinate rehabilitation resources to provide continuous services. The current industry generally adopts the following technical paths: 1. At the level of multi-source data integration: Traditional systems mainly rely on structured data of the Hospital Information System (HIS). Although they can integrate basic information such as medication records, they have insufficient ability to fuse non-medical Internet of Things data such as smart home electricity consumption frequency and wearable device positioning, resulting in limited patient behavior analysis in a single dimension; for example, the implicit association between abnormal night lighting and medication compliance is often ignored, making risk warnings lag behind the deterioration of actual physiological indicators.

[0003] 2. At the level of resource scheduling decision-making: The allocation of community rehabilitation resources is mostly based on a fixed schedule or manual experience, lacking dynamic prediction of service demand fluctuations. Although existing digital twin technologies can simulate resource allocation, their models are not linked with patients' real-time behavior data, resulting in a significant deviation between the virtual deduction results and actual needs, and unable to effectively alleviate the contradiction of coexistence of resource idleness and shortage.

[0004] 3. At the level of patient participation incentive: Existing point systems mostly adopt a one-way redemption mode, with fixed uses of points and unable to reflect the dynamic changes in resource supply and demand; for example, during periods of shortage of rehabilitation therapists, services are still redeemed at a fixed ratio, which not only exacerbates resource occupation, but also makes it difficult to guide patients to actively optimize their health behaviors through the point leverage.

[0005] To improve the real-time nature of data integration, some solutions alleviate the synchronization delay by increasing the data transmission frequency of wearable devices, but this leads to a sharp increase in the risk of patient privacy leakage, and the computing power load of edge devices exceeds the bearing limit of the medical Internet of Things gateway; in terms of optimizing resource scheduling, some research introduces reinforcement learning algorithms to predict service demand, but their model training depends on the static features of historical data and is difficult to adapt to the dynamic evolution of patients' behavior patterns. Instead, it exacerbates the management cost due to frequent schedule adjustments. Therefore, how to achieve deep coupling of cross-domain behavior characteristics while ensuring data security and construct a two-way dynamic optimization mechanism for resource allocation and patient participation has become the technical problem to be solved by the present invention. Summary of the Invention

[0006] The present invention provides a method and system for constructing a team for whole-course case management, and its main purpose is to solve the problems of decision-making lag caused by multi-modal data islands, service imbalance caused by rigid resource scheduling, and the difficulty of the one-way point mechanism to stimulate patients' active health management behaviors.

[0007] To achieve the above-mentioned purpose, the present invention provides a team building method for full-course case management, comprising the following steps: Step 1: Integrate the patient medication information recorded in the medical information system and the patient life habit data from non-medical IoT devices through a cross-domain patient behavior association engine. Non-medical IoT devices include smart home devices and wearable health monitoring devices. Smart home devices at least include a power monitoring module and a lighting monitoring module. Wearable health monitoring devices at least include a heart rate monitoring module and a location monitoring module. Use a natural language processing module to perform semantic analysis and feature extraction on medication information and life habit data, and construct a patient behavior map that includes both medical behavior features and non-medical behavior features of the patient. The patient behavior map is used to characterize the patient's behavior patterns in the medical and life fields and the potential association between the two. Step 2: Based on the reverse reasoning algorithm, the real-time physiological data from the wearable health monitoring device is analyzed. When abnormal fluctuations in the physiological data are monitored, such as a sudden increase in heart rate or an activity range exceeding a preset threshold, the possible causes of the abnormal fluctuations are reversely deduced. The reverse deduction at least associates and analyzes the non-medical behavior characteristics in the patient's behavior map, such as associating the timestamp of the sudden increase in heart rate with the positioning data of the smart bracelet to determine whether the patient is outside the scope of daily activities and may not carry medication, or associating the abnormal frequency of use of night lighting equipment with the use of insomnia drugs in the medication record, inferring the relationship between insomnia and the patient's medication compliance, and generating a graded intervention warning signal containing potential behavioral inducements; Step 3: Build a dynamic resource sandbox simulation system. Before making a decision, the system simulates the impact of different resource allocation plans on the satisfaction of patients' service needs. Resources include at least community rehabilitation workers. The simulation process is based on historical patient service data and predicted future service needs. It allocates different quantities and types of service resources in a virtual environment, predicts service gaps and service redundancies within a preset time frame (e.g., the next 72 hours), and generates a prediction report on the effect of resource allocation. Step 4: Use the elastic priority algorithm to dynamically adjust the priority of the patient service task queue according to the patient's risk level and the resource virtual occupancy rate predicted by the resource dynamic sandbox simulation system. For high-risk patients, prioritize the service resources that have been simulated and verified in the virtual sandbox and have a lower resource occupancy rate to ensure that high-risk patients can get the required services in a timely manner. Step 5: Deploy a points system based on blockchain technology. The points system uses smart contracts to realize the divisible combination of points and the dynamic pricing of community service resources, allowing patients to split the points they earn into different proportions to redeem different services. For example, part of the follow-up points can be redeemed for registration priority, and the remaining part can be used for drug discounts. At the same time, the smart contract automatically adjusts the points redemption ratio according to the supply and demand of community service resources. When the resources of rehabilitation therapists are tight, the points required for redemption will be increased, and vice versa.

[0008] Preferably, in step 2, the reverse reasoning algorithm further includes: when it is monitored that the frequency of electricity consumption in the smart home increases abnormally during non-normal sleep periods, the patient's medication records are correlated and analyzed, and if there is a record of the use of sedatives and hypnotic drugs, a warning signal is generated that insomnia may cause a deviation in the medication time; when the patient's geographic location data is monitored by a wearable health monitoring device to exceed a preset distance threshold of their daily activity range and the duration exceeds a preset time, a risk mark for not carrying regular medications is automatically triggered.

[0009] Preferably, in step 3, the resource dynamic sandbox simulation system also considers the disease characteristics and service preferences of patients in different communities when simulating the effect of resource allocation, and optimizes the resource allocation plan in a targeted manner to improve resource utilization efficiency and service satisfaction.

[0010] Preferably, in step 4, the elastic priority algorithm also considers the patient's recent service history, appointment preferences, and geographic location distribution of community service resources when adjusting the priority of the patient service task queue, thereby achieving more refined resource matching and task scheduling.

[0011] Preferably, in step 5, the smart contract nesting mechanism allows patients to combine different types of points to redeem higher value or more personalized service packages, such as combining follow-up points and health behavior points to redeem a home visit service by a family doctor.

[0012] Preferably, in step 5, the dynamic pricing of community service resources is also adjusted in combination with factors such as the professional level of the service provider, the length of service, and the patient's evaluation feedback, so as to achieve fairer and more efficient resource allocation.

[0013] Preferably, the method further includes step 6, wherein personalized service recommendations, risk warning information, and redeemable points and service options are displayed to patients through a user interface, and patients are allowed to make service reservations and redeem points according to their own needs.

[0014] Preferably, the user interface also integrates augmented reality technology to provide patients with visual rehabilitation guidance, medication reminders, and community service navigation information.

[0015] Preferably, in step 1, the cross-domain patient behavior association engine also integrates patient health-related discussion data from social media platforms, uses sentiment analysis and topic modeling techniques to mine patients' potential health needs and psychological states, and incorporates the analysis results into the patient behavior map to more comprehensively understand the patients' behavior patterns.

[0016] A team building system for full-course case management, comprising: A cross-domain patient behavior association module, configured to integrate patients' medication information from a medical information system and patients' living habit data from non-medical Internet of Things devices, and use a natural language processing module to construct a patient behavior map including patients' medical behavior characteristics and non-medical behavior characteristics; A reverse reasoning analysis module, configured to analyze real-time physiological data from wearable health monitoring devices, reverse-derive the incentives that may cause abnormal fluctuations, and generate a hierarchical intervention warning signal including potential behavior incentives; A resource dynamic sandbox deduction module, configured to simulate the impact of different resource allocation schemes on the satisfaction of patients' service needs before decision-making, and generate a prediction report on the resource allocation effect; A flexible priority scheduling module, configured to dynamically adjust the priority of the patient service task queue according to the risk level of the patient and the virtual resource occupancy rate predicted by the resource dynamic sandbox deduction module; A blockchain points management module, configured to realize the split combination of points and the dynamic pricing of community service resources through smart contracts; A data interaction and display module, configured to display personalized service recommendations, risk warning information, and redeemable points and service options to patients, and receive patients' service reservation and points redemption operation requests.

[0017] Compared with the problems in the background art, the beneficial effects of the present invention are: 1. Through the collaborative analysis of medication records, smart home, and wearable device data by the cross-domain data fusion engine, the system can capture behavior association patterns that are difficult to detect in traditional medical monitoring. For example, considering the hidden association between abnormal night lighting and medication compliance in practice, this multi-modal feature cross-analysis avoids the limitation of a single data dimension, enabling the risk warning mechanism to reverse-derive potential incentives from the patient's living scenario and trigger precise intervention before the occurrence of abnormal physiological indicators, effectively avoiding the warning lag problem caused by data islands in traditional solutions.

[0018] 2. The resource virtual sandbox identifies conflict nodes in community rehabilitation resource scheduling in advance by simulating the service gap distributions of different resource allocation scenarios. Combined with the dynamic marking of high-risk patients using the elastic priority algorithm, the system can intelligently reorganize the service queue under resource constraints. This closed-loop mechanism of prediction - simulation - adjustment enables the resource allocation plan to have self-optimization capabilities, avoiding the rigid defects of traditional scheduling and ensuring the feasibility of service strategies through virtual verification.

[0019] 3. Through the intelligent contract nesting of splitable points and dynamic pricing rules, the system converts patient behavior data into quantifiable and adjustable service redemption rights. The point combination mechanism not only reflects the cumulative value of individual healthy behaviors but also guides the flow of community resources to the most optimized nodes through the adjustment of the supply-demand linked redemption ratio. This design avoids the one-way redemption mode of traditional point systems, constructs a positive feedback loop for patient participation while ensuring privacy and security, and through the real-time parsing of multi-source heterogeneous data based on a lightweight semantic distillation model, the system can still maintain the accuracy of behavior intention recognition in a resource-constrained environment. Through the reverse reasoning mechanism triggered by timestamp deviation, the system can reconstruct the patient behavior time sequence and correct data synchronization errors, enabling the dynamic sandbox deduction to be based on a reliable data foundation. This synergistic effect of edge intelligence and semantic enhancement ensures the stable implementation of complex decision-making logic in clinical scenarios. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] Figure 1 is the data integration and processing flowchart of the present invention; Figure 2 is the timing diagram of the reverse reasoning and warning process of the present invention; Figure 3 is the flowchart for constructing the patient behavior map of the present invention.

[0021] The implementation, functional features, and advantages of the present invention will be further described with reference to the embodiments and the accompanying drawings. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0022] It should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention.

[0023] The embodiments of the present application provide a method for constructing a team for full-course case management, including the following steps: Step 1: Integrate the patient medication information recorded in the medical information system and the patient life habit data from non-medical IoT devices through a cross-domain patient behavior association engine. Non-medical IoT devices include smart home devices and wearable health monitoring devices. Smart home devices at least include a power monitoring module and a lighting monitoring module. Wearable health monitoring devices at least include a heart rate monitoring module and a location monitoring module. Use a natural language processing module to perform semantic analysis and feature extraction on medication information and life habit data, and construct a patient behavior map that includes both medical behavior features and non-medical behavior features of the patient. The patient behavior map is used to characterize the patient's behavior patterns in the medical and life fields and the potential association between the two. Step 2: Based on the reverse reasoning algorithm, the real-time physiological data from the wearable health monitoring device is analyzed. When abnormal fluctuations in the physiological data are monitored, such as a sudden increase in heart rate or an activity range exceeding a preset threshold, the possible causes of the abnormal fluctuations are reversely deduced. The reverse deduction at least associates and analyzes the non-medical behavior characteristics in the patient's behavior map, such as associating the timestamp of the sudden increase in heart rate with the positioning data of the smart bracelet to determine whether the patient is outside the scope of daily activities and may not carry medication, or associating the abnormal frequency of use of night lighting equipment with the use of insomnia drugs in the medication record, inferring the relationship between insomnia and the patient's medication compliance, and generating a graded intervention warning signal containing potential behavioral inducements; Step 3: Build a dynamic resource sandbox simulation system. Before making a decision, the system simulates the impact of different resource allocation plans on the satisfaction of patients' service needs. Resources include at least community rehabilitation workers. The simulation process is based on historical patient service data and predicted future service needs. It allocates different quantities and types of service resources in a virtual environment, predicts service gaps and service redundancies within a preset time frame (e.g., the next 72 hours), and generates a prediction report on the effect of resource allocation. Step 4: Use the elastic priority algorithm to dynamically adjust the priority of the patient service task queue according to the patient's risk level and the resource virtual occupancy rate predicted by the resource dynamic sandbox simulation system. For high-risk patients, prioritize the service resources that have been simulated and verified in the virtual sandbox and have a lower resource occupancy rate to ensure that high-risk patients can get the required services in a timely manner. Step 5: Deploy a points system based on blockchain technology. The points system uses smart contracts to realize the divisible combination of points and the dynamic pricing of community service resources, allowing patients to split the points they earn into different proportions to redeem different services. For example, part of the follow-up points can be redeemed for registration priority, and the remaining part can be used for drug discounts. At the same time, the smart contract automatically adjusts the points redemption ratio according to the supply and demand of community service resources. When the resources of rehabilitation therapists are tight, the points required for redemption will be increased, and vice versa.

[0024] Preferably, in step 2, the reverse reasoning algorithm further includes: when it is monitored that the frequency of electricity consumption in the smart home increases abnormally during non-normal sleep periods, the patient's medication records are correlated and analyzed, and if there is a record of the use of sedatives and hypnotic drugs, a warning signal is generated that insomnia may cause a deviation in the medication time; when the patient's geographic location data is monitored by a wearable health monitoring device to exceed a preset distance threshold of their daily activity range and the duration exceeds a preset time, a risk mark for not carrying regular medications is automatically triggered.

[0025] Preferably, in step 3, the resource dynamic sandbox simulation system also considers the disease characteristics and service preferences of patients in different communities when simulating the effect of resource allocation, and optimizes the resource allocation plan in a targeted manner to improve resource utilization efficiency and service satisfaction.

[0026] Preferably, in step 4, the elastic priority algorithm also considers the patient's recent service history, appointment preferences, and geographic location distribution of community service resources when adjusting the priority of the patient service task queue, thereby achieving more refined resource matching and task scheduling.

[0027] Preferably, in step 5, the smart contract nesting mechanism allows patients to combine different types of points to redeem higher value or more personalized service packages, such as combining follow-up points and health behavior points to redeem a home visit service by a family doctor.

[0028] Preferably, in step 5, the dynamic pricing of community service resources is also adjusted in combination with factors such as the professional level of the service provider, the length of service, and the patient's evaluation feedback, so as to achieve fairer and more efficient resource allocation.

[0029] Preferably, step 6 is also included, wherein personalized service recommendations, risk warning information, and redeemable points and service options are displayed to patients through the user interface, and patients are allowed to make service reservations and redeem points according to their own needs.

[0030] Preferably, the user interface also integrates augmented reality technology to provide patients with visual rehabilitation guidance, medication reminders, and community service navigation information.

[0031] Preferably, in step 1, the cross-domain patient behavior association engine also integrates patient health-related discussion data from social media platforms, uses sentiment analysis and topic modeling techniques to explore patients' potential health needs and psychological states, and integrates the analysis results into the patient behavior map to more comprehensively understand the patient's behavior patterns.

[0032] A team-building system for full-course case management, including: A cross - domain patient behavior association module, which is used to integrate the patient's medication information from the medical information system and the patient's lifestyle data from non - medical Internet of Things devices, and uses the natural language processing module to construct a patient behavior map including the patient's medical behavior characteristics and non - medical behavior characteristics; A reverse reasoning analysis module, which is used to analyze the real - time physiological data from wearable health monitoring devices, reverse - deduce the incentives that may cause abnormal fluctuations, and generate a hierarchical intervention warning signal including potential behavior incentives; A resource dynamic sandbox deduction module, which is used to simulate the impact of different resource allocation schemes on the satisfaction of patient service needs before decision - making, and generate a prediction report on the effect of resource allocation; An elastic priority scheduling module, which is used to dynamically adjust the priority of the patient service task queue according to the patient's risk level and the predicted virtual resource occupancy rate of the resource dynamic sandbox deduction module; A blockchain integral management module, which is used to realize the split combination of integral and the dynamic pricing of community service resources through smart contracts; A data interaction and display module, which is used to display personalized service recommendations, risk warning information, and redeemable integral and service options to patients, and receive the service reservation and integral redemption operation requests from patients.

[0033] Embodiment 1: In this embodiment, through a cross - domain patient behavior association engine, the patient's medication records from the medical information system (HIS) and the patient's lifestyle data from non - medical Internet of Things devices (such as smart home, wearable health monitoring devices) are integrated. The smart home devices include, for example, at least one power consumption monitoring module and a lighting monitoring module, and the wearable health monitoring devices include at least a heart rate monitoring module and a location monitoring module. The natural language processing (NLP) module is used to perform semantic analysis and feature extraction on the medication records and lifestyle data, so as to construct a comprehensive patient behavior map; this behavior map can represent the behavior characteristics of patients in the medical field and the life field, as well as the potential associations between these behavior characteristics. The patient's medical behavior characteristics include medication conditions, treatment processes, etc., and the life behavior characteristics include sleep quality, activity range, environmental factors, etc.

[0034] In the real-time monitoring of patients' physiological data, when abnormal fluctuations occur in heart rate or other physiological indicators, the reverse inference algorithm is activated to analyze whether the abnormality is caused by lifestyle or behavioral incentives. For example, if the heart rate suddenly increases, the system will automatically correlate the smart home data during that time period to analyze whether the patient is outside the normal activity range or has missed taking medicine on time. Through the correlation analysis with the patient's behavior map, the possible behavioral incentives are deduced in reverse, and then corresponding warning signals are generated. During this process, the system will combine the patient's non-medical behavior characteristics, such as the usage frequency of night lighting equipment and the use of insomnia drugs in the medication record, to reverse-deduce the possible medication deviation caused by insomnia symptoms and generate targeted intervention tips.

[0035] Based on the patient's behavior and physiological data, the system will use the resource dynamic sandbox deduction model to simulate the deployment plans of different community rehabilitation resources. Considering the service data of historical patients and the prediction of future service needs, by simulating the effects of different resource allocation plans, the service gaps and surplus resources are identified, and finally a prediction report of service resources is generated. This sandbox deduction system helps decision-makers optimize resource allocation in the case of limited actual resources by simulating service needs and resource deployment in a virtual environment, and predicts service gaps or redundancy phenomena. This process ensures that service resources can be allocated in the optimal way, avoiding the inaccuracy and inefficiency problems in traditional manual scheduling methods. In addition, the system dynamically adjusts the patient's service task queue through the elastic priority algorithm. According to the patient's risk level and resource virtual occupancy rate, the algorithm automatically adjusts the priority of high-risk patients. High-risk patients will be preferentially allocated the verified service resources with a lower resource occupancy rate in the virtual sandbox. In this way, the system can ensure that high-risk patients can obtain the required services in a timely manner, avoiding unfairness or delays in resource allocation.

[0036] During the resource scheduling process, the algorithm will also consider factors such as the patient's historical service records, appointment preferences, and the geographical distribution of community service resources to achieve refined resource matching and task scheduling. Through this flexible priority mechanism, the system can dynamically adjust according to actual needs and resource conditions, improving resource utilization efficiency and service timeliness. In the patient service points system, the system realizes the split combination of points and the dynamic pricing of community service resources through smart contracts. Patients participate in the accumulation of health behavior points through the data of smart home and wearable devices and use these points to exchange for medical services. The smart contract in the system dynamically adjusts the point exchange ratio according to the supply and demand of community service resources. When the resources of rehabilitation therapists are in short supply, the system will increase the points required for exchange, and vice versa. Through this mechanism, the system can guide patients to actively participate in health management while optimizing resource allocation. By displaying personalized service recommendations, risk warning information, and exchangeable points and service options through the user interface, patients can make service appointments and point exchange operations according to their own needs. The user interface also integrates augmented reality technology to provide patients with visual rehabilitation guidance, medication reminders, and community service navigation information, enhancing the user experience and service personalization.

[0037] Embodiment 2: This embodiment combines Figures 1 to 3 , and describes a method and system for constructing a team for full-course case management. As Figure 1 shown, data is first collected from four data sources, which are: medical information system records, smart home device data, wearable health monitoring device data, and social media platform data. Among them, the medical information system records provide patient medication information, both the smart home device data and the wearable health monitoring device data provide lifestyle data, while the social media platform data provides health-related discussion data. These data are integrated into a cross-domain patient behavior association engine, which receives the medication and lifestyle data from the natural language processing module and the social media data from sentiment analysis and topic modeling respectively. The natural language processing module performs semantic analysis and feature extraction on the data, while sentiment analysis and topic modeling analyze the potential needs and psychological states of the data. The processed data finally forms a patient behavior map, which contains the medical and non-medical behavior characteristics and associations of the patient and is finally output.

[0038] As Figure 2 shown, Figure 2It involves a wearable device, a behavior graph engine, a reverse reasoning module, an early warning decision-making system, a blockchain network, and a family member terminal. The process starts with the wearable device recording real-time heart rate data (120 bpm). Subsequently, the wearable device sends a request for behavior correlation analysis to the behavior graph engine. After receiving the request, the behavior graph engine first retrieves the medication record (sedative not taken), then retrieves the location data (outside the normal activity range), and then sends the returned correlation features to the reverse reasoning module. Based on the received correlation features, the reverse reasoning module generates a secondary early warning (medication not taken + abnormal activity), and records this early warning information in the early warning event (timestamp + risk value). This event record is then block-confirmed through the blockchain network. At the same time, the early warning decision-making system sends a command to the wearable device to trigger a vibration reminder to take medicine immediately, and sends a location notification to the family member terminal.

[0039] As Figure 3 shown, the process first starts, and then proceeds to the step of integrating cross-domain data sources. The data sources include medication information provided by the medical information system and lifestyle data provided by the non-medical Internet of Things. After the integrated data undergoes data preprocessing, natural language processing is performed, and semantic analysis is carried out. Next, the process enters the step of extracting behavior features. The extracted behavior features are used to construct a patient behavior graph, which includes medical behavior features, non-medical behavior features, and potential associations between behaviors. The entire graph is represented by the label "graph". Finally, the graph is stored and updated, and the process ends.

[0040] Embodiment 3: In the present invention, the main task of the cross-domain data fusion engine is to integrate the medication information from the medical information system and the patient's lifestyle data from the non-medical Internet of Things devices. This process performs semantic analysis and feature extraction on the medical data and lifestyle data through the natural language processing NLP module to construct a patient behavior graph. The behavior graph includes the patient's medical behavior features such as medication records and treatment processes, and non-medical behavior features such as sleep quality, activity range, and environmental factors, thereby reflecting the patient's behavior patterns and potential associations in the medical and lifestyle domains. When the sensor detects abnormal patient physiological data such as a sharp increase in heart rate, the reverse reasoning module will be triggered to analyze the possible causes of this abnormal fluctuation. The system will reverse-derive the possible behavioral causes by correlating the non-medical behavior features in the patient behavior graph, such as the frequency of using night lighting in the smart home and the patient's activity range. For example, the system may find that the sudden increase in heart rate is abnormally correlated with the frequency of using lighting devices in the smart home at night, and thus infer that the patient may have taken relevant medications due to insomnia. At this time, the system will automatically generate a warning signal and guide the medical staff to carry out appropriate interventions.

[0041] The resource dynamic sandbox deduction module of the present invention aims to optimize the allocation plan of community rehabilitation resources. During the resource allocation process, this module will simulate different resource allocation plans and predict the service demand and resource usage in the next period (such as within 72 hours). This simulation process is based on historical patient data and future predictions, taking into account the availability of resources and the fluctuations in patient needs, generating a service gap report and an excess resource report. This report provides a reference for decision-makers so that, in the case of limited resources, service resources can be reasonably allocated to avoid resource waste or service shortages. At the same time, the elastic priority algorithm dynamically adjusts the priority of the task queue according to the risk level of patients and the virtual resource occupancy rate. For high-risk patients, the system will preferentially allocate verified service resources with a lower resource occupancy rate. The implementation of the elastic priority algorithm takes into account factors such as the patient's historical service records, appointment preferences, and service duration, thus achieving more refined resource matching and task scheduling.

[0042] The integral system of the present invention realizes the dynamic pricing and splitting combination of points through smart contracts. The behavior data of each patient, such as medication records and health monitoring data, will be converted into points. These points can be used to exchange different types of service resources, such as registration priorities, drug discounts, community health services, etc. The smart contract mechanism enables the point exchange ratio to be adjusted in real time according to the supply and demand situation of service resources. When the resources of community rehabilitation therapists are in short supply, the system will automatically increase the points required for exchange, and vice versa. This system not only optimizes resource allocation but also promotes patients' active participation in health management through an integral incentive mechanism.

[0043] Example 4: First, integrate the medication records from the medical information system HIS and the lifestyle data from non-medical Internet of Things devices such as smart homes and wearable health monitoring devices through a cross-domain patient behavior association engine. Use the natural language processing NLP module to perform semantic analysis and feature extraction on medical data and non-medical data. During the feature extraction process, special attention is paid to the following variables: medication records, including drug names, medication times, dosages, etc.; non-medical behavior characteristics, including the electricity consumption frequency of smart homes, lighting patterns, exercise amounts, heart rates, positions recorded by wearable devices, etc. After fusing these data, a patient behavior map is generated. This map can comprehensively represent the behavior characteristics and potential associations of patients in medical and daily life, serving as the basis for subsequent behavior prediction and intervention.

[0044] When abnormal fluctuations are detected in the patient's real-time physiological data such as heart rate or range of motion, the reverse inference algorithm will be activated; this algorithm analyzes the potential causes behind the abnormal fluctuations, mainly through the following steps: Heart rate sudden increase analysis, if the heart rate shows a significant increase during the monitoring period, the system first associates the timestamp with smart home data such as lighting frequency to determine whether the patient is within an unusual range of activities. If the patient's location exceeds their regular activity area, it is speculated that the patient may not have taken their medication on time, and a risk warning signal is then generated; Sleep disorder speculation, if the usage frequency of the lighting equipment at night is abnormally high, the system conducts an associated analysis of the patient's medication records. If it is found that sedative drugs are used, it is speculated that insomnia may cause a deviation in the medication time, and a targeted intervention prompt is generated. The reverse inference algorithm is not limited to analyzing physiological data abnormalities, but also combines non-medical data in the patient's behavior profile for comprehensive judgment to ensure that the behavioral causes leading to physiological fluctuations can be accurately located and avoid reasoning based on a single data source alone.

[0045] To ensure the dynamic allocation of resources, the system simulates the effects of different resource allocation schemes before making a decision; this process is based on historical patient data and future service demand forecasts, and the specific steps are as follows: Simulating resource allocation, according to the predicted patient needs and historical data, the virtual sandbox system will simulate resource allocation schemes to identify service gaps and redundancies. Priority scheduling algorithm, using an elastic priority algorithm, combines the patient's risk level and resource occupancy rate to dynamically adjust the priority of the task queue; this algorithm takes into account the patient's historical service records, appointment preferences, and the geographical distribution of service resources to ensure that resources are allocated first in the most urgent situations. During this process, the virtual occupancy rate is an important parameter that describes the usage of each resource in the virtual sandbox. Based on the virtual occupancy rates of different resources, the algorithm can dynamically adjust the resource allocation strategy to meet different service demands.

[0046] Based on blockchain technology, the points system realizes the split combination and dynamic pricing of points through smart contracts; the specific operation is as follows: Points system, patients generate points through the data of wearable devices and smart home devices, and exchange the points for medical services according to the points. Patients can choose to exchange for different types of services, such as registration priority, drug discounts, etc.; The smart contract adjusts the exchange ratio. The smart contract automatically adjusts the points required for exchange according to the supply and demand situation of service resources. When resources are in short supply, the system increases the points required for exchange, and vice versa, reduces the point requirement. This mechanism ensures that community service resources can be flexibly allocated according to actual needs by guiding patients to optimize their health behaviors.

[0047] Example 5: At the data integration level, the cross-domain patient behavior correlation engine, as the core of information aggregation, actively obtains data from multiple channels. It not only receives standard medication records from the Hospital Information System (HIS), which contain key information such as drug names, usage, dosage, and time; but also accesses lifestyle data generated by smart home devices and wearable health monitoring devices. Smart home devices can provide indirect behavior information such as the patient's daily routine and electricity usage patterns, such as the duration of night lighting. Wearable devices, on the other hand, continuously monitor physiological data such as heart rate, activity level, sleep duration, and location information. In addition, in order to more comprehensively understand the patient's potential needs and psychological state, the engine also obtains health-related discussion data posted by the patient on public social media platforms. All of this raw data is first preprocessed, including cleaning, deduplication, and format unification, and then sent to the Natural Language Processing (NLP) module for in-depth analysis. The NLP module performs semantic analysis and entity extraction on text data such as medication orders and social media content, identifying technical features related to the patient's health, such as emotional tendencies, health topics of concern, and attitudes towards treatment plans. The processed structured and unstructured data is used to construct and continuously update a multi-modal patient behavior map. This map represents entities such as patients, their behaviors, physiological states, medications, environmental factors, and social activities, and their interrelationships in the form of nodes and edges.

[0048] The reverse inference analysis module works closely with the patient behavior map to analyze the underlying causes behind abnormal physiological data. When the wearable device captures a significant deviation of the patient's real-time physiological indicators from their individual baseline or the medical normal range, the reverse inference mechanism is triggered. For example, if it is monitored that the patient's heart rate continues to rise, the system will, according to the current time point, search for other related behavior data in the patient behavior map, such as whether the activity state at that time is resting or exercising, whether the geographical location is outside the usual activity area, whether there are specific medication records in the recent period, such as stimulants or sedatives, and even whether there are signs of anxiety or stress in the patient's emotional state reflected in the map. By comparing these associated information with the preset inference rule set, the system attempts to rule out superficial causes and reverse-deduce more likely inducing factors. For example, if the heart rate increase occurs at night, and the map shows that the patient has had sleep disorders recently (judged by the duration of night lighting and the sleep quality monitored by the wearable device) and has used sedative-hypnotic drugs, the inference module may conclude that the abnormal heart rate may be related to anxiety caused by insomnia or confusion in medication time, and generate corresponding risk warning signals. These are all extended implementation methods known to those of ordinary skill in the art.

[0049] The resource dynamics sandbox simulation module plays a key role before planning the service resource allocation plan. It builds a virtual environment to simulate the service resources of a specific community, including the supply capacity of medical staff, equipment, venues, etc. and the service demand predicted based on the patient behavior map and early warning information. This sandbox does not perform complex mathematical calculations, but simulates the process of discrete events. The simulation process takes into account the type, quantity, and available time period of resources, as well as the patient's service preferences and disease characteristics, in order to evaluate the pros and cons of different plans. For example, for the predicted regular follow-up needs of diabetic patients, the sandbox will simulate whether the existing doctor resources can meet these needs within the specified time period; for the home care needs triggered by sudden high-risk warnings, the sandbox simulates whether the nursing resources can respond in time. After the simulation, the system will generate a predictive report, pointing out the possible service gaps or resource redundancies in the simulated scenario, providing a basis for the final resource allocation decision.

[0050] The elastic priority scheduling module is responsible for dynamically managing the patient's service task queue. When the system receives a new service request or the status of an existing task changes, the module will adjust the execution order of the task based on a comprehensive evaluation mechanism. The evaluation mechanism takes multiple factors into consideration. The first is the patient's current risk level. Tasks for high-risk patients are usually given higher priority. Secondly, combined with the resource virtual occupancy rate predicted by the resource dynamic sandbox deduction module, if a certain key resource is expected to be very tight during a certain period of time, and a high-priority task happens to need this resource, the priority of this task will be further increased to ensure priority protection of resources. In addition, the algorithm will also refer to the patient's historical service records, their appointment preferences reflected in the map, and the geographical location of the required service resources. While meeting the urgency and importance, it improves the accuracy of resource matching and the efficiency of scheduling. The entire scheduling process is dynamically adjusted and can respond flexibly based on the information received in real time and the simulation prediction results.

[0051] The blockchain-based points management module provides a flexible and transparent incentive and service redemption mechanism. Patients can accumulate health points by performing specific behaviors that contribute to their own health (such as taking medicine on time, exercising regularly, actively participating in community activities, etc., and the completion of these behaviors is confirmed from multi-source data through the behavior correlation engine). These points are recorded on the blockchain, and the processes of accumulation, transfer, and use are transparent and traceable. The points system allows different types of points to be combined and used. For example, the points obtained from daily health check-ins can be combined with the points obtained from achieving phased rehabilitation goals to redeem higher-value services such as expert remote consultations. A dynamic pricing mechanism is adopted when redeeming community service resources with points; as an automated carrier for implementing rules, smart contracts will adjust the points redemption ratio according to the real-time supply and demand situation of community service resources. For example, if the reservation demand for a certain rehabilitation training service surges, resulting in resource shortages, the smart contract will automatically increase the number of points required to redeem this service; conversely, if the resource utilization rate is low, the required points will be reduced. This dynamic adjustment aims to guide patients' behaviors through economic signals, encourage the redemption of services during off-peak resource periods, thereby optimizing the overall allocation efficiency of resources, and motivating patients to manage their own health more proactively. These are all extended implementation methods known to those of ordinary skill in the art.

[0052] For those skilled in the art, it is obvious that the present invention is not limited to the details of the above exemplary embodiments, and without departing from the spirit or basic characteristics of the present invention, the present invention can be implemented in other specific forms.

[0053] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit them. Although the present invention has been described in detail with reference to the preferred embodiments, those of ordinary skill in the art should understand that the technical solutions of the present invention can be modified or equivalently replaced without departing from the spirit and scope of the technical solutions of the present invention.

Claims

1. A team building method for full-course case management, characterized by: The following steps are involved: Step 1: Integrate the patient medication information recorded in the medical information system and the patient life habit data from non-medical Internet of Things devices through a cross-domain patient behavior association engine. The non-medical Internet of Things devices include smart home devices and wearable health monitoring devices. The smart home devices include a lighting monitoring module, and the wearable health monitoring devices include at least a heart rate monitoring module and a location monitoring module. Use a natural language processing module to perform semantic analysis and feature extraction on the medication information and life habit data, and construct a patient behavior map containing the patient's medical behavior characteristics and non-medical behavior characteristics. The patient behavior map is used to characterize the patient's behavior patterns in the medical and life fields and the potential association between the two. Step 2: Based on the reverse reasoning algorithm, the real-time physiological data from the wearable health monitoring device is analyzed. When abnormal fluctuations in the physiological data are monitored, the possible causes of the abnormal fluctuations are reversely deduced. The reverse deduction at least associates and analyzes the non-medical behavior characteristics in the patient's behavior map to determine whether the patient is outside the scope of daily activities and may not carry medications, or associates the abnormal frequency of use of night lighting equipment with the use of insomnia drugs in the medication records, infers the relationship between insomnia and the patient's medication compliance, and generates a graded intervention warning signal containing potential behavioral inducements; Step 3: Build a dynamic resource sandbox simulation system. Before making a decision, the system simulates the impact of different resource allocation plans on the satisfaction of patients' service needs. Resources include at least community rehabilitation workers. The simulation process is based on historical patient service data and predicted future service needs. It allocates different quantities and types of service resources in a virtual environment, predicts service gaps and service redundancies within a preset time range, and generates a prediction report on the effect of resource allocation. Step 4: Use the elastic priority algorithm to dynamically adjust the priority of the patient service task queue according to the patient's risk level and the resource virtual occupancy rate predicted by the resource dynamic sandbox simulation system. For high-risk patients, prioritize the service resources that have been simulated and verified in the virtual sandbox and have a lower resource occupancy rate; Step 5, deploy a points system based on blockchain technology. The points system realizes the divisible combination of points and the dynamic pricing of community service resources through smart contracts, allowing patients to split the points they have earned into different proportions for redemption of different services. At the same time, the smart contract automatically adjusts the points redemption ratio according to the supply and demand relationship of community service resources. When the resources of rehabilitation therapists are tight, the points required for redemption will be increased, and vice versa.

2. A team building method for full course case management according to claim 1, characterized in that: In step 2, the reverse reasoning algorithm further includes: when it is monitored that the power consumption frequency of the smart home increases abnormally during non-normal sleep periods, the patient's medication records are correlated and analyzed. If there is a record of the use of sedatives and hypnotic drugs, a warning signal is generated that insomnia may cause a deviation in medication time; when the patient's geographic location data is monitored by a wearable health monitoring device to exceed a preset distance threshold of their daily activity range and the duration exceeds a preset time, a risk mark for not carrying regular medications is automatically triggered.

3. A team building method for full course case management according to claim 2, characterized in that: In step 4, the elastic priority algorithm also considers the patient's recent service history, appointment preferences, and geographic location distribution of community service resources when adjusting the priority of the patient service task queue.

4. A team building method for full course case management according to claim 1, characterized in that: It also includes step 6, which displays personalized service recommendations, risk warning information, and redeemable points and service options to patients through the user interface, and allows patients to make service appointments and redeem points according to their own needs.

5. A team building method for full course case management according to claim 4, characterized in that: The user interface also integrates augmented reality technology to provide patients with visual rehabilitation guidance, medication reminders, and community service navigation information.

6. A team building method for full course case management according to claim 1, characterized in that: In step 1, the cross-domain patient behavior association engine also integrates patient health-related discussion data from social media platforms, uses sentiment analysis and topic modeling techniques to explore patients' potential health needs and psychological states, and integrates the analysis results into the patient behavior map.

7. A team building system for full-course case management, characterized by: include: The cross-domain patient behavior association module is used to integrate patient medication information from the medical information system and patient lifestyle data from non-medical IoT devices, and use the natural language processing module to build a patient behavior map that includes patient medical behavior characteristics and non-medical behavior characteristics; The reverse reasoning analysis module is used to analyze the real-time physiological data from wearable health monitoring devices, reversely deduce the possible causes of abnormal fluctuations, and generate graded intervention warning signals containing potential behavioral inducements; The resource dynamic sandbox simulation module is used to simulate the impact of different resource allocation plans on patient service demand satisfaction before making decisions, and generate a prediction report on the resource allocation effect; The elastic priority scheduling module is used to dynamically adjust the priority of the patient service task queue according to the patient's risk level and the resource virtual occupancy rate predicted by the resource dynamic sandbox deduction module; The blockchain points management module is used to realize the splittable combination of points and dynamic pricing of community service resources through smart contracts; The data interaction and display module is used to display personalized service recommendations, risk warning information, redeemable points and service options to patients, and receive patients' service appointment and point redemption operation requests.

Images