Informatization management system and method for stroke patients in neurology department

Abstract

The invention discloses a neurology stroke patient informatization management system and method, and the method comprises the following steps: building a standardized multi-mode data set through collecting images, physiological signals, gene information, electronic medical records and other data; a causal reasoning method is adopted to identify pathological features and causal relationships thereof, a condition causal graph is generated, confounding factors are eliminated, and data analysis precision is optimized; based on a cross-modal representation learning method, extracting sharing features of the multi-modal data, and establishing a personalized health portrait; the influence of different intervention measures on disease development is evaluated by combining a time sequence causal reasoning method, and a rehabilitation path is dynamically adjusted by utilizing a reinforcement learning method; an optimal rehabilitation path is calculated through a causal dynamic programming method, feedback data are collected in real time in the rehabilitation process, and an intervention strategy is optimized. The intelligent level and the precision level of stroke patient management can be improved, personalized intervention and rehabilitation path optimization are achieved, and the rehabilitation effect of the patient is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of information management, and in particular to an information management system and method for neurology stroke patients. Background Art

[0002] With the rapid development of medical informatization, the management methods of neurological stroke patients are also evolving. However, the existing stroke patient management system still has many technical bottlenecks in information integration, personalized intervention and rehabilitation path optimization. The disease progression of stroke patients involves multimodal data, and the heterogeneity and information isolation between different data sources seriously affect the effective use of data. In addition, the existing management methods lack accurate disease analysis and personalized rehabilitation intervention strategies, which makes it difficult to optimize the rehabilitation effect of patients.

[0003] In the existing technology, traditional stroke management systems usually rely on rule-based decision models based on a single data source or experience-guided rehabilitation programs. These methods often fail to provide efficient and accurate management solutions when faced with complex and changing disease progression. Specifically, the existing technology has the following main defects: 1. Data silos and inconsistency: The existing stroke management system cannot effectively integrate multimodal data such as images, physiological signals, and genetic information, resulting in data isolation and information inconsistency, which affects the accuracy of disease analysis.

[0004] 2. Lack of causal reasoning analysis: Traditional disease assessment methods mostly rely on statistical models, which make it difficult to eliminate confounding factors and cannot accurately model the evolution of the disease, affecting the scientific nature of personalized intervention decisions.

[0005] 3. Difficulty in aligning cross-modal data: Existing methods lack an effective cross-modal learning mechanism, making it difficult to extract shared features from different data sources, resulting in inaccurate individual health portraits and the inability to provide accurate personalized intervention plans.

[0006] 4. Insufficient optimization of intervention measures: Traditional stroke management mostly adopts static rehabilitation pathway planning and fails to make full use of intelligent optimization technologies such as reinforcement learning. As a result, the rehabilitation intervention strategy lacks dynamic adjustment capabilities and is difficult to adapt to real-time changes in the patient's condition.

[0007] 5. Lack of long-term optimization of rehabilitation pathways: Existing rehabilitation programs do not fully consider the long-term effects of different interventions and lack causal dynamic programming methods to optimize rehabilitation pathways, which may lead to prolonged rehabilitation periods or unsatisfactory recovery effects.

[0008] Therefore, how to provide an information management system and method for neurological stroke patients is an urgent problem that technical personnel in this field need to solve. Summary of the invention

[0009] One purpose of the present invention is to propose an information management system and method for neurological stroke patients. The present invention makes full use of multimodal data fusion, temporal causal reasoning and dynamic rehabilitation path optimization technology, and describes in detail the methods of stroke patient condition evolution modeling, personalized intervention optimization and dynamic adjustment of rehabilitation paths. It has the advantages of high data integration, high intervention accuracy and intelligent rehabilitation optimization.

[0010] According to an embodiment of the present invention, the neurology stroke patient information management system and method include the following steps: S1, data acquisition module, collects imaging data, physiological signals, genetic information and electronic medical record data of neurological stroke patients, and performs standardized processing and storage, removes redundant information, and constructs a structured multimodal data set; S2, causal reasoning analysis module, based on structured multimodal data sets, uses structural causal learning methods to analyze pathological characteristics, eliminate confounding factors, build a causal model of disease evolution, and generate a disease causal graph; S3, cross-modal learning module, based on the disease causal graph, uses a self-supervised cross-modal representation learning method to extract shared features of structured multimodal datasets, perform feature alignment and multimodal mapping, and generate a patient health profile; S4, disease evolution analysis module, based on the patient's health profile and disease causal diagram, combines temporal causal reasoning to evaluate the impact of different intervention measures on the development of the disease and form an intervention plan; S5, data fusion module, performs cross-modal consistency check on the disease causal diagram, patient health profile and intervention plan, uses adaptive weighting strategy to adjust data contribution, and generates rehabilitation path decision data; S6, rehabilitation optimization decision module, based on rehabilitation pathway decision data, uses causal dynamic programming methods to simulate the long-term impact of different intervention plans, and dynamically adjusts the intervention plan during the execution of the rehabilitation pathway.

[0011] Optionally, the S2 specifically includes: S21, a data parsing unit, receiving a structured multimodal data set, including imaging data, physiological signals, genetic information, and electronic medical record data, performing modality separation on the structured multimodal data set, extracting time series features and static features, and generating a standardized pathology data matrix; S22, causal relationship modeling unit, based on the standardized pathological data matrix, uses structural causal learning methods to build a causal model of disease evolution, identify the causal relationship between variables, and generate a causal association matrix ,in Representation variables right The causal influence weight of the variable and variables They represent disease-related variables respectively; S23, confounding factor elimination unit, based on the causal association matrix, uses the causal intervention inference method to calculate the direct causal effects between different variables to eliminate confounding factors and generate an adjusted causal association matrix: ; in, In the causal relationship matrix Under the constraints, eliminate confounding factors After the influence, the variable About the disease status The causal effect of Under the combined effects of variables, confounding factors and causal association matrix, the disease state The probability of In the causal relationship matrix Under the constraints, confounding factors The prior probability distribution of ; S24, disease evolution modeling unit, based on the adjusted causal association matrix , combined with time series analysis methods to calculate the disease status , and establish a temporal causal relationship network to evaluate the disease development path; S25, causal graph generation unit, constructs a disease causal graph based on the disease evolution causal model ,in is a set of variables, The causal relationship is the time-dependent causal relationship of the temporal causal relationship network, and the disease causal graph is passed to the cross-modal learning module.

[0012] Optionally, the S23 specifically includes: S231, confounding factor identification unit, based on causal association matrix , identify confounding factors , using the conditional independence test method to evaluate the variables and disease status The independence between them is calculated by calculating the conditional independence probability Identify confounding factors Whether to change the variable and disease status The causal relationship between them has an impact; S232, causal effect calculation unit, based on the causal association matrix and confounding factors , using causal intervention inference methods to calculate variables About the disease status Direct causal effect, eliminating confounding factors The influence of the factors on the causal effect is generated after eliminating the confounding factors. S233, an adjustment matrix generation unit, generating an adjusted causal association matrix based on the causal effect calculation results after eliminating confounding factors ,in Eliminate confounding factors Finally, the Bayesian updating method is used to optimize the adjusted causal association matrix Calculation.

[0013] Optionally, the S3 specifically includes: S31, cross-modal feature extraction unit, receiving disease causal graph , based on structured multimodal datasets, construct a standardized feature matrix ; S32. Use self-supervised cross-modal representation learning method to build a cross-modal feature mapping model , and standardize the feature matrix Perform feature encoding and extract potential shared feature representation : ; in, is the potential shared feature representation, is a cross-modal feature mapping model, It is a cause-effect diagram of the disease; S33, feature alignment unit, based on potential shared feature representation , calculate the feature similarity between different modal data, and use the cross-modal contrast learning loss function for optimization, so that the feature distribution distance of different modal features of the same patient in the potential shared feature representation is minimized, and the feature distribution distance of different patients is maximized, and the optimized shared feature representation is generated ; S34, health portrait generation unit, based on optimized shared feature representation , using weighted feature fusion strategy to generate patient health portraits : ; in, is the feature fusion weight, is the fusion bias term, Create a health portrait for the patient.

[0014] Optionally, the S4 specifically includes: S41, disease status modeling unit, based on patient health profile Extract individual pathological feature vectors , construct the disease feature space, combine the temporal causal reasoning method, calculate the causal association weights between different pathological features, and quantify the evolution of the disease in continuous time steps; S42, intervention plan generation unit, based on the disease feature space, constructs a set of intervention measures ,in For intervention measures, For the type of intervention, define the intervention impact function Calculate the effect of different interventions on individual pathological feature vectors The impact value of the intervention measures was constructed : ; in, To measure interventions Individual pathological feature vector Contribution of S43, reinforcement learning optimization unit, based on the intervention evaluation matrix , using reinforcement learning methods to optimize intervention measures, calculate the long-term return value of intervention measures, and select the optimal disease intervention path through decision optimization functions : ; in, Represented in the individual pathological feature vector Intervention measures The long-term return value obtained, the time window for intervention evaluation; S44, intervention path update unit, based on individual pathological feature vector And the long-term return value of the intervention measures, dynamically adjust the intervention measures to form an intervention plan.

[0015] Optionally, the S6 specifically includes: S61, rehabilitation status modeling unit, receiving rehabilitation path decision data, and constructing a rehabilitation status set ,in Indicates that the patient is at time The state of recovery at all times; S62, causal dynamic programming unit, based on the recovery state set and intervention packages , using causal dynamic programming methods to simulate the impact of different rehabilitation paths on long-term recovery effects, and define the optimal value function of the rehabilitation path , and calculate the optimal value of the rehabilitation path for each rehabilitation state : ; in, Indicates recovery Implement intervention measures The immediate recovery benefits obtained, is the discount factor, which measures the balance between current rehabilitation benefits and long-term rehabilitation impacts. is the patient's recovery status at the next moment after the intervention, Indicates recovery status The optimal value of the rehabilitation path; S63, a rehabilitation path optimization unit, calculating an optimal rehabilitation path based on the optimal value of the rehabilitation path for each rehabilitation state; S64, personalized adjustment unit, uses the adaptive update mechanism to adjust the optimal rehabilitation path, evaluates the patient's recovery progress through dynamic causal reasoning, optimizes the intervention plan, and adapts the intervention plan to individual recovery needs.

[0016] Optionally, a method for information management of stroke patients in the neurology department, the method comprising the following steps: S71. Collect imaging data, physiological signals, genetic information and electronic medical record data of stroke patients in the Department of Neurology, convert and normalize the data format, remove redundant and abnormal data, and construct a structured multimodal data set; S72. Based on structured multimodal data sets, a structural causal learning method is used to construct a causal model of disease evolution, identify variables and their causal relationships, generate a disease causal graph, eliminate confounding factors, and optimize the causal relationship matrix; S73. Based on the disease causal graph, a self-supervised cross-modal representation learning method is used to extract shared features of structured multimodal datasets, perform cross-modal feature alignment and multimodal mapping, and generate a patient health profile. S74. Based on the patient's health profile and disease causal graph, combined with the temporal causal reasoning method, predict the impact of different intervention measures on the development of the disease, and use reinforcement learning methods to optimize the disease intervention measures and form an intervention plan; S75. Perform cross-modal consistency check on the disease causal diagram, patient health profile and intervention plan, use adaptive weighting strategy to adjust data contribution, and generate rehabilitation path decision data; S76. Based on the rehabilitation pathway decision data, combined with the causal dynamic programming method, simulate the long-term impact of different intervention plans, calculate the optimal rehabilitation pathway, and dynamically adjust the intervention plan during the implementation of the rehabilitation pathway.

[0017] The beneficial effects of the present invention are: (1) The present invention provides a deep understanding and dynamic adaptation of the progression of stroke patients' conditions by combining causal reasoning, cross-modal learning and reinforcement learning optimization, enabling the system to build an accurate causal model of the condition based on multimodal data, identify the causal relationship between pathological features, and eliminate confounding factors, thereby improving the accuracy of condition assessment. By using cross-modal representation learning methods, it can efficiently align and fuse features of multimodal data such as images, physiological signals, genetic data and electronic medical records to form an accurate individualized health portrait to support personalized intervention decisions.

[0018] (2) The present invention uses temporal causal reasoning and reinforcement learning optimization, combined with individual health portraits and disease causal graphs, to evaluate the impact of different intervention measures on the development of the disease in real time and optimize the personalized intervention path. This not only improves the accuracy of stroke intervention strategies, but also can adaptively adjust intervention plans during the dynamic evolution of the patient's condition, ensure the selection of the optimal rehabilitation path, and improve the patient's long-term rehabilitation effect.

[0019] (3) The present invention simulates the long-term effects of different rehabilitation intervention programs through a causal dynamic programming method, constructs an optimized rehabilitation path decision model, and dynamically updates the intervention strategy during the rehabilitation execution process. Reinforcement learning is used to optimize the rehabilitation path, so that the system can make intelligent decisions based on the patient's real-time condition changes, ensure the continuous optimization of the rehabilitation program, and improve the rehabilitation quality and management efficiency of stroke patients. This method reduces the limitations of traditional rehabilitation paths that rely on experience-based decision-making and realizes a personalized, intelligent, and dynamically optimized stroke management program. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] The accompanying drawings are used to provide a further understanding of the present invention and constitute a part of the specification. Together with the embodiments of the present invention, they are used to explain the present invention and do not constitute a limitation of the present invention. In the accompanying drawings: Figure 1 This is the system architecture diagram of the neurology stroke patient information management system proposed by the present invention; Figure 2 This is a flow chart of the information management method for neurological stroke patients proposed by the present invention. DETAILED DESCRIPTION

[0021] The present invention will now be described in further detail with reference to the accompanying drawings. These drawings are simplified schematic diagrams, which only illustrate the basic structure of the present invention in a schematic manner, and therefore only show the components related to the present invention.

[0022] refer to Figure 1-2 , the information management system and method for stroke patients in neurology department, including the following steps: S1, data acquisition module, collects imaging data, physiological signals, genetic information and electronic medical record data of neurological stroke patients, and performs standardized processing and storage, removes redundant information, and constructs a structured multimodal data set; This implementation method collects imaging data, physiological signals, genetic information and electronic medical record data from neurological stroke patients, where imaging data include MRI, CT and PET-CT, physiological signals include blood pressure, electrocardiogram, electroencephalogram, blood oxygen saturation, etc., genetic information includes APOE, CYP2C19, MTHFR and other genes, and electronic medical record data integrates the patient's basic medical history, medication records, laboratory tests, neurological function scores and rehabilitation training conditions, etc., and uses data standardization processing technology to remove redundant information to ensure the consistency of data formats of different data sources, and constructs a structured multimodal data set to support subsequent causal reasoning analysis and personalized health portrait generation. This method can effectively integrate heterogeneous data, improve data quality and integrity, provide a reliable data basis for condition assessment and precise intervention of stroke patients, and achieve more accurate condition monitoring and dynamic management.

[0023] S2, causal reasoning analysis module, based on structured multimodal data sets, uses structural causal learning methods to analyze pathological characteristics, eliminate confounding factors, build a causal model of disease evolution, and generate a disease causal graph; S3, cross-modal learning module, based on the disease causal graph, uses a self-supervised cross-modal representation learning method to extract shared features of structured multimodal datasets, perform feature alignment and multimodal mapping, and generate a patient health profile; S4, disease evolution analysis module, based on the patient's health profile and disease causal diagram, combines temporal causal reasoning to evaluate the impact of different intervention measures on the development of the disease and form an intervention plan; S5, data fusion module, performs cross-modal consistency check on the disease causal diagram, patient health profile and intervention plan, uses adaptive weighting strategy to adjust data contribution, and generates rehabilitation path decision data; This implementation adopts a cross-modal consistency verification mechanism to ensure that the disease causal diagram, health portrait and intervention plan remain consistent during the multimodal data fusion process. The distribution similarity between different modal data is calculated based on the statistical consistency test method, and the maximum mean difference is used to measure the cross-modal distribution deviation of imaging data, physiological signals, genetic information and electronic medical record data. The cross-modal feature alignment is optimized through an adversarial training strategy, and a feature mapping network is constructed using a cross-modal comparative learning method. The feature mutual information between different data sources is calculated to ensure that the representation of different modal data in the shared space is consistent. An adaptive weighted fusion strategy is used to optimize the weight distribution of each modal data, reduce the impact of abnormal data on the overall model, and improve the reliability of disease assessment and rehabilitation path decision-making, thereby improving the accuracy and effectiveness of personalized rehabilitation plans.

[0024] S6, rehabilitation optimization decision module, based on rehabilitation pathway decision data, uses causal dynamic programming methods to simulate the long-term impact of different intervention plans, and dynamically adjusts the intervention plan during the execution of the rehabilitation pathway.

[0025] In this implementation, S2 specifically includes: S21, a data parsing unit, receiving a structured multimodal data set, including imaging data, physiological signals, genetic information, and electronic medical record data, performing modality separation on the structured multimodal data set, extracting time series features and static features, and generating a standardized pathology data matrix; S22, causal relationship modeling unit, based on the standardized pathological data matrix, uses structural causal learning methods to build a causal model of disease evolution, identify the causal relationship between variables, and generate a causal association matrix ,in Representation variables right The causal influence weight of the variable and variables They represent disease-related variables respectively; S23, confounding factor elimination unit, based on the causal association matrix, uses the causal intervention inference method to calculate the direct causal effects between different variables to eliminate confounding factors and generate an adjusted causal association matrix: ; in, In the causal relationship matrix Under the constraints, eliminate confounding factors After the influence, the variable About the disease status The causal effect of Under the combined effects of variables, confounding factors and causal association matrix, the disease state The probability of In the causal relationship matrix Under the constraints, confounding factors The prior probability distribution of ; This formula is used to calculate the change in the disease state under specific intervention measures and eliminate the influence of confounding factors on causal relationships. Its principle is based on the intervention calculation method in causal reasoning. Through summation operations, the influence of confounding factors is marginalized so that the final calculation results can truly reflect the direct impact of intervention measures on the disease state. Specifically, the formula first calculates the probability of the disease state under the joint action of intervention variables and confounding factors, and then combines the prior probability of confounding factors to finally obtain the probability distribution of the disease state after eliminating the influence of confounding factors, thereby optimizing the accuracy of disease prediction and improving the scientific nature of intervention decisions.

[0026] S24, disease evolution modeling unit, based on the adjusted causal association matrix , combined with time series analysis methods to calculate the disease status , and establish a temporal causal relationship network to evaluate the disease development path; S25, causal graph generation unit, constructs a disease causal graph based on the disease evolution causal model ,in is a set of variables, The causal relationship is the time-dependent causal relationship of the temporal causal relationship network, and the disease causal graph is passed to the cross-modal learning module.

[0027] This implementation effectively solves the problems in the prior art of strong data heterogeneity, inaccurate disease analysis, and failure to eliminate the influence of confounding factors, improves the accuracy of disease modeling, and provides a solid data foundation and causal reasoning support for personalized intervention decisions.

[0028] In this implementation manner, the S23 specifically includes: S231, confounding factor identification unit, based on causal association matrix , identify confounding factors , using the conditional independence test method to evaluate the variables and disease status The independence between them is calculated by calculating the conditional independence probability Identify confounding factors Whether to change the variable and disease status The causal relationship between them has an impact; S232, causal effect calculation unit, based on the causal association matrix and confounding factors , using causal intervention inference methods to calculate variables About the disease status Direct causal effect, eliminating confounding factors The influence of the factors on the causal effect is generated after eliminating the confounding factors. S233, an adjustment matrix generation unit, generating an adjusted causal association matrix based on the causal effect calculation results after eliminating confounding factors ,in Eliminate confounding factors Finally, the Bayesian updating method is used to optimize the adjusted causal association matrix Calculation.

[0029] This implementation method identifies confounding factors based on the causal association matrix, and evaluates the independence between variables and disease status through the conditional independence test method, thereby determining the impact of confounding factors on causal relationships, using the causal intervention reasoning method to calculate the direct causal effect of variables on disease status, eliminating the interference of confounding factors on causal reasoning, ensuring the accuracy of causal reasoning, and using the Bayesian update method to optimize the causal association matrix to generate an adjusted causal association matrix to improve the stability and reliability of causal reasoning. This implementation method effectively improves the accuracy of disease evolution analysis, ensures the scientific nature of intervention decisions, avoids the causal reasoning bias caused by the influence of confounding factors in traditional statistical methods, and provides a more reliable theoretical basis for personalized disease assessment and precise intervention of stroke patients.

[0030] In this implementation, S3 specifically includes: S31, cross-modal feature extraction unit, receiving disease causal graph , based on structured multimodal datasets, construct a standardized feature matrix ; S32. Use self-supervised cross-modal representation learning method to build a cross-modal feature mapping model , and standardize the feature matrix Perform feature encoding and extract potential shared feature representation : ; in, is the potential shared feature representation, is a cross-modal feature mapping model, It is a cause-effect diagram of the disease; S33, feature alignment unit, based on potential shared feature representation , calculate the feature similarity between different modal data, and use the cross-modal contrast learning loss function for optimization, so that the feature distribution distance of different modal features of the same patient in the potential shared feature representation is minimized, and the feature distribution distance of different patients is maximized, and the optimized shared feature representation is generated ; This implementation is based on potential shared feature representation, calculates the feature similarity between different modal data (such as images, physiological signals, genetic data, and electronic medical records), and uses cross-modal contrast learning loss function for optimization, so that the distribution distance of multimodal features of the same patient in the shared feature space is minimized, while the distribution distance of features of different patients is maximized, thereby improving the accuracy of feature alignment. This method ensures that cross-modal data can be effectively integrated in a unified feature space, reduces information loss and feature deviation, improves the accuracy of individual health portraits, and provides more accurate data support for subsequent disease assessment, intervention optimization, and rehabilitation path decision-making, thereby enhancing the system's intelligent and personalized management capabilities for stroke patients.

[0031] S34, health portrait generation unit, based on optimized shared feature representation , using weighted feature fusion strategy to generate patient health portraits : ; in, is the feature fusion weight, is the fusion bias term, Create a health portrait for the patient.

[0032] This implementation method realizes the construction of personalized health portraits for stroke patients through cross-modal feature extraction, feature alignment and weighted fusion strategies, constructs a standardized feature matrix based on the disease causal graph and multimodal data, and uses a self-supervised cross-modal representation learning method to extract shared feature representations, thereby enhancing the information interoperability between different modal data. The cross-modal contrast learning loss function is used to optimize feature alignment so that the feature distribution of multimodal data of the same patient in the shared feature space is as close as possible, while the data features of different patients remain distinguished, thereby improving the accuracy of patient representation. Through the weighted feature fusion strategy, combined with the disease causal graph, personalized key pathological features are extracted to generate accurate health portraits. This implementation method can effectively integrate and align heterogeneous medical data, ensure the integrity and consistency of health portraits, and improve the accuracy of disease management, thereby optimizing personalized stroke intervention strategies and rehabilitation pathways.

[0033] In this implementation, S4 specifically includes: S41, disease status modeling unit, based on patient health profile Extract individual pathological feature vectors , construct the disease feature space, combine the temporal causal reasoning method, calculate the causal association weights between different pathological features, and quantify the evolution of the disease in continuous time steps; This implementation method extracts individual pathological feature vectors based on the patient's health portrait, constructs a disease feature space, and combines the temporal causal reasoning method to calculate the causal association weights between different pathological features, quantify the dynamic impact between disease variables, and by constructing the disease feature space, this method can accurately depict the changing trend of the disease state at different time steps, thereby identifying the evolution pattern of key pathological features and predicting the disease development trajectory. This method improves the accuracy of disease assessment, so that disease modeling is not limited to static data analysis, but can dynamically adjust the correlation of disease states, providing a more scientific reference for personalized intervention and rehabilitation path optimization.

[0034] S42, intervention plan generation unit, based on the disease feature space, constructs a set of intervention measures ,in For intervention measures, For the type of intervention, define the intervention impact function Calculate the effect of different interventions on individual pathological feature vectors The impact value of the intervention measures was constructed : ; in, To measure interventions Individual pathological feature vector Contribution of S43, reinforcement learning optimization unit, based on the intervention evaluation matrix , using reinforcement learning methods to optimize intervention measures, calculate the long-term return value of intervention measures, and select the optimal disease intervention path through decision optimization functions : ; in, Represented in the individual pathological feature vector Intervention measures The long-term return value obtained, the time window for intervention evaluation; This formula is used to calculate the optimal choice of different rehabilitation interventions for the patient's long-term rehabilitation path. The intervention strategy is optimized through reinforcement learning methods, so that the system can dynamically adjust the rehabilitation path at multiple time steps. Its principle is based on the idea of ​​value iteration, which calculates the cumulative rehabilitation returns that can be obtained by performing different interventions under a specific rehabilitation state, and introduces a discount coefficient to balance short-term and long-term rehabilitation benefits, thereby ensuring that the system gives priority to the path that can bring the greatest long-term rehabilitation effect. This method can effectively improve the adaptability of rehabilitation strategies, optimize and adjust the rehabilitation path as the patient's condition changes dynamically, thereby improving personalized rehabilitation effects.

[0035] S44, intervention path update unit, based on individual pathological feature vector And the long-term return value of the intervention measures, dynamically adjust the intervention measures to form an intervention plan.

[0036] This implementation method constructs the disease feature space based on the patient's health portrait, and combines the temporal causal reasoning method to quantify the evolution of the disease state, so as to accurately predict the development trend of the disease. The system evaluates the impact of different intervention measures on individual pathological characteristics through the intervention measure impact function, and constructs an intervention measure evaluation matrix to provide a quantitative basis for personalized intervention plans. It uses reinforcement learning methods to optimize intervention decisions, calculates the long-term reward values ​​of different intervention paths, and selects the optimal intervention path through the decision optimization function to ensure that the intervention measures can effectively promote patient rehabilitation. The system dynamically adjusts the intervention path based on the patient's real-time pathological feature vector and the long-term reward value of the intervention measures to form an individualized rehabilitation plan. This implementation method not only improves the accuracy and adaptability of stroke patient management by accurately modeling disease changes, optimizing the decision-making process of intervention measures, and combining reinforcement learning to achieve personalized dynamic adjustment, but also optimizes the rehabilitation path, making the patient's rehabilitation process more efficient and scientific.

[0037] In this implementation manner, S6 specifically includes: S61, rehabilitation status modeling unit, receiving rehabilitation path decision data, and constructing a rehabilitation status set ,in Indicates that the patient is at time The state of recovery at all times; S62, causal dynamic programming unit, based on the recovery state set and intervention packages , using causal dynamic programming methods to simulate the impact of different rehabilitation paths on long-term recovery effects, and define the optimal value function of the rehabilitation path , and calculate the optimal value of the rehabilitation path for each rehabilitation state : ; in, Indicates recovery Implement intervention measures The immediate recovery benefits obtained, is the discount factor, which measures the balance between current rehabilitation benefits and long-term rehabilitation impacts. is the patient's recovery status at the next moment after the intervention, Indicates recovery status The optimal value of the rehabilitation path; This formula is used to calculate the optimal value of the rehabilitation path and measure the impact of different rehabilitation interventions on the long-term recovery process. Its principle is based on causal dynamic programming. It accumulates immediate rehabilitation benefits and considers the long-term benefits of future rehabilitation status to balance short-term intervention effects and long-term rehabilitation goals. The discount coefficient is used to adjust the weights of current benefits and future benefits to ensure that the optimization process pays attention to current rehabilitation progress and reasonably plans long-term rehabilitation paths. This calculation method enables the rehabilitation plan to dynamically adapt to the patient's condition evolution trend and optimize the selection of intervention measures, thereby improving the accuracy and long-term effectiveness of personalized rehabilitation paths.

[0038] S63, a rehabilitation path optimization unit, calculating an optimal rehabilitation path based on the optimal value of the rehabilitation path for each rehabilitation state; S64, personalized adjustment unit, uses the adaptive update mechanism to adjust the optimal rehabilitation path, evaluates the patient's recovery progress through dynamic causal reasoning, optimizes the intervention plan, and adapts the intervention plan to individual recovery needs.

[0039] This implementation method quantifies the key state changes in the patient's rehabilitation process by constructing a rehabilitation state set and a state transition matrix to ensure the scientific nature and continuity of rehabilitation path optimization. Based on the causal reinforcement decision model, the impact of different rehabilitation intervention paths on long-term recovery effects is simulated, and the optimal value of each rehabilitation state is calculated, so that the system can accurately evaluate the long-term effects of different intervention measures. By optimizing the rehabilitation strategy, calculating the optimal rehabilitation path, and combining the adaptive update mechanism of the personalized adjustment unit, dynamic causal reasoning is used to evaluate the patient's recovery progress in real time, ensuring that the intervention strategy can be dynamically optimized as the patient's condition changes. The present invention improves the accuracy of rehabilitation path decisions, makes the rehabilitation plan for stroke patients more personalized, intelligent and dynamic, optimizes the rehabilitation effect, and at the same time reduces ineffective or inefficient medical interventions and improves the utilization of medical resources.

[0040] In this embodiment, the informatization management method for neurological stroke patients comprises the following steps: S71. Collect imaging data, physiological signals, genetic information and electronic medical record data of stroke patients in the Department of Neurology, convert and normalize the data format, remove redundant and abnormal data, and construct a structured multimodal data set; S72. Based on structured multimodal data sets, a structural causal learning method is used to construct a causal model of disease evolution, identify variables and their causal relationships, generate a disease causal graph, eliminate confounding factors, and optimize the causal relationship matrix; S73. Based on the disease causal graph, a self-supervised cross-modal representation learning method is used to extract shared features of structured multimodal datasets, perform cross-modal feature alignment and multimodal mapping, and generate a patient health profile. S74. Based on the patient's health profile and disease causal graph, combined with the temporal causal reasoning method, predict the impact of different intervention measures on the development of the disease, and use reinforcement learning methods to optimize the disease intervention measures and form an intervention plan; S75. Perform cross-modal consistency check on the disease causal diagram, patient health profile and intervention plan, use adaptive weighting strategy to adjust data contribution, and generate rehabilitation path decision data; S76. Based on the rehabilitation pathway decision data, combined with the causal dynamic programming method, simulate the long-term impact of different intervention plans, calculate the optimal rehabilitation pathway, and dynamically adjust the intervention plan during the implementation of the rehabilitation pathway.

[0041] Embodiment 1: In order to verify the feasibility of the present invention, the present invention is applied to the stroke patient management system of the Department of Neurology of a tertiary hospital. The hospital admitted a large number of patients with acute ischemic stroke and cerebral hemorrhage, with complicated conditions and long rehabilitation cycles. The traditional management method mainly relies on the experience of doctors and single data analysis, lacks personalized and precise intervention, resulting in large differences in patient rehabilitation effects. Some patients fail to adjust the rehabilitation plan in time, resulting in limited functional recovery or increased risk of recurrence. For this reason, the hospital decided to introduce the information management method for stroke patients in the Department of Neurology of the present invention to improve the precision and intelligence level of patient management and achieve personalized intervention and dynamic rehabilitation optimization.

[0042] In this system, the hospital establishes individualized health records for all stroke patients, and uses multimodal data acquisition equipment to obtain patients' imaging data, physiological signals, genetic information, and electronic medical record data during hospitalization and rehabilitation. All data are updated in real time through the hospital information management system and sensor equipment, and uploaded to the database of the stroke management system. After the data is standardized, it enters the causal reasoning analysis module, which identifies the causal relationship between pathological feature variables based on the causal learning method, generates a disease causal diagram, and automatically eliminates confounding factors to improve the accuracy of disease analysis.

[0043] In the disease assessment phase, the system extracts shared features of multimodal data based on cross-modal learning methods to form a personalized health portrait, and combines temporal causal reasoning methods to predict the disease evolution trend. For different patients, the system can dynamically evaluate their post-stroke neurological function recovery path and optimize personalized rehabilitation intervention plans in combination with reinforcement learning algorithms. For example, the system compares the long-term effects of different interventions, recommends the optimal rehabilitation path, and makes adjustments based on the patient's real-time data during the rehabilitation process, so that the rehabilitation plan always meets the patient's dynamic recovery needs.

[0044] In order to verify the actual effect of the present invention, the hospital selected 400 stroke patients for comparative experiments, of which 200 patients were treated with traditional stroke management methods and 200 patients were treated with the intelligent management method of the present invention. The experimental period was 6 months, and the performance of the two management methods in terms of rehabilitation progress, functional recovery, and medical intervention optimization were compared. The experimental data are shown in the following table: Table 1 Comparison of the effects of traditional stroke management methods and the method of the present invention

[0045] It can be seen from the experimental data that the intelligent stroke management method of the present invention significantly improves the patient's rehabilitation effect. The Barthel index is improved by 45.8% compared with the traditional method, indicating that the patient's daily living ability recovers faster. Because the system can automatically optimize the rehabilitation path, the hospital stay is reduced by 4.6 days, which improves the hospital's bed turnover rate. The dynamic adjustment of personalized rehabilitation plans has increased the matching degree of the intervention plan from 74.6% to 91.8%, and the frequency of rehabilitation plan adjustment has increased by 3.1 times, making the patient's rehabilitation strategy more targeted and adaptable. In addition, the system can respond to changes in the patient's condition in real time, shorten the intervention response time, reduce the stroke recurrence rate, and improve the patient's rehabilitation satisfaction.

[0046] In the specific application process, the system provides precise intervention according to the individual needs of different patients. For example, a 65-year-old male patient suffered from hemiplegia after stroke. The initial rehabilitation plan recommended 45 minutes of lower limb rehabilitation training every day, but the system found through time series analysis that the patient's muscle strength recovered faster, so the plan was adjusted in the third week of rehabilitation to increase the walking training time and combine it with neurostimulation therapy. Another 58-year-old female patient suffered from swallowing dysfunction after stroke. The system found through multimodal analysis combined with causal reasoning that her recovery was slow, automatically adjusted the nutritional intervention strategy, and recommended additional swallowing rehabilitation training, so that her swallowing function was basically restored after 8 weeks.

[0047] In summary, the present invention achieves accurate management of stroke patients and improves rehabilitation efficiency and medical resource utilization through real-time data collection, causal reasoning analysis, cross-modal learning, personalized health portrait construction, reinforcement learning optimization and dynamic adjustment of rehabilitation paths. The system can dynamically adapt to changes in the patient's condition and provide stroke patients with personalized, intelligent and optimized rehabilitation management solutions, thereby improving the patient's rehabilitation quality and reducing the stroke recurrence rate, providing medical institutions with an efficient and scientific stroke management model.

[0048] The above description is only a preferred specific implementation manner of the present invention, but the protection scope of the present invention is not limited thereto. Any technician familiar with the technical field can make equivalent replacements or changes according to the technical scheme and inventive concept of the present invention within the technical scope disclosed by the present invention, which should be covered by the protection scope of the present invention.

Claims

1. The information management system for stroke patients in the Department of Neurology is characterized by: The steps include: S1, data acquisition module, collects imaging data, physiological signals, genetic information and electronic medical record data of neurological stroke patients, and performs standardized processing and storage, removes redundant information, and constructs a structured multimodal data set; S2, causal reasoning analysis module, based on structured multimodal data sets, uses structural causal learning methods to analyze pathological characteristics, eliminate confounding factors, build a causal model of disease evolution, and generate a disease causal graph; S3, cross-modal learning module, based on the disease causal graph, uses a self-supervised cross-modal representation learning method to extract shared features of structured multimodal datasets, perform feature alignment and multimodal mapping, and generate a patient health profile; S4, disease evolution analysis module, based on the patient's health profile and disease causal diagram, combines temporal causal reasoning to evaluate the impact of different intervention measures on the development of the disease and form an intervention plan; S5, data fusion module, performs cross-modal consistency check on the disease causal diagram, patient health profile and intervention plan, uses adaptive weighting strategy to adjust data contribution, and generates rehabilitation path decision data; S6, rehabilitation optimization decision module, based on rehabilitation pathway decision data, uses causal dynamic programming methods to simulate the long-term impact of different intervention plans, and dynamically adjusts the intervention plan during the execution of the rehabilitation pathway.

2. The neurological stroke patient information management system according to claim 1, characterized in that: The S2 specifically includes: S21, a data parsing unit, receiving a structured multimodal data set, including imaging data, physiological signals, genetic information, and electronic medical record data, performing modality separation on the structured multimodal data set, extracting time series features and static features, and generating a standardized pathology data matrix; S22, causal relationship modeling unit, based on the standardized pathological data matrix, uses structural causal learning methods to build a causal model of disease evolution, identify the causal relationship between variables, and generate a causal association matrix ,in Representation variables right The causal influence weight of the variable and variables They represent disease-related variables respectively; S23, confounding factor elimination unit, based on the causal association matrix, uses the causal intervention inference method to calculate the direct causal effects between different variables to eliminate confounding factors and generate an adjusted causal association matrix: ； in, In the causal relationship matrix Under the constraints, eliminate confounding factors After the influence, the variable About the disease status The causal effect of Under the combined effects of variables, confounding factors and causal association matrix, the disease state The probability of In the causal relationship matrix Under the constraints, confounding factors The prior probability distribution of ; S24, disease evolution modeling unit, based on the adjusted causal association matrix , combined with time series analysis methods to calculate the disease status , and establish a temporal causal relationship network to evaluate the disease development path; S25, causal graph generation unit, constructs a disease causal graph based on the disease evolution causal model ,in is a set of variables, The causal relationship is the time-dependent causal relationship of the temporal causal relationship network, and the disease causal graph is passed to the cross-modal learning module.

3. The neurological stroke patient information management system according to claim 2, characterized in that: The S23 specifically includes: S231, confounding factor identification unit, based on causal association matrix , identify confounding factors , using the conditional independence test method to evaluate the variables and disease status The independence between them is calculated by calculating the conditional independence probability Identify confounding factors Whether to change the variable and disease status The causal relationship between them has an impact; S232, causal effect calculation unit, based on the causal association matrix and confounding factors , using causal intervention inference methods to calculate variables About the disease status Direct causal effect, eliminating confounding factors The influence of the factors on the causal effect is generated after eliminating the confounding factors. S233, an adjustment matrix generation unit, generating an adjusted causal association matrix based on the causal effect calculation results after eliminating confounding factors ,in Eliminate confounding factors Finally, the Bayesian updating method is used to optimize the adjusted causal association matrix Calculation.

4. The neurological stroke patient information management system according to claim 2, characterized in that: The S3 specifically includes: S31, cross-modal feature extraction unit, receiving disease causal graph , based on structured multimodal datasets, construct a standardized feature matrix ; S32. Use self-supervised cross-modal representation learning method to build a cross-modal feature mapping model , and standardize the feature matrix Perform feature encoding and extract potential shared feature representation : ； in, is the potential shared feature representation, is a cross-modal feature mapping model, It is a cause-effect diagram of the disease; S33, feature alignment unit, based on potential shared feature representation , calculate the feature similarity between different modal data, and use the cross-modal contrast learning loss function for optimization, so that the feature distribution distance of different modal features of the same patient in the potential shared feature representation is minimized, and the feature distribution distance of different patients is maximized, and the optimized shared feature representation is generated ; S34, health portrait generation unit, based on optimized shared feature representation , using weighted feature fusion strategy to generate patient health portraits : ； in, is the feature fusion weight, is the fusion bias term, Create a health portrait for the patient.

5. The neurological stroke patient information management system according to claim 1, characterized in that: The S4 specifically includes: S41, disease status modeling unit, based on patient health profile Extract individual pathological feature vectors , construct the disease feature space, combine the temporal causal reasoning method, calculate the causal association weights between different pathological features, and quantify the evolution of the disease in continuous time steps; S42, intervention plan generation unit, based on the disease feature space, constructs a set of intervention measures ,in For intervention measures, For the type of intervention, define the intervention impact function Calculate the effect of different interventions on individual pathological feature vectors The impact value of intervention measures was constructed : ； in, To measure interventions Individual pathological feature vector Contribution of S43, reinforcement learning optimization unit, based on the intervention evaluation matrix , using reinforcement learning methods to optimize intervention measures, calculate the long-term return value of intervention measures, and select the optimal disease intervention path through decision optimization functions : ； in, Represented in the individual pathological feature vector Intervention measures The long-term return value obtained, the time window for intervention evaluation; S44, intervention path update unit, based on individual pathological feature vector And the long-term return value of the intervention measures, dynamically adjust the intervention measures to form an intervention plan.

6. The neurological stroke patient information management system according to claim 5, characterized in that: The S6 specifically includes: S61, rehabilitation status modeling unit, receiving rehabilitation path decision data, and constructing a rehabilitation status set ,in Indicates that the patient is at time The state of recovery at all times; S62, causal dynamic programming unit, based on the recovery state set and intervention packages , using causal dynamic programming methods to simulate the impact of different rehabilitation paths on long-term recovery effects, and define the optimal value function of the rehabilitation path , and calculate the optimal value of the rehabilitation path for each rehabilitation state : ； in, In recovery state Implement intervention measures The immediate recovery benefits obtained, is the discount factor, which measures the balance between current rehabilitation benefits and long-term rehabilitation impacts. is the patient's recovery status at the next moment after the intervention, Indicates recovery status The optimal value of the rehabilitation path; S63, a rehabilitation path optimization unit, calculating an optimal rehabilitation path based on the optimal value of the rehabilitation path for each rehabilitation state; S64, personalized adjustment unit, uses the adaptive update mechanism to adjust the optimal rehabilitation path, evaluates the patient's recovery progress through dynamic causal reasoning, optimizes the intervention plan, and adapts the intervention plan to individual recovery needs.

7. A method for information management of neurological stroke patients, applied to the information management system for neurological stroke patients according to any one of claims 1 to 6, characterized in that: The steps include: S71. Collect imaging data, physiological signals, genetic information and electronic medical record data of stroke patients in the Department of Neurology, convert and normalize the data format, remove redundant and abnormal data, and construct a structured multimodal data set; S72. Based on structured multimodal data sets, a structural causal learning method is used to construct a causal model of disease evolution, identify variables and their causal relationships, generate a disease causal graph, eliminate confounding factors, and optimize the causal relationship matrix; S73. Based on the disease causal graph, a self-supervised cross-modal representation learning method is used to extract shared features of structured multimodal datasets, perform cross-modal feature alignment and multimodal mapping, and generate a patient health profile. S74. Based on the patient's health profile and disease causal graph, combined with the temporal causal reasoning method, predict the impact of different intervention measures on the development of the disease, and use reinforcement learning methods to optimize the disease intervention measures and form an intervention plan; S75. Perform cross-modal consistency check on the disease causal diagram, patient health profile and intervention plan, use adaptive weighting strategy to adjust data contribution, and generate rehabilitation path decision data; S76. Based on the rehabilitation pathway decision data, combined with the causal dynamic programming method, simulate the long-term impact of different intervention plans, calculate the optimal rehabilitation pathway, and dynamically adjust the intervention plan during the implementation of the rehabilitation pathway.

Images