Dynamic tracking and analyzing method and system for injury condition of sick and wounded

Abstract

The invention provides a method and a system for dynamically tracking and analyzing the injury condition of the sick and wounded. The method comprises the following steps: receiving a real-time injury data stream through a multi-source medical monitoring device, and generating a comprehensive wounded health record; on the basis of the comprehensive sick and wounded health archives, a deep time sequence generative adversarial network algorithm is applied, a potential time evolution mode is modeled to predict the future health trend, a causal inference technology is adopted, a causal relationship chain in sick and wounded health data is explored, it is ensured that an intervention measure effect is reflected, and a future health trend prediction result is generated; on the basis of the future health trend prediction result, an adaptive multi-scale map attention network algorithm is used, a social network interaction mode is captured, a personalized treatment planning method is adopted, optimal rehabilitation path planning is achieved, and a personalized treatment scheme is generated; and based on the personalized treatment scheme, continuously updating the health file of the sick and wounded, and generating a dynamic injury condition tracking database. According to the technical scheme provided by the invention, the injury tracking efficiency and accuracy are remarkably improved.

Description

Technical Field

[0001] The embodiments of the present application relate to the field of medical information technology, and in particular to a method and system for dynamically tracking and analyzing the injuries of patients and the sick. Background Art

[0002] With the rapid development of medical information technology and the widespread application of big data and artificial intelligence technologies, in the modern medical environment, the health management of the injured and sick needs to cover multiple scenarios from hospitals to homes, including emergency treatment, hospitalization, rehabilitation care, and remote monitoring. These application scenarios put forward real-time, comprehensive, and personalized requirements for technology.

[0003] At present, common management plans for the injured and sick mainly rely on traditional electronic medical record systems and limited remote monitoring technologies, usually including the use of fixed medical equipment to regularly collect the physiological parameters of the injured and sick, and store them in local or cloud databases; using simple statistical analysis methods, such as mean and standard deviation, to conduct preliminary processing of the historical data of the injured and sick; formulating treatment plans based on the doctor's experience and conventional clinical guidelines, without sufficient consideration of individual differences and social environmental factors.

[0004] Although the existing solutions have met the basic management needs of the injured and sick to a certain extent, there are still significant deficiencies; traditional methods are difficult to capture subtle changes in the health status of the injured and sick, especially in the development of complex diseases, which limits the ability of early warning and precise intervention, resulting in insufficient accuracy in dynamic tracking of injuries. In addition, the existing solutions fail to fully utilize deep learning algorithms to model the potential evolution of the health status of the injured and sick, and the prediction of future health trends is not accurate enough; the impact of external social environmental factors on the rehabilitation of the injured and sick is rarely considered, making the design of personalized treatment plans lack comprehensiveness and flexibility. Summary of the invention

[0005] The embodiments of the present application provide a method and system for dynamically tracking and analyzing the injuries of the wounded and sick, so as to solve the problem of insufficient accuracy of dynamic tracking of injuries in the prior art.

[0006] In a first aspect, an embodiment of the present application provides a method for dynamically tracking and analyzing the condition of an injured or sick person, comprising:

[0007] Through multi-source medical monitoring equipment, real-time injury and illness data streams are received, historical medical records of patients from different medical institutions are integrated, and comprehensive health records of patients are generated;

[0008] Based on the comprehensive health records of the injured and sick, a deep time series generative adversarial network algorithm is used to model the potential time evolution pattern of the health status of the injured and sick to predict future health trends. Causal inference technology is used to explore the causal relationship chain in the health data of the injured and sick to ensure that the prediction model reflects the effect of intervention measures and generates future health trend prediction results;

[0009] Based on the future health trend prediction results, an adaptive multi-scale graph attention network algorithm is used to integrate multi-level social environment influencing factors, capture social network interaction patterns, adopt personalized treatment planning methods, dynamically adjust treatment strategies, achieve optimal rehabilitation path planning, and generate personalized treatment plans;

[0010] Based on the personalized treatment plan, the health records of the injured and sick are continuously updated, all changes and responses during treatment are recorded, and a dynamic injury tracking database is generated.

[0011] Optionally, based on the comprehensive health records of the injured and sick, a deep time series generative adversarial network algorithm is used to model the potential time evolution pattern of the health status of the injured and sick to predict future health trends, and causal inference technology is used to explore the causal relationship chain in the health data of the injured and sick to ensure that the prediction model reflects the effect of intervention measures and generate future health trend prediction results, including:

[0012] Based on the comprehensive health records of the injured and sick, pre-analyze the historical and real-time health data of the injured and sick, identify key time series features, and generate a time series feature set;

[0013] Based on the time series feature set, a deep time series generative adversarial network algorithm is used to model the potential time evolution pattern of the health status of the injured and sick, simulate the changes in the health status of the injured and sick in different time periods, and generate a health status evolution model;

[0014] Based on the health status evolution model, causal inference technology is used to explore and process the causal relationship chain in the health data of the injured and sick, build a causal map, ensure that the effect of the intervention measures is reflected, and generate a causal relationship chain report;

[0015] Based on the causal chain report, counterfactual reasoning is used to simulate changes in health status under different intervention measures, optimize and adjust model parameters, and generate future health trend prediction results.

[0016] Optionally, based on the time series feature set, a deep time series generative adversarial network algorithm is used to model the potential time evolution pattern of the health status of the injured and sick, simulate the changes in the health status of the injured and sick in different time periods, and generate a health status evolution model, including:

[0017] Based on the time series feature set, the time series decomposition technology is used to separate the time series feature set into trend and random components, identify key evolution laws and abnormal points in different time periods, and generate a time series evolution law diagram;

[0018] Based on the time series evolution law diagram, a deep time series generative adversarial network algorithm is used to initialize the generator parameters, introduce a random noise vector, model the potential time evolution pattern of the health status of the injured and sick, simulate the changes in the health status of the injured and sick in different time periods, and generate health status evolution samples;

[0019] Based on the health status evolution samples, the synthetic samples are distinguished from the actual historical data through a binary classification task, the parameters in the generation process are adjusted according to the feedback, and the weights are updated through a cyclic iteration to generate an optimized health status evolution result;

[0020] Based on the optimized health status evolution results, mean square error and mean absolute error are selected as evaluation indicators, and multiple rounds of verification and fine-tuning are performed to generate a health status evolution model.

[0021] Optionally, based on the health status evolution model, causal inference technology is used to explore and process the causal relationship chain in the health data of the injured and sick, build a causal graph, ensure that the effect of the intervention measures is reflected, and generate a causal relationship chain report, including:

[0022] Based on the health status evolution model, the causal relationship between historical and real-time health data of the injured and sick is explored, and the structural equation modeling method is used to identify potential causal relationships and generate a preliminary causal map;

[0023] Based on the preliminary causal map, the effects of different intervention measures are evaluated through the propensity score matching method, the actual impact of each intervention measure on the change of health status is quantified, and the quantitative results of the intervention measure effects are generated;

[0024] Based on the quantitative results of the intervention measures, combined with the specific conditions and treatment responses of the injured and sick, the correlation between different intervention measures and changes in health status is analyzed to generate a prediction model for the impact of intervention measures;

[0025] Based on the intervention impact prediction model, all causal relationships and intervention effect information are deeply integrated to generate a causal chain report.

[0026] Optionally, based on the future health trend prediction results, an adaptive multi-scale graph attention network algorithm is used to integrate multi-level social environment influencing factors, capture social network interaction patterns, adopt a personalized treatment planning method, dynamically adjust the treatment strategy, achieve optimal rehabilitation path planning, and generate a personalized treatment plan, including:

[0027] Based on the future health trend prediction results, integrate multi-level social environmental influencing factors, evaluate the impact of external factors, and generate a social environmental impact assessment report;

[0028] Based on the social environment impact assessment report, combined with the specific conditions of the injured and sick and their social network interaction patterns, an adaptive multi-scale graph attention network algorithm is used to analyze the interaction relationship and time-varying laws between individuals, quantify the influence of social interaction factors, and generate a social interaction impact assessment report;

[0029] Based on the social interaction impact assessment report, a personalized treatment planning method is adopted to consider the personal preferences and social interaction characteristics of the injured and sick, simulate the effects of different treatment strategies, select the optimal treatment path, and generate a personalized intervention strategy;

[0030] Based on the personalized intervention strategy, a multi-dimensional efficacy evaluation system is constructed to conduct regular evaluation and feedback to generate personalized treatment plans.

[0031] Optionally, based on the social environment impact assessment report, combined with the specific condition of the injured and sick and the social network interaction pattern, an adaptive multi-scale graph attention network algorithm is used to analyze the interaction relationship and time-varying law between individuals, quantify the influence of social interaction factors, and generate a social interaction impact assessment report, including:

[0032] Based on the social environmental impact assessment report, a comprehensive analysis of the specific conditions and social interaction patterns of the injured and sick is conducted to generate a comprehensive input data set;

[0033] Based on the comprehensive input data set, an adaptive multi-scale graph attention network algorithm is used to set graph structure nodes to represent individuals, assign edge weights to reflect the intensity and nature of interactions, capture static social network structures, dynamically reflect time-varying trends in interactions, and generate a graph structure model;

[0034] Based on the graph structure model, multi-scale analysis is performed to capture social interaction patterns at different scales, identify key figures and relationship chains, and generate interaction pattern analysis results;

[0035] Based on the interaction pattern analysis results, the influence of each social interaction factor is quantified, the positive and negative effects of each interaction type are evaluated, and a social interaction impact evaluation report is generated.

[0036] Optionally, based on the personalized treatment plan, the health records of the injured and sick are continuously updated, all changes and responses during treatment are recorded, and a dynamic injury tracking database is generated, including:

[0037] Based on the personalized treatment plan, the daily treatment process of the injured and sick is carefully recorded to ensure that all treatment-related data are systematically collected and a detailed treatment log is generated;

[0038] Based on the detailed treatment log and combined with the improvement of the patient's self-reported symptoms, the patient's physiological index changes during treatment are monitored in real time to form a continuous health status record and generate a detailed physiological response data set;

[0039] Based on the detailed physiological response data set, the recovery process of the injured and sick is evaluated in stages, the treatment effects at different stages are quantified, and compared and analyzed with the preset rehabilitation goals to generate a stage-by-stage evaluation report;

[0040] Based on the periodic assessment reports, the health records of the injured and sick are iteratively updated to generate a dynamic injury tracking database.

[0041] In a second aspect, the present application embodiment provides a system for dynamically tracking and analyzing the condition of a patient, including:

[0042] The receiving module is used to receive real-time injury and illness data streams through multi-source medical monitoring equipment, integrate historical medical records of the injured and sick between different medical institutions, and generate comprehensive health records of the injured and sick;

[0043] A prediction module is used to model the potential time evolution pattern of the health status of the injured and sick based on the comprehensive health records of the injured and sick, using a deep time series generative adversarial network algorithm to predict future health trends, and to use causal inference technology to explore the causal relationship chain in the health data of the injured and sick, to ensure that the prediction model reflects the effect of intervention measures, and to generate future health trend prediction results;

[0044] An adjustment module is used to use an adaptive multi-scale graph attention network algorithm based on the future health trend prediction results, integrate multi-level social environment influencing factors, capture social network interaction patterns, adopt a personalized treatment planning method, dynamically adjust the treatment strategy, achieve optimal rehabilitation path planning, and generate a personalized treatment plan;

[0045] The update module is used to continuously update the health records of the injured and sick based on the personalized treatment plan, record all changes and responses during the treatment, and generate a dynamic injury tracking database.

[0046] In a third aspect, an embodiment of the present application provides a computing device, comprising a processing component and a storage component; the storage component stores one or more computer instructions; the one or more computer instructions are used to be called and executed by the processing component to implement a method for dynamic tracking and analysis of injuries of injured and sick persons as described in the first aspect.

[0047] In a fourth aspect, an embodiment of the present application provides a computer storage medium storing a computer program, which, when executed by a computer, implements a method for dynamically tracking and analyzing the injuries of injured and sick persons as described in the first aspect.

[0048] In an embodiment of the present application, a real-time injury and illness data stream is received through multi-source medical monitoring equipment, and the historical medical records of the injured and sick between different medical institutions are integrated to generate a comprehensive health file of the injured and sick; based on the comprehensive health file of the injured and sick, a deep time series generative adversarial network algorithm is used to model the potential time evolution pattern of the health status of the injured and sick to predict future health trends, and causal inference technology is used to explore the causal relationship chain in the health data of the injured and sick to ensure that the prediction model reflects the effect of intervention measures and generates future health trend prediction results; based on the future health trend prediction results, an adaptive multi-scale graph attention network algorithm is used to integrate multi-level social environment influencing factors, capture social network interaction patterns, adopt personalized treatment planning methods, dynamically adjust treatment strategies, achieve optimal rehabilitation path planning, and generate personalized treatment plans; based on the personalized treatment plan, the health files of the injured and sick are continuously updated, all changes and responses during treatment are recorded, and a dynamic injury tracking database is generated. By integrating the data streams of multi-source medical monitoring equipment with historical medical records between different medical institutions, a detailed and continuous health record can be generated for each patient. This not only helps to fully understand the evolution of the patient's health status, but also supports accurate predictions based on deep time series generative adversarial networks and causal inference technology, effectively improving the accuracy of predictions of future health trends of the patient and ensuring that the prediction model can reflect the effectiveness of intervention measures. Finally, by dynamically tracking changes in injuries and continuously updating health records, a closed-loop management from data collection to the formulation of personalized treatment plans is achieved, greatly improving the efficiency and accuracy of health management of the patient.

[0049] Furthermore, by pre-analyzing the historical and real-time health data of the injured and sick, key time series features were identified, further enhancing the accuracy and reliability of the health status evolution model. Using the deep time series generative adversarial network algorithm to model the potential time evolution pattern, and combining causal inference technology to explore the causal chain in health data, it can ensure that the prediction model not only reflects the current health trends, but also accurately evaluates the effectiveness of intervention measures. The causal chain report and optimized future health trend prediction results generated by this process provide a scientific basis for the formulation of subsequent personalized treatment plans, greatly improving the rationality and effectiveness of predictions and decisions.

[0050] Furthermore, an adaptive multi-scale graph attention network is adopted to fuse multi-level social environment impact factors, taking into account the influence of external factors on the rehabilitation of the wounded and sick. By quantifying the degree of influence of social interaction factors and simulating the effects of different treatment strategies, this method not only promotes the customization of personalized treatment plans but also ensures that the selected treatment path is optimal. A multi-dimensional efficacy evaluation system is constructed for regular evaluation and feedback, enabling the treatment plan to be flexibly adjusted according to the actual situation, thereby maximizing the treatment effect and improving the rehabilitation quality of the wounded and sick. This approach not only enhances the pertinence and effectiveness of treatment but also provides a solid foundation for achieving long-term and effective rehabilitation management.

[0051] These aspects or other aspects of the present application will be more clearly understood in the following description of the embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0052] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the following will briefly introduce the drawings required for the description of the embodiments or the prior art. Obviously, the drawings in the following description are some embodiments of the present application. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0053] Figure 1 It is a flowchart of a method for dynamically tracking and analyzing the injury conditions of the wounded and sick provided by an embodiment of the present application;

[0054] Figure 2 It is a schematic structural diagram of a system for dynamically tracking and analyzing the injury conditions of the wounded and sick provided by an embodiment of the present application;

[0055] Figure 3 It is a schematic structural diagram of a computing device provided by an embodiment of the present application. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0056] In order to enable those skilled in the art to better understand the solution of the present application, the following will clearly and completely describe the technical solutions in the embodiments of the present application with reference to the drawings in the embodiments of the present application.

[0057] In some of the processes described in the specification and claims of this application and the above-mentioned figures, multiple operations that appear in a specific order are included, but it should be clearly understood that these operations may not be executed in the order in which they appear in this article or executed in parallel. The serial numbers of the operations, such as 101, 102, etc., are only used to distinguish between different operations, and the serial numbers themselves do not represent any execution order. In addition, these processes may include more or fewer operations, and these operations may be executed in sequence or in parallel. It should be noted that the descriptions of "first", "second", etc. in this article are used to distinguish different messages, devices, modules, etc., do not represent the order of precedence, and do not limit the "first" and "second" to be different types.

[0058] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative work are within the scope of protection of this application.

[0059] Figure 1 A flowchart of a method for dynamically tracking and analyzing the condition of a patient is provided for an embodiment of the present application, such as Figure 1 As shown, the method includes:

[0060] 101. Receive real-time injury and illness data streams through multi-source medical monitoring equipment, integrate historical medical records of patients from different medical institutions, and generate comprehensive health records of patients;

[0061] In this step, multi-source medical monitoring equipment includes electrocardiographs, blood pressure monitors, blood glucose meters, smart wearable devices, etc. These devices can continuously collect physiological parameters of the injured and sick, such as heart rate, blood pressure, blood oxygen saturation, etc.

[0062] Real-time injury and illness data stream refers to the physiological parameters of the injured and sick transmitted in real time by multi-source medical monitoring equipment, which is used to instantly reflect the current health status of the injured and sick.

[0063] Historical medical records cover the medical records, diagnostic reports, imaging data and other information generated when the injured and sick visit different medical institutions, and are used to provide information on the injured and sick’s past health status and treatment history.

[0064] The comprehensive health records of the injured and sick integrate the real-time injury and illness data stream with the data set of historical medical records. It not only contains the basic information of the injured and sick (such as age and gender), but also includes detailed medical history (such as previous diseases, surgical experience) and treatment response records. This comprehensive health record provides a solid data foundation for subsequent analysis.

[0065] In the embodiment of the present application, it is assumed that a hospital first introduces a variety of intelligent medical monitoring equipment to collect the vital signs data of hospitalized patients in real time; secondly, the hospital information system establishes a data sharing mechanism with multiple cooperative medical institutions to ensure that the historical medical records of the patients can be seamlessly connected; thirdly, an advanced data processing platform is used to clean, standardize and integrate all collected data; finally, a comprehensive health record of the patients is generated that integrates real-time monitoring data and historical medical records, and a detailed electronic medical record system is established for each patient.

[0066] 102. Based on the comprehensive health records of the injured and sick, use the deep time series generative adversarial network algorithm to model the potential time evolution pattern of the health status of the injured and sick to predict future health trends, use causal inference technology to explore the causal relationship chain in the health data of the injured and sick, ensure that the prediction model reflects the effect of intervention measures, and generate future health trend prediction results;

[0067] In this step, the deep time series generative adversarial network algorithm is an algorithm that combines the generative adversarial network (GAN) with the time series modeling technology to simulate the potential pattern of the health status of the injured and sick over time. The algorithm optimizes the model parameters through the competition mechanism of the generator and the discriminator, and can capture complex time series characteristics.

[0068] Time evolution patterns are the patterns of changes in the health status of the injured and sick over time, which are used to predict future health trends. These patterns may include different situations such as worsening, improvement or stability of the condition.

[0069] Causal inference technology uses statistical methods to identify causal relationships between variables in data. It aims to explore causal chains in health data and ensure that predictive models can reflect the effects of interventions. This method can distinguish between correlation and causality, thereby providing more reliable predictions.

[0070] A causal chain is a series of cause-effect relationships that describes how one factor directly or indirectly affects another factor. In health data of the sick and injured, the causal chain helps understand which factors actually affect health outcomes and evaluate the effectiveness of interventions.

[0071] The future health trend prediction results are prediction outputs generated based on time evolution patterns and causal chains. They can not only predict the future health status of the injured and sick, but also evaluate the possible effects of different intervention measures, thereby providing a basis for personalized treatment.

[0072] In the embodiment of the present application, it is assumed that a research institution first applies a deep time series generative adversarial network algorithm to model the time evolution pattern of the health status of the injured and sick based on the existing comprehensive health records of the injured and sick; secondly, causal inference technology is used to identify key health influencing factors and their interactions, and a causal graph is constructed; thirdly, the changes in the health status of the injured and sick under different intervention measures are simulated according to the causal graph, and the model parameters are optimized and adjusted; finally, detailed future health trend prediction results are generated, and a personalized health management plan is formulated for each injured and sick.

[0073] Optionally, in step 102, based on the comprehensive health records of the injured and sick, a deep time series generative adversarial network algorithm is used to model the potential time evolution pattern of the health status of the injured and sick to predict future health trends, and causal inference technology is used to explore the causal relationship chain in the health data of the injured and sick to ensure that the prediction model reflects the effect of intervention measures, and generate future health trend prediction results, including: based on the comprehensive health records of the injured and sick, the historical and real-time health data of the injured and sick are pre-analyzed, key time series features are identified, and a time series feature set is generated; based on the time series feature set, a deep time series generative adversarial network algorithm is used to model the potential time evolution pattern of the health status of the injured and sick, simulate the changes in the health status of the injured and sick in different time periods, and generate a health status evolution model; based on the health status evolution model, causal inference technology is used to explore the causal relationship chain in the health data of the injured and sick, construct a causal map, ensure that the effect of intervention measures is reflected, and generate a causal relationship chain report; based on the causal relationship chain report, counterfactual reasoning is used to simulate the changes in health status under different intervention measures, optimize and adjust model parameters, and generate future health trend prediction results.

[0074] In this step, the time series feature set is a set of key features identified by pre-analyzing the historical and real-time health data of the injured and sick. These features can be the trend of changes in physiological indicators, abnormal values ​​or periodic patterns, etc., which are used for further modeling and processing.

[0075] The health status evolution model is a model built based on a time series feature set using a deep time series generative adversarial network algorithm to simulate the changes in the health status of the injured and sick in different time periods. This model helps predict future health trends and evaluate the effectiveness of intervention measures.

[0076] The causal graph is a chart generated after exploring the causal chain in the health data of the injured and sick through causal inference technology. It shows the direct and indirect effects between various factors and helps understand which factors actually affect health outcomes.

[0077] The causal chain report is a document summarizing the causal map, which describes in detail the causal chain in the health data of the injured and sick, ensuring that the predictive model can reflect the effect of the intervention measures.

[0078] Counterfactual reasoning is a statistical method used to simulate possible outcomes under different hypothetical conditions. In this context, it is used to evaluate the impact of different interventions on the health status of the injured and sick, thereby optimizing the adjustment of model parameters.

[0079] Firstly, based on the comprehensive health records of the wounded and sick, the historical and real-time health data of the wounded and sick were pre-analyzed to identify the key time series features and generate a time series feature set; secondly, based on the time series feature set, the deep time series generative adversarial network algorithm was used to model the potential time evolution pattern of the health status of the wounded and sick, simulate the changes in different time periods, and form a health status evolution model; thirdly, based on the health status evolution model, causal inference technology was used to explore the causal chain in the health data and construct a causal graph to ensure that the model can accurately reflect the effect of the intervention measures; finally, according to the causal chain report, counterfactual reasoning was used to simulate the changes in health status under different intervention measures, optimize and adjust the model parameters, and generate future health trend prediction results.

[0080] Optionally, based on the time series feature set, a deep time series generative adversarial network algorithm is used to model the potential time evolution pattern of the health status of the injured and sick, simulate the changes in the health status of the injured and sick in different time periods, and generate a health status evolution model, including: based on the time series feature set, separating into trend and random components through time series decomposition technology, identifying key evolution laws and abnormal points in different time periods, and generating a time series evolution law graph; based on the time series evolution law graph, a deep time series generative adversarial network algorithm is used to initialize generator parameters, introduce random noise vectors, model the potential time evolution pattern of the health status of the injured and sick, simulate the changes in the health status of the injured and sick in different time periods, and generate health status evolution samples; based on the health status evolution samples, a binary classification task is used to distinguish between synthetic samples and actual historical data, parameters in the generation process are adjusted according to feedback, and weights are updated in a loop iteration to generate an optimized health status evolution result; based on the optimized health status evolution result, mean square error and mean absolute error are selected as evaluation indicators, multiple rounds of verification and fine-tuning are performed to generate a health status evolution model.

[0081] Based on the health status evolution model, causal inference technology is used to explore and process the causal chain in the health data of the injured and sick, construct a causal map, ensure that the effects of intervention measures are reflected, and generate a causal chain report, including: based on the health status evolution model, causal relationship exploration is carried out on the historical and real-time health data of the injured and sick, and the structural equation modeling method is used to identify potential causal relationships, and a preliminary causal map is generated; based on the preliminary causal map, the effects of different intervention measures are evaluated through the propensity score matching method, the actual impact of each intervention measure on the change of health status is quantified, and the quantitative results of the effects of intervention measures are generated; based on the quantitative results of the effects of intervention measures, combined with the specific conditions of the injured and sick and the treatment response, the correlation between different intervention measures and changes in health status is analyzed, and an intervention measure impact prediction model is generated; based on the intervention measure impact prediction model, all causal relationships and intervention measure effect information are deeply integrated to generate a causal chain report.

[0082] In this step, time series decomposition technology is a method to decompose complex time series data into trend components and random components, helping to identify key evolution patterns and anomalies in different time periods.

[0083] The time series evolution law diagram is a diagram generated based on the time series decomposition technology, which shows the evolution law and abnormal conditions of the health status of the injured and sick in different time periods. It provides an intuitive reference for subsequent modeling.

[0084] The health status evolution samples are simulated by introducing random noise vectors and adjusting the generator parameters to simulate the changes in the health status of the injured and sick in different time periods.

[0085] The binary classification task is to distinguish between synthetic samples and actual historical data. It is used to evaluate the performance of the generator and adjust the parameters in the generation process based on feedback. The cycle is iterated to update the weights and finally generate an optimized health state evolution result.

[0086] Mean square error and mean absolute error are commonly used evaluation indicators to measure the difference between predicted values ​​and true values. Multiple rounds of verification and fine-tuning ensure the accuracy and stability of the model.

[0087] The health status evolution model is the final model generated after multiple rounds of verification and fine-tuning, which can accurately simulate the changes in the health status of the injured and sick in different time periods.

[0088] Structural equation modeling is a statistical method used to identify causal relationships between variables, construct preliminary causal maps, and show the direct and indirect effects between various factors.

[0089] Propensity score matching is a method for evaluating the effectiveness of interventions. It quantifies the actual impact of each intervention on changes in health status by matching individuals under similar conditions.

[0090] The intervention impact prediction model is a prediction model generated by deeply integrating all relevant information based on the causal chain and intervention effect information, and is used to guide personalized treatment plans.

[0091] In the embodiments of the present application, firstly, based on the time series feature set, the trend and random components are separated by time series decomposition technology, the key evolution laws and abnormal points in different time periods are identified, and a time series evolution law diagram is generated; secondly, based on the time series evolution law diagram, the deep time series generative adversarial network algorithm is used to initialize the generator parameters, and a random noise vector is introduced to model the potential time evolution pattern of the health status of the injured and sick, simulate the changes in different time periods, and generate health status evolution samples; thirdly, based on the health status evolution samples, the synthetic samples and the actual historical data are distinguished by a binary classification task, the parameters in the generation process are adjusted according to the feedback, the weights are updated in a loop iteratively, and an optimized health status evolution result is generated; finally, based on the optimized health status evolution result, the mean square error and the mean absolute error are selected as evaluation indicators, and multiple rounds of verification and fine-tuning are carried out to finally generate a health status evolution model.

[0092] Suppose a smart medical platform aims to improve the accuracy of dynamic tracking and analysis of the injuries of the wounded and sick, and introduces advanced deep time series generative adversarial network algorithms and causal inference technology to optimize personalized treatment plans; first, based on the time series feature set, the platform uses time series decomposition technology to separate the trend component and random component of the health status of the wounded and sick, applies the deep time series generative adversarial network algorithm to initialize the generator parameters, and introduces random noise vectors to model the potential time evolution pattern of the health status of the wounded and sick, successfully simulates the changes in different time periods, and generates multiple health status evolution samples; secondly, the platform distinguishes the synthetic samples from the actual historical data through the binary classification task, continuously adjusts the parameters in the generation process according to the feedback, and iterates in a cycle to Update the weights, and based on the optimization results, select mean square error and mean absolute error as evaluation indicators, and conduct multiple rounds of verification and fine-tuning; thirdly, based on the health status evolution model, the platform explores the causal relationship between the historical and real-time health data of the injured and sick, applies the structural equation modeling method to identify potential causal relationships, and evaluates the effects of different intervention measures through the propensity score matching method. The actual impact of each intervention measure on the change in health status is quantified, and the quantitative results of the intervention effect are generated; finally, combined with the specific condition of the injured and sick and the treatment response, the correlation between different intervention measures and changes in health status is analyzed, and all causal relationships and intervention effect information are deeply integrated to generate a detailed causal chain report, which provides a scientific basis for personalized treatment plans.

[0093] This application takes into account that the existing technology has problems such as inaccurate modeling of the evolution pattern of the health status of the injured and the failure to fully consider historical influences and periodic errors when predicting future health trends, so the invention embodiment proposes this optional solution. By introducing a deep time series generative adversarial network algorithm combined with frequency domain analysis and nonlinear conversion functions, it aims to solve the above technical problems, improve prediction accuracy and enhance the generalization ability of the model.

[0094] Optionally, based on the time series evolution law graph, a deep time series generative adversarial network algorithm is used to initialize generator parameters, introduce a random noise vector, model the potential time evolution pattern of the health status of the injured and sick, simulate the changes in the health status of the injured and sick in different time periods, and generate health status evolution samples, including:

[0095] Based on the time series evolution law diagram, the sliding window technology is used to extract feature fragments in different time spans, and the periodicity and trend components in the time series are captured by applying frequency domain analysis to identify local changes at multiple scales to generate potential feature representations;

[0096] The latent feature representation is calculated using the following formula:

[0097]

[0098] Among them, h t is the potential feature representation; x t is the time series data of the health status of the injured and sick at time t; z is the introduced random noise vector; D(z) is the preprocessing function of the noise vector used to enhance the generalization ability of the model; W g , A g , B g and V g is the weight matrix of the generator; b g is the bias term; σ is the activation function ReLU or Sigmoid, which is used to introduce nonlinearity;

[0099] Based on the potential feature representation, a nonlinear conversion function and a dynamic weight adjustment mechanism are introduced to map to the future health status change space, and a historical impact factor is introduced to quantify the impact of past health status on future evolution to generate a prediction increment;

[0100] The predicted increment is calculated by the following formula:

[0101]

[0102] Among them, Δp t+1 It is the prediction increment rather than the directly generated sample; h t is the potential feature representation; T is a complex nonlinear conversion function that accepts multiple inputs to simulate the changing trend of the health status of the injured and sick in different time periods; Ch is the coefficient matrix reflecting historical influence; λ is the attenuation factor used to adjust the influence of historical data; t is the time step; F(h t ) is a function that further processes the latent feature representation in order to capture deeper patterns; η, ω, and φ are the amplitude, angular frequency, and phase difference, respectively, which are used to define the periodic error term η·cos(ω·t+φ), which is used to simulate the uncertainty of unobserved data points; α and β are parameters that adjust the exponential decay term in the denominator; γ and δ are parameters that adjust the additional periodic fraction; θ is the angular frequency of the additional periodic fraction; ||h t || is the norm of the latent feature representation;

[0103] Based on the predicted increment, the existing health status data is fused through the Kalman filtering method to eliminate the noise in the predicted increment, simulate the effects of different intervention measures, evaluate the specific impact on the evolution of the health status of the injured and sick, and generate a health status evolution sample.

[0104] The method aims to use sliding window technology and frequency domain analysis to capture the periodicity and trend components in the time series, and map the potential features to the future health state change space through complex nonlinear conversion functions, quantify the impact of past health states on future evolution, and introduce periodic error terms to simulate the uncertainty of unobserved data points. In addition, the existing health state data is fused through the Kalman filter method to eliminate the noise in the prediction increment, thereby generating a more reliable health state evolution sample.

[0105] In the latent feature representation, random noise is introduced into the term W g ·σ(A g ·(x t +D(z))):By introducing the random noise vector z, the model's generalization ability is enhanced, so that the model can still maintain a high prediction accuracy when facing unseen data; nonlinear transformation term The nonlinear transformation introduces additional dynamics, increasing the model's ability to capture complex patterns; the periodic variation term V g ·sin(π·x t ): Simulate the periodic changes in the health status of the injured and sick, such as physiological phenomena such as circadian rhythms, to ensure that the model can capture these regular changes;

[0106] Among them, W g ,A g ,B g ,V g is the weight matrix of the generator, which is automatically learned through the training process; b g is the bias term, which is set when the model is initialized; σ is the activation function ReLU or Sigmoid, which is selected according to the specific application scenario; x tis the time series data of the health status of the injured at time t, collected from the real-time monitoring equipment of the injured; z is the introduced random noise vector, generated by the random number generator; D(z) is the preprocessing function of the pre-constructed noise vector, which is used to enhance the generalization ability of the model;

[0107] In the prediction increment, the nonlinear transformation term T(h t ,C h exp(-λ t) x t ,F(h t )):Accept multiple inputs through complex nonlinear conversion function T to simulate the changing trend of the health status of the injured and sick in different time periods, ensuring that the model can capture long-term and short-term changes; attenuate the historical impact term C h exp(-λ t) x t : The influence of historical data is adjusted through exponential decay, so that the more recent historical data has a greater influence, while the influence of more distant data gradually weakens; Periodic error term η·cos(ω·t+φ): Define the periodic error term to simulate the uncertainty of unobserved data points and ensure that the model can adapt to actual fluctuations; Exponential decay term Adjust the parameters of the exponential decay term in the denominator to make the model more robust to outliers; additional periodic fractions Adjust the parameters of the additional periodic fractions to further refine the model's ability to capture periodic changes;

[0108] Among them, C h ,λ,η,ω,φ,α,β,γ,δ,θ are automatically learned through the training process; h t is the potential feature representation, which is calculated by the potential feature representation formula; x t It is the time series data of the health status of the injured and sick at time t, collected from the real-time monitoring equipment of the injured and sick;

[0109] Suppose a smart medical platform is used in the cardiac rehabilitation center of a large hospital, specifically to optimize the personalized treatment plan for patients after heart surgery and track their recovery progress in real time;

[0110] Assume that all parameters used in formula 1 are assumed; x t =0.78; z=0.35; D(z)=0.22; W g =0.82; A g =0.64; B g =0.55; V g =0.91; b g =0.21;

[0111] h t=0.82·σ(0.64·(0.78+0.22)+0.55·log(0.78 2 +1)+0.21)+0.91·sin(π·0.78)=0.76;

[0112] Assumption C h =0.88; λ=0.12; t=3; F(h t )=0.74; η=0.45; ω=0.32; φ=0.11; α=0.23; β=0.31; γ=0.12; δ=0.06; θ=0.21;

[0113]

[0114] Assuming the threshold is set to 0.45, since the calculated result 0.47 is greater than the set threshold, it indicates that the health status of the patient after heart surgery will be significantly improved in the future. The above process not only improves the accuracy of dynamic tracking of the patient's injury, but also provides solid data support and technical guarantee for the effective implementation of personalized treatment plans, ensuring that each patient can obtain the most optimized rehabilitation path. Through the above steps, the accurate monitoring and scientific planning of the patient's rehabilitation process is ensured, the reliability and scientific nature of the rescue operation is improved, and the accuracy and response speed of the entire health monitoring system are enhanced.

[0115] 103. Based on the above-mentioned future health trend prediction results, an adaptive multi-scale graph attention network algorithm is used to integrate multi-level social environment influencing factors, capture social network interaction patterns, adopt personalized treatment planning methods, dynamically adjust treatment strategies, achieve optimal rehabilitation path planning, and generate personalized treatment plans;

[0116] In this step, the adaptive multi-scale graph attention network algorithm is a graph attention network algorithm that integrates multi-level social environment influencing factors. It is used to capture social network interaction patterns and dynamically adjust treatment strategies. The algorithm can process graph structure data at different scales and assign different weights to nodes and edges to better reflect actual social relationships.

[0117] Multi-level social environmental influencing factors include family support, community resources, work environment and other factors, which are used to evaluate the impact of external factors on the rehabilitation of the injured and sick. These factors can be measured in quantitative and qualitative ways, such as the support of family members and the availability of community medical services.

[0118] Social network interaction patterns reflect the interactions between patients and their social circle members and are used to quantify the influence of social interaction factors. These patterns can help understand the relationship between individual behavior and social support, and how they work together on the recovery process of patients.

[0119] The personalized treatment planning method is a method of formulating personalized treatment plans based on the specific condition and social interaction characteristics of the injured and patients to ensure that the selected treatment path is optimal. This method takes into account the personal preferences, lifestyle and social environment of the injured and patients to improve the effectiveness of treatment and patient compliance.

[0120] The personalized treatment plan combines the personal preferences and social interaction characteristics of the injured and patients, simulates the effects of different treatment strategies, and selects the optimal treatment path to guide the rehabilitation process of the injured and patients. The plan not only includes drug treatment, but may also involve various forms of intervention measures such as psychological counseling and physical therapy.

[0121] In the embodiment of the present application, it is assumed that a medical institution first evaluates the impact of the social environment of the injured and sick on their rehabilitation based on the prediction results of future health trends, and generates a social environment impact assessment report; secondly, based on the specific condition of the injured and sick and the social network interaction pattern, an adaptive multi-scale graph attention network algorithm is used to analyze the interaction relationship and time-varying laws between individuals, and quantify the influence of social interaction factors; thirdly, a personalized treatment planning method is adopted, taking into account the personal preferences and social interaction characteristics of the injured and sick, simulating the effects of different treatment strategies, and selecting the optimal treatment path; finally, a multi-dimensional efficacy evaluation system is constructed, regular evaluation and feedback are provided, and personalized treatment plans are generated to ensure the effectiveness and pertinence of the treatment.

[0122] Optionally, the method in step 103, based on the future health trend prediction results, uses an adaptive multi-scale graph attention network algorithm, integrates multi-level social environment influencing factors, captures social network interaction patterns, adopts personalized treatment planning methods, dynamically adjusts treatment strategies, achieves optimal rehabilitation path planning, and generates personalized treatment plans, including: based on the future health trend prediction results, integrates multi-level social environment influencing factors, evaluates the impact of external factors, and generates a social environment impact assessment report; based on the social environment impact assessment report, combined with the specific condition of the injured and the social network interaction pattern, uses an adaptive multi-scale graph attention network algorithm to analyze the interaction relationship and time-varying laws between individuals, quantifies the degree of influence of social interaction factors, and generates a social interaction impact assessment report; based on the social interaction impact assessment report, adopts a personalized treatment planning method, considers the personal preferences and social interaction characteristics of the injured, simulates the effects of different treatment strategies, selects the optimal treatment path, and generates a personalized intervention strategy; based on the personalized intervention strategy, constructs a multidimensional efficacy evaluation system, conducts regular evaluation and feedback, and generates a personalized treatment plan.

[0123] Among them, based on the social environment impact assessment report, combined with the specific condition of the injured and the social network interaction pattern, the adaptive multi-scale graph attention network algorithm is used to analyze the interaction relationship and time-varying law between individuals, quantify the influence of social interaction factors, and generate a social interaction impact assessment report, including: based on the social environment impact assessment report, comprehensively analyze the specific condition of the injured and the social interaction pattern, and generate a comprehensive input data set; based on the comprehensive input data set, use the adaptive multi-scale graph attention network algorithm, set graph structure nodes to represent individuals, assign edge weights to reflect the intensity and nature of the interaction, capture the static social network structure, dynamically reflect the time-varying trend of the interaction, and generate a graph structure model; based on the graph structure model, perform multi-scale analysis, capture social interaction patterns at different scales, identify key figures and relationship chains, and generate interaction pattern analysis results; based on the interaction pattern analysis results, quantify the influence of each social interaction factor, evaluate the positive and negative effects of each interaction type, and generate a social interaction impact assessment report.

[0124] In this step, multi-level social environmental influencing factors including family support, community resources, work environment and other factors are used to evaluate the impact of external factors on the rehabilitation of the injured and sick.

[0125] The social and environmental impact assessment report is a detailed report generated based on the prediction results of future health trends to evaluate the impact of external social and environmental factors on the rehabilitation of the injured and sick.

[0126] The comprehensive input dataset is a data set that combines social and environmental impact assessment reports, specific conditions of the injured and sick, and social interaction patterns, and is used to build a graph structure model.

[0127] Graph structure models are graphical representations generated based on comprehensive input data sets, where nodes represent individuals and edge weights reflect the intensity and nature of interactions.

[0128] The results of the interaction pattern analysis are the key figures and relationship chains of social interaction patterns at different scales identified through multi-scale analysis. These results help quantify the influence of each social interaction factor and evaluate the positive and negative effects of each type of interaction.

[0129] The social interaction impact assessment report is a detailed report generated based on the results of interaction pattern analysis, which quantifies the impact of each social interaction factor and evaluates the positive and negative effects of each interaction type.

[0130] The personalized treatment planning method is a method to develop personalized treatment plans based on the specific conditions and social interaction characteristics of the injured and patients, ensuring that the selected treatment path is the best.

[0131] The personalized intervention strategy combines the personal preferences and social interaction characteristics of the injured and patients, simulates the effects of different treatment strategies, and selects the optimal treatment path to guide the rehabilitation process of the injured and patients.

[0132] The multidimensional efficacy evaluation system is a comprehensive evaluation framework that regularly evaluates the effectiveness of personalized treatment plans and makes adjustments based on feedback. The system ensures that the treatment plan can be continuously optimized to achieve the best rehabilitation effect.

[0133] In the embodiments of the present application, firstly, based on the prediction results of future health trends, multi-level social environment influencing factors are integrated to evaluate the impact of external factors on the rehabilitation of the injured and sick, and generate a social environment impact assessment report; secondly, based on the social environment impact assessment report, combined with the specific condition of the injured and sick and the social network interaction pattern, an adaptive multi-scale graph attention network algorithm is used to analyze the interaction relationship and time-varying law between individuals, quantify the influence of social interaction factors, and generate a social interaction impact assessment report; thirdly, based on the social interaction impact assessment report, a personalized treatment planning method is adopted, taking into account the personal preferences and social interaction characteristics of the injured and sick, simulating the effects of different treatment strategies, selecting the optimal treatment path, and generating a personalized intervention strategy; finally, based on the personalized intervention strategy, a multi-dimensional efficacy evaluation system is constructed, regular evaluation and feedback are carried out, and a personalized treatment plan is generated.

[0134] Suppose a high-end intelligent medical platform aims to improve the accuracy of dynamic tracking and analysis of the injuries of the wounded and sick, and optimize personalized treatment plans through advanced adaptive multi-scale graph attention network algorithms and personalized treatment planning methods; first, based on the prediction results of future health trends, the platform integrates multi-level social environment influencing factors, evaluates the impact of external factors on the rehabilitation of the wounded and sick, and combines the specific conditions of the wounded and sick with the social network interaction patterns to generate a comprehensive input data set through comprehensive analysis; secondly, based on this data set, the adaptive multi-scale graph attention network algorithm is applied, the graph structure nodes are set to represent individuals, and the edge weights are assigned to reflect the intensity and nature of the interaction, capturing the static The social network structure dynamically reflects the time-varying trend of interaction, conducts multi-scale analysis, captures social interaction patterns at different scales, identifies key figures and relationship chains, and generates interaction pattern analysis results; secondly, based on the interaction pattern analysis results, quantifies the influence of each social interaction factor, evaluates the positive and negative effects of each interaction type, adopts a personalized treatment planning method, considers the personal preferences and social interaction characteristics of the injured and sick, simulates the effects of different treatment strategies, selects the optimal treatment path, and generates a personalized intervention strategy; finally, based on the personalized intervention strategy, a multi-dimensional efficacy evaluation system is constructed, regular evaluation and feedback are carried out, and ultimately a personalized treatment plan is generated.

[0135] This application takes into account that the existing technology does not adequately capture the interaction patterns of the injured and sick in social networks and does not adequately combine static and dynamic factors, so the invention embodiment proposes this optional solution. By introducing an adaptive multi-scale graph attention network algorithm, it aims to solve the above technical problems, improve the model's ability to understand complex social network structures, and accurately reflect the time-varying trends of interactions.

[0136] Optionally, based on the comprehensive input data set, an adaptive multi-scale graph attention network algorithm is used to set graph structure nodes to represent individuals, assign edge weights to reflect interaction intensity and nature, capture static social network structure, dynamically reflect interaction time-varying trends, and generate a graph structure model, including:

[0137] Based on the comprehensive input data set, the node attribute values ​​are standardized, and potential association patterns between nodes are identified through statistical analysis;

[0138] Perform similarity measurement to quantify the closeness between nodes and identify dense subgraphs in the network to generate edge weights;

[0139] The edge weight is calculated using the following formula:

[0140]

[0141] Among them, w ij is the edge weight, representing the strength and nature of the interaction; s ij is the association measure between nodes i and j in the static social network structure; t ij is a time factor that dynamically reflects the time-varying trend of interaction; α, β, γ are adaptive adjustment parameters used to capture the influence of different scales; x ik ,x jk is the eigenvalue of nodes i and j on the kth feature dimension; k is the index of the feature dimension; K is the total dimension of the feature vector;

[0142] Based on the edge weights, the information flow between nodes is re-weighted to ensure that the strength of information transmission matches the closeness of the relationship between nodes. The influence of neighboring nodes is collected through an aggregation mechanism, and a nonlinear activation function is introduced to capture complex dependencies to generate an updated representation.

[0143] The updated representation is calculated using the following formula:

[0144]

[0145] in, is the updated representation of node i at layer l+1; j and k are the indexes of each node in the neighbor set N(i) of node i; N(i) is the neighbor set of node i; w ij is the edge weight, representing the strength and nature of the interaction; and is the query function, which measures the similarity between nodes; is the gating function used to regulate the information flow and takes into account the time factor t ij ; b i The bias vector provides independent learning capabilities for each node; σ is the activation function ReLU or Sigmoid, which is used to increase the nonlinearity of the expression to better capture complex patterns; ⊙ is the element-wise multiplication operation; λ and μ are the parameters of the time decay effect; d ij is the shortest path distance between nodes i and j; is the current representation of node i at layer l; is the current representation of node j at layer l;

[0146] Based on the update representation, the loss function is minimized through the back-propagation method, the weights and other parameters are continuously adjusted to converge to the optimal state, the overall network parameter configuration is optimized, regularization technology is introduced to prevent overfitting, and a graph structure model is generated.

[0147] This method aims to use standardized node attribute values ​​and identify potential association patterns through statistical analysis, perform similarity measurement to quantify the closeness of nodes, and identify dense subgraphs to generate edge weights; calculate edge weights through complex formulas to ensure that the intensity of information transmission matches the closeness of the relationship between nodes; finally, minimize the loss function through the backpropagation method, continuously adjust the weights and other parameters to converge to the optimal state, optimize the overall network parameter configuration, and prevent overfitting; improve the modeling accuracy of the social network interaction patterns of the wounded and sick, ensure that the prediction model can reflect the effect of intervention measures, and quantify the impact of different scales through complex formulas.

[0148] In the edge weight, the static association measure α·log(s ij +1): By measuring the association s between nodes i and j in the static social network structure ij , capturing long-term stable social relationships; the time factor term β·sin(t ij ): through time factor t ij Reflects the time-varying trend of interaction and captures the short-term volatility of social relationships; characteristic distance term The similarity between nodes is measured by the difference in feature dimensions to ensure that the model can capture the influence of multi-dimensional features;

[0149] Among them, s ij ,t ij ,α,β,γ are automatically learned through the training process; x ik ,x jk is the eigenvalue of nodes i and j on the kth feature dimension, extracted from the health records of the injured and sick; K is the total dimension of the feature vector, determined according to the actual application.

[0150] In the update representation, the weighted information flow term The information of neighbor nodes is integrated in a weighted manner to ensure that the strength of information transmission matches the closeness of the relationship between nodes; the bias term b i :Provide independent learning capabilities for each node to enhance the expressiveness of the model; time decay effect term Introducing the time decay effect so that more recent historical data has a greater influence, while the influence of more distant data gradually weakens;

[0151] in, b i ,λ,μ,d ij Learned automatically through the training process; is the current representation of nodes j and i at layer l, calculated from edge weights; N(i) is the set of neighbors of node i, determined by the social network structure; σ is the function form selected according to the specific application scenario;

[0152] Assume there are two participants A and B, the frequency of online communication in the past month is 0.6 (after standardization), the time factor (reflecting the time-varying trend of interaction) is 0.4; the adaptive adjustment parameters α = 0.5, β = 0.3, γ = 0.2; the total number of feature dimensions K = 3; and the eigenvalues ​​in each feature dimension are x Ak =[0.8,0.7,0.9],x Bk =[0.7,0.6,0.8];

[0153]

[0154] Assume that for node A, the current representation at layer l is The neighbor set N(A) includes nodes B and C, where Query Function Gating Function Bias vector b A = 0.2, time decay effect parameter λ = 0.3, μ = 0.4; the shortest path distance d between nodes A and B and C AB =0.5,d AC =0.6;

[0155]

[0156] Suppose a smart medical platform is applied to a community center focusing on mental health management, aiming to optimize psychological counseling services by analyzing the interaction patterns between patients and tracking their mental health status in real time;

[0157] Assuming that the threshold is set to 0.9, since the calculated result 0.89 is less than the set threshold, it indicates that the mental health of this participant may need more attention and support in the future. This is because the lower update representation value reflects that the support intensity obtained by the participant in his social network is relatively weak, which may affect his mental health recovery progress. Through the above steps, the effective monitoring and timely intervention of the mental health status of the participants are ensured, the quality and efficiency of psychological counseling services are improved, the accuracy and response speed of the entire mental health monitoring system are enhanced, and each participant is guaranteed to receive the rehabilitation guidance and support that best suits him.

[0158] 104. Based on the personalized treatment plan, the health records of the injured and sick are continuously updated, all changes and responses during the treatment are recorded, and a dynamic injury tracking database is generated.

[0159] In this step, the patient's health record is continuously updated to add new information to the patient's health record during the patient's treatment process, ensuring that all the latest treatment progress, physiological changes and responses are recorded in a timely manner. This step ensures that the health record is always up to date and supports real-time decision-making and evaluation.

[0160] Changes and responses during treatment are to systematically collect and save various changes in the injured and patients during treatment and their responses to treatment measures, including changes in physiological indicators (such as blood pressure, heart rate), improvement or deterioration of symptoms, side effects of drugs, etc., for subsequent analysis and evaluation.

[0161] The injury dynamic tracking database is a structured and easy-to-query database specifically used to store and manage all relevant data of the injured and sick throughout the entire treatment cycle. The database not only records the basic information and treatment process of the injured and sick, but also includes detailed information such as the specific content, time point, medication status and dosage adjustment of each treatment.

[0162] In the embodiment of the present application, it is assumed that a medical institution first begins to implement a treatment plan for a specific patient based on a personalized treatment plan; secondly, during the treatment process, medical staff continuously updates the health records of the patient, recording the specific content, time point, medication and dosage adjustment of each treatment; thirdly, the system automatically records all physiological indicator changes, self-reported symptom improvement and other detailed information of the patient during the treatment; finally, all these data are integrated into a dynamic injury tracking database, which not only records the entire process of the patient from admission to discharge, but also supports subsequent research and data analysis, providing a reference for future clinical decision-making.

[0163] Optionally, in step 104, based on the personalized treatment plan, the health records of the injured and sick are continuously updated, all changes and responses during the treatment are recorded, and a dynamic injury tracking database is generated, including: based on the personalized treatment plan, the daily treatment process of the injured and sick is recorded in detail to ensure the systematic collection of all treatment-related data and generate a detailed treatment log; based on the detailed treatment log, combined with the improvement of the injured and sick self-reported symptoms, the changes in physiological indicators of the injured and sick during treatment are monitored in real time to form a continuous health status record and generate a detailed physiological response data set; based on the detailed physiological response data set, the recovery process of the injured and sick is evaluated in stages, the treatment effects at different stages are quantified, and compared and analyzed with the preset rehabilitation goals to generate a staged evaluation report; based on the staged evaluation report, the health records of the injured and sick are iteratively updated to generate a dynamic injury tracking database.

[0164] In this step, the detailed treatment log is the result of meticulously recording the daily treatment process of the patient, ensuring the systematic collection of all treatment-related data, including the specific content, time point, medication and dosage adjustment of each treatment, to support subsequent analysis and evaluation.

[0165] Continuous health status records are formed by real-time monitoring of changes in physiological indicators during treatment, combined with the improvement of the patient's self-reported symptoms. This helps to form a coherent time series that comprehensively reflects the patient's health evolution process.

[0166] The detailed physiological response data set is a data set generated by continuous health status records, which specifically includes various physiological changes (such as blood pressure and heart rate) that occur during the treatment of the injured and sick and their responses to treatment measures.

[0167] The phased assessment report is a document generated after a phased assessment of the recovery process of the injured or sick based on a detailed physiological response data set. It quantifies the treatment effects at different stages and compares and analyzes them with the preset rehabilitation goals, providing a basis for adjusting the treatment strategy.

[0168] Iteratively update the health records of the injured and sick to continuously update the health records of the injured and sick based on the findings in the periodic assessment reports to ensure that all the latest health information is recorded in a timely manner.

[0169] In the embodiments of the present application, firstly, based on the personalized treatment plan, the daily treatment process of the patient is recorded in detail to ensure the systematic collection of all treatment-related data and generate a detailed treatment log; secondly, based on the detailed treatment log and in combination with the improvement of the patient's self-reported symptoms, the changes in physiological indicators of the patient during treatment are monitored in real time to form a continuous health status record and generate a detailed physiological response data set; thirdly, based on the detailed physiological response data set, the patient's recovery process is evaluated in stages, the treatment effects at different stages are quantified, and compared with the preset rehabilitation goals to generate a stage-by-stage evaluation report; finally, based on the stage-by-stage evaluation report, the patient's health records are iteratively updated to generate a dynamic injury tracking database.

[0170] Suppose a large hospital aims to improve the accuracy of dynamic tracking and analysis of the injuries of the wounded and sick through systematic data collection and analysis, and ensure the effective implementation of personalized treatment plans; first, based on the personalized treatment plan, the platform carefully records the daily treatment process of the wounded and sick, ensures the systematic collection of all treatment-related data, and generates detailed treatment logs, covering the specific content, time points, medication conditions, dosage adjustments and other information of each treatment; secondly, based on these detailed treatment logs, combined with the improvement of symptoms reported by the wounded and sick, the changes in physiological indicators of the wounded and sick during treatment are monitored in real time, forming a continuous health status record, and generating detailed life records. First, based on the detailed physiological response datasets, these datasets comprehensively reflect the health evolution of the injured and sick during the treatment; secondly, based on the detailed physiological response datasets, the platform conducts a phased evaluation of the recovery process of the injured and sick, quantifies the treatment effects at different stages, and compares and analyzes them with the preset rehabilitation goals, and generates phased evaluation reports. These reports provide a scientific basis for adjusting treatment strategies; finally, based on the phased evaluation reports, the platform iteratively updates the health records of the injured and sick, ensuring that all the latest health information is recorded in a timely manner, and finally generates a complete injury dynamic tracking database, which provides valuable data resources for future clinical research and decision-making.

[0171] In summary, steps 101 to 104 cover the complete process from the collection of patient information, preliminary diagnosis, feature recognition to the generation of personalized treatment plans, aiming to provide an efficient, accurate and comprehensive intelligent medical support system to meet the needs of dynamic tracking and personalized treatment of patients.

[0172] Figure 2 A schematic diagram of a system for dynamically tracking and analyzing the condition of a patient is provided for the present application. Figure 2 As shown, the device comprises:

[0173] The receiving module 21 is used to receive real-time injury and illness data streams through multi-source medical monitoring equipment, integrate historical medical records of the injured and sick between different medical institutions, and generate comprehensive health records of the injured and sick;

[0174] The prediction module 22 is used to use the deep time series generative adversarial network algorithm to model the potential time evolution pattern of the health status of the injured and sick based on the comprehensive health records of the injured and sick, so as to predict future health trends, and use causal inference technology to explore the causal relationship chain in the health data of the injured and sick, ensure that the prediction model reflects the effect of the intervention measures, and generate future health trend prediction results;

[0175] An adjustment module 23 is used to use an adaptive multi-scale graph attention network algorithm based on the future health trend prediction results, integrate multi-level social environment influencing factors, capture social network interaction patterns, adopt a personalized treatment planning method, dynamically adjust the treatment strategy, achieve optimal rehabilitation path planning, and generate a personalized treatment plan;

[0176] The updating module 24 is used to continuously update the health records of the injured and sick based on the personalized treatment plan, record all changes and responses during the treatment, and generate a dynamic injury tracking database.

[0177] Figure 2 The system for tracking and analyzing the dynamic condition of the injured and sick can be executed Figure 1 The implementation principle and technical effect of the method for dynamic tracking and analysis of the injuries of the injured and sick described in the embodiment shown are not repeated here. The specific way in which each module and unit performs operations in the system for dynamic tracking and analysis of the injuries of the injured and sick in the above embodiment has been described in detail in the embodiment of the method, and will not be elaborated here.

[0178] In one possible design, Figure 2 The system for dynamically tracking and analyzing the injuries of the injured and sick in the embodiment shown can be implemented as a computing device, such as Figure 3 As shown, the computing device may include a storage component 31 and a processing component 32;

[0179] The storage component 31 stores one or more computer instructions, wherein the one or more computer instructions are called and executed by the processing component 32 .

[0180] The processing component 32 is used to: receive real-time injury and illness data streams through multi-source medical monitoring equipment, integrate historical medical records of the injured and sick between different medical institutions, and generate comprehensive health records of the injured and sick; based on the comprehensive health records of the injured and sick, use the deep time series generative adversarial network algorithm to model the potential time evolution pattern of the health status of the injured and sick to predict future health trends, use causal inference technology to explore the causal relationship chain in the health data of the injured and sick, ensure that the prediction model reflects the effect of intervention measures, and generate future health trend prediction results; based on the future health trend prediction results, use the adaptive multi-scale graph attention network algorithm, integrate multi-level social environment influencing factors, capture social network interaction patterns, use personalized treatment planning methods, dynamically adjust treatment strategies, achieve optimal rehabilitation path planning, and generate personalized treatment plans; based on the personalized treatment plans, continuously update the health records of the injured and sick, record all changes and responses during treatment, and generate a dynamic injury tracking database.

[0181] The processing component 32 may include one or more processors to execute computer instructions to complete all or part of the steps in the above method. Of course, the processing component may also be implemented by one or more application-specific integrated circuits (ASICs), digital signal processors (DSPs), digital signal processing devices (DSPDs), programmable logic devices (PLDs), field programmable gate arrays (FPGAs), controllers, microcontrollers, microprocessors or other electronic components to perform the above method.

[0182] The storage component 31 is configured to store various types of data to support operations at the terminal. The storage component can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic memory, flash memory, magnetic disk or optical disk.

[0183] Of course, the computing device may also include other components, such as input / output interfaces, display components, communication components, etc.

[0184] The input / output interface provides an interface between the processing component and the peripheral interface module, which may be an output device, an input device, etc.

[0185] The communication component is configured to facilitate, among other things, wired or wireless communications between the computing device and other devices.

[0186] Among them, the computing device can be a physical device or an elastic computing host provided by a cloud computing platform, etc. In this case, the computing device can refer to a cloud server, and the above-mentioned processing components, storage components, etc. can be basic server resources rented or purchased from the cloud computing platform.

[0187] The present application also provides a computer storage medium storing a computer program, wherein the computer program can achieve the above-mentioned Figure 1 A method for dynamically tracking and analyzing the condition of a wounded or sick person in the illustrated embodiment.

[0188] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the systems, devices and units described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0189] The device embodiments described above are merely illustrative, wherein the units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the scheme of this embodiment. Ordinary technicians in this field can understand and implement it without paying creative labor.

[0190] Through the description of the above implementation methods, those skilled in the art can clearly understand that each implementation method can be implemented by means of software plus a necessary general hardware platform, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solution is essentially or the part that contributes to the prior art can be embodied in the form of a software product, and the computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, a disk, an optical disk, etc., including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in each embodiment or some parts of the embodiments.

[0191] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit it. Although the present application has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims

1. A method for dynamically tracking and analyzing the condition of a wounded or sick person, characterized in that: include: Through multi-source medical monitoring equipment, real-time injury and illness data streams are received, historical medical records of patients from different medical institutions are integrated, and comprehensive health records of patients are generated; Based on the comprehensive health records of the injured and sick, a deep time series generative adversarial network algorithm is used to model the potential time evolution pattern of the health status of the injured and sick to predict future health trends. Causal inference technology is used to explore the causal relationship chain in the health data of the injured and sick to ensure that the prediction model reflects the effect of intervention measures and generates future health trend prediction results; Based on the future health trend prediction results, an adaptive multi-scale graph attention network algorithm is used to integrate multi-level social environment influencing factors, capture social network interaction patterns, adopt personalized treatment planning methods, dynamically adjust treatment strategies, achieve optimal rehabilitation path planning, and generate personalized treatment plans; Based on the personalized treatment plan, the health records of the injured and sick are continuously updated, all changes and responses during treatment are recorded, and a dynamic injury tracking database is generated.

2. The method according to claim 1, characterized in that Based on the comprehensive health records of the injured and sick, the deep time series generative adversarial network algorithm is used to model the potential time evolution pattern of the health status of the injured and sick to predict future health trends. The causal inference technology is used to explore the causal relationship chain in the health data of the injured and sick to ensure that the prediction model reflects the effect of the intervention measures and generate future health trend prediction results, including: Based on the comprehensive health records of the injured and sick, pre-analyze the historical and real-time health data of the injured and sick, identify key time series features, and generate a time series feature set; Based on the time series feature set, a deep time series generative adversarial network algorithm is used to model the potential time evolution pattern of the health status of the injured and sick, simulate the changes in the health status of the injured and sick in different time periods, and generate a health status evolution model; Based on the health status evolution model, causal inference technology is used to explore and process the causal relationship chain in the health data of the injured and sick, build a causal map, ensure that the effect of the intervention measures is reflected, and generate a causal relationship chain report; Based on the causal chain report, counterfactual reasoning is used to simulate changes in health status under different intervention measures, optimize and adjust model parameters, and generate future health trend prediction results.

3. The method according to claim 2, characterized in that Based on the time series feature set, a deep time series generative adversarial network algorithm is used to model the potential time evolution pattern of the health status of the injured and sick, simulate the changes in the health status of the injured and sick in different time periods, and generate a health status evolution model, including: Based on the time series feature set, the time series decomposition technology is used to separate the time series feature set into trend and random components, identify key evolution laws and abnormal points in different time periods, and generate a time series evolution law diagram; Based on the time series evolution law diagram, a deep time series generative adversarial network algorithm is used to initialize the generator parameters, introduce a random noise vector, model the potential time evolution pattern of the health status of the injured and sick, simulate the changes in the health status of the injured and sick in different time periods, and generate health status evolution samples; Based on the health status evolution samples, the synthetic samples are distinguished from the actual historical data through a binary classification task, the parameters in the generation process are adjusted according to the feedback, and the weights are updated through a cyclic iteration to generate an optimized health status evolution result; Based on the optimized health status evolution results, mean square error and mean absolute error are selected as evaluation indicators, and multiple rounds of verification and fine-tuning are performed to generate a health status evolution model.

4. The method according to claim 2, characterized in that: Based on the health status evolution model, causal inference technology is used to explore and process the causal relationship chain in the health data of the injured and sick, build a causal map, ensure that the effect of the intervention measures is reflected, and generate a causal relationship chain report, including: Based on the health status evolution model, the causal relationship between historical and real-time health data of the injured and sick is explored, and the structural equation modeling method is used to identify potential causal relationships and generate a preliminary causal map; Based on the preliminary causal map, the effects of different interventions are evaluated through the propensity score matching method, the actual impact of each intervention on the change of health status is quantified, and the quantitative results of the effects of the interventions are generated; Based on the quantitative results of the intervention measures, combined with the specific conditions and treatment responses of the injured and sick, the correlation between different intervention measures and changes in health status is analyzed to generate a prediction model for the impact of intervention measures; Based on the intervention impact prediction model, all causal relationships and intervention effect information are deeply integrated to generate a causal chain report.

5. The method according to claim 1, characterized in that Based on the future health trend prediction results, the adaptive multi-scale graph attention network algorithm is used to integrate multi-level social environment influencing factors, capture social network interaction patterns, adopt personalized treatment planning methods, dynamically adjust treatment strategies, achieve optimal rehabilitation path planning, and generate personalized treatment plans, including: Based on the future health trend prediction results, integrate multi-level social environmental influencing factors, evaluate the impact of external factors, and generate a social environmental impact assessment report; Based on the social environment impact assessment report, combined with the specific conditions of the injured and sick and their social network interaction patterns, an adaptive multi-scale graph attention network algorithm is used to analyze the interaction relationship and time-varying laws between individuals, quantify the influence of social interaction factors, and generate a social interaction impact assessment report; Based on the social interaction impact assessment report, a personalized treatment planning method is adopted to consider the personal preferences and social interaction characteristics of the injured and sick, simulate the effects of different treatment strategies, select the optimal treatment path, and generate a personalized intervention strategy; Based on the personalized intervention strategy, a multi-dimensional efficacy evaluation system is constructed to conduct regular evaluation and feedback to generate personalized treatment plans.

6. The method according to claim 5, characterized in that Based on the social environment impact assessment report, combined with the specific conditions of the injured and sick and the social network interaction pattern, the adaptive multi-scale graph attention network algorithm is used to analyze the interaction relationship and time-varying law between individuals, quantify the influence of social interaction factors, and generate a social interaction impact assessment report, including: Based on the social environmental impact assessment report, a comprehensive analysis of the specific conditions and social interaction patterns of the injured and sick is conducted to generate a comprehensive input data set; Based on the comprehensive input data set, an adaptive multi-scale graph attention network algorithm is used to set graph structure nodes to represent individuals, assign edge weights to reflect the intensity and nature of interactions, capture static social network structures, dynamically reflect time-varying trends in interactions, and generate a graph structure model; Based on the graph structure model, multi-scale analysis is performed to capture social interaction patterns at different scales, identify key figures and relationship chains, and generate interaction pattern analysis results; Based on the interaction pattern analysis results, the influence of each social interaction factor is quantified, the positive and negative effects of each interaction type are evaluated, and a social interaction impact evaluation report is generated.

7. The method according to claim 1, characterized in that Based on the personalized treatment plan, the health records of the injured and sick are continuously updated, all changes and responses during treatment are recorded, and a dynamic injury tracking database is generated, including: Based on the personalized treatment plan, the daily treatment process of the injured and sick is carefully recorded to ensure that all treatment-related data are systematically collected and a detailed treatment log is generated; Based on the detailed treatment log and combined with the improvement of the patient's self-reported symptoms, the patient's physiological index changes during treatment are monitored in real time to form a continuous health status record and generate a detailed physiological response data set; Based on the detailed physiological response data set, the recovery process of the injured and sick is evaluated in stages, the treatment effects at different stages are quantified, and compared and analyzed with the preset rehabilitation goals to generate a stage-by-stage evaluation report; Based on the periodic assessment reports, the health records of the injured and sick are iteratively updated to generate a dynamic injury tracking database.

8. A system for tracking and analyzing the dynamic condition of the injured and sick, characterized in that: include: The receiving module is used to receive real-time injury and illness data streams through multi-source medical monitoring equipment, integrate historical medical records of the injured and sick between different medical institutions, and generate comprehensive health records of the injured and sick; A prediction module is used to model the potential time evolution pattern of the health status of the injured and sick based on the comprehensive health records of the injured and sick, using a deep time series generative adversarial network algorithm to predict future health trends, and to use causal inference technology to explore the causal relationship chain in the health data of the injured and sick, to ensure that the prediction model reflects the effect of intervention measures, and to generate future health trend prediction results; An adjustment module is used to use an adaptive multi-scale graph attention network algorithm based on the future health trend prediction results, integrate multi-level social environment influencing factors, capture social network interaction patterns, adopt a personalized treatment planning method, dynamically adjust the treatment strategy, achieve optimal rehabilitation path planning, and generate a personalized treatment plan; The update module is used to continuously update the health records of the injured and sick based on the personalized treatment plan, record all changes and responses during the treatment, and generate a dynamic injury tracking database.

9. A computing device, characterized in that It comprises a processing component and a storage component; the storage component stores one or more computer instructions; the one or more computer instructions are used to be called and executed by the processing component to implement a method for dynamic tracking and analysis of the injuries of the injured and sick as described in any one of claims 1 to 7.

10. A computer storage medium, characterized in that: A computer program is stored, and when the computer program is executed by a computer, a method for dynamically tracking and analyzing the injury condition of a patient as claimed in any one of claims 1 to 7 is implemented.

Images