Intelligent patient monitoring and predicting clinical data analysis system

Abstract

The invention discloses an intelligent patient monitoring and predicting clinical data analysis system, which belongs to the field of medical analysis and comprises a real-time monitoring module, a data acquisition module and a knowledge base building module. According to the method, the nursing quality of the patient can be improved, medical resources can be effectively saved, the processing capacity of complex medical data is enhanced, the scientificity and rationality of clinical decisions are improved, a personalized and timely health management scheme is provided for the patient, a hospital can be more prospective and initiative in disease management, and the disease management efficiency is improved. Therefore, the overall medical service level is improved, uncertainty and patient individualized demand differences existing in clinical data can be effectively handled, and scheme suggestions with higher diversity and flexibility are provided for medical staff during decision making, so that the whole system has real-time performance, and long-term effect optimization across time dimensions can be realized; the treatment effect and experience of the patient can be improved, and a scientific basis is provided for reasonable distribution of medical resources.

Description

Technical Field

[0001] The present invention relates to the field of medical analysis, and in particular to a clinical data analysis system for intelligent patient monitoring and prediction. Background Art

[0002] In recent years, with the aging of the global population and the rising incidence of chronic diseases, the medical system is facing increasing pressure. With the development of medical big data and the advancement of intelligent technology, traditional patient monitoring and health risk management have gradually shifted to the direction of intelligence and data-driven. In the modern medical system, massive clinical data, including patients' vital signs, medical records, laboratory test results, etc., contain rich health information. These data are not only of great value for the judgment of the current condition, but also can be used to predict the health risks and development trends of patients. However, due to the complexity, high dimensionality and strong heterogeneity of medical data, traditional data analysis methods are difficult to fully extract the feature information and make accurate predictions. At the same time, how to provide doctors with personalized treatment plans through effective algorithms for the personalized needs of different patients has become an urgent problem to be solved, and has become one of the hot spots of research and application in the medical field.

[0003] The existing clinical data analysis system cannot save medical resources while improving the quality of patient care, and has a low ability to process complex medical data, which reduces the scientificity and rationality of clinical decision-making. In addition, the existing clinical data analysis system cannot cope with the uncertainty in clinical data and the differences in individual patient needs, and cannot achieve long-term effect optimization across time dimensions, which is not conducive to improving patients' treatment effects and experience. For this reason, we propose a clinical data analysis system for intelligent patient monitoring and prediction. Summary of the invention

[0004] The purpose of the present invention is to solve the defects in the prior art and to propose an intelligent patient monitoring and prediction clinical data analysis system.

[0005] In order to achieve the above object, the present invention adopts the following technical solutions:

[0006] A clinical data analysis system for intelligent patient monitoring and prediction, including a real-time monitoring module, a data acquisition module, a knowledge base building module, an integration and extraction module, a health prediction module, a simulation verification module, a visualization display module, a clinical decision support module, an abnormal alarm module, an optimization application module and a decision optimization module;

[0007] The real-time monitoring module is used to continuously monitor the patient's physiological indicators and compare them with historical data;

[0008] The data acquisition module is used to automatically collect patient data from various medical devices and external electronic medical record systems;

[0009] The knowledge base building module is used to associate the patient's clinical data with medical knowledge to construct a clinical medicine map;

[0010] The integration and extraction module is used to clean and integrate patient data from different sources, analyze data features, and extract key clinical feature information;

[0011] The health prediction module is used to analyze the extracted features and predict the health risks and development trends of the patients;

[0012] The simulation verification module is used to evaluate the performance of the health prediction module in different scenarios;

[0013] The visualization display module is used to visualize the prediction results and analysis information;

[0014] The clinical decision support module is used to provide medical staff with targeted treatment plans based on the prediction results;

[0015] The abnormal alarm module is used to automatically issue an alarm when abnormalities occur in the monitoring data;

[0016] The optimization application module is used to optimize the parameter information of each module of the system;

[0017] The decision optimization module is used to recommend the most suitable personalized treatment plan based on the patient's real-time health data and dynamically adjust the treatment plan.

[0018] As a further solution of the present invention, the integrated extraction module extracts key clinical feature information in the following specific steps:

[0019] S1.1: Detect the patient data of each patient in the record, and compare the characteristic values ​​of each data record to determine whether there are exactly the same records, and delete the duplicate patient data, leaving only one copy, and detect the missing values ​​in each group of patient data. If the missing field of the patient data is less than the preset threshold, the field is directly deleted. If the missing numerical data in the patient data is less than the preset filling value, the missing data is filled by mean filling;

[0020] S1.2: Arrange the patient data in ascending order, calculate the first quartile Q1, the second quartile Q2, and the third quartile Q3 of the arranged patient data, and calculate the interquartile range IQR by the formula IQR=Q3-Q1, and then determine the outlier range [Q1-1.5×IQR, Q3+1.5×IQR] by the quartile rule. If there is patient data that exceeds the outlier range, mark the patient data as outlier data and remove it;

[0021] S1.3: Pass Calculate the Z score of the remaining patient data that cannot be sorted in ascending order, where represents the standardized score of the patient data, Represents the patient data value, represents the mean of the patient data, Represents the standard deviation of the patient's data. If the absolute Z score of the patient's data is greater than 3, the patient is considered abnormal and excluded;

[0022] S1.4: Check the same patient data in different data sources, unify their names and units, use timestamp alignment strategy, integrate records in different data sources by time, and unify the data format to obtain the integrated patient data set, and then standardize each group of data in the patient data set;

[0023] S1.5: The linear correlation between the characteristics of each group of patients and the target variables set by medical staff is calculated using the Pearson correlation coefficient, and the features whose correlation with the target is higher than the preset correlation threshold are screened. PCA is used to reduce the feature dimension, and then the information gain of each group of patient characteristics is calculated. The features whose information gain is higher than the preset gain threshold are screened, and the impact of each group of patient characteristics on the variance of the target variable is analyzed, and the features whose impact on the target variance is higher than the preset impact value are extracted.

[0024] As a further solution of the present invention, the health prediction module predicts the health risks and development trends of patients in the following specific steps:

[0025] S2.1: Arrange the extracted patient features of each group in chronological order, and normalize the patient features of each group to make the data range of each feature consistent. Then convert the processed patient features into a three-dimensional format, i.e., [number of samples, number of time steps, number of features], and build a bidirectional GRU network architecture consisting of an input layer, a Bi-GRU layer, and an output layer to construct the required prediction model;

[0026] S2.2: Divide the processed patient features into a training set and a validation set, and transmit the patient features in the training set to the prediction model. The prediction model input layer receives the preprocessed patient features and performs forward GRU unit calculation and backward GRU unit calculation on the patient features through the Bi-GRU layer.

[0027] S2.3: The output layer receives the output data of the Bi-GRU layer, and combines the hidden states of the forward GRU unit and the backward GRU unit to form the final output of the bidirectional GRU. The error between the predicted value and the true label is calculated through the mean square error loss function. The gradient of the loss function to the network parameters of each layer of the prediction model is calculated through the back propagation algorithm, and the parameters are updated through the Adam optimizer based on the calculated gradient.

[0028] S2.4: Input the patient characteristics in the validation set into the trained prediction model, calculate the model's loss value and accuracy performance indicators on the validation set, and evaluate its prediction effect on the patient's health risk and development trend. If the performance indicator does not reach the preset target value, retrain and verify the prediction model until the model performance indicator reaches the preset target value;

[0029] S2.5: Deploy the trained model in the clinical system, input the latest patient data into the prediction model for forward propagation calculation, input the forward propagation results into the output layer of the model, and the output layer performs classification calculation on the received data;

[0030] S2.6: If the target is binary classification, use the sigmoid activation function to output the risk probability. If the target is multi-classification or regression, use the softmax activation function or linear activation function to output the corresponding prediction value, and interpret the prediction results of each group to determine the risk score and make a judgment;

[0031] 2.7: Re-collect data at fixed intervals and update model predictions. When the prediction results show that the risk value is higher than the set safety threshold, the alarm module automatically alerts medical staff and records the corresponding patient data for further analysis.

[0032] As a further solution of the present invention, the specific calculation formula for the forward GRU unit calculation in S2.2 is as follows:

[0033]

[0034] In the formula, represents the update gate, which is used to control the update amount of the previous hidden state; Represents the Sigmoid activation function, which is used to map the input value to between 0 and 1; Represents the weight matrix of the update gate; Representative The hidden state of time steps; Represents the current time step input; Represents the bias term of the update gate; Represents the reset gate, which is used to control the amount of forgetting; Represents the weight matrix of the reset gate; Represents the bias term for the reset gate; represents the candidate hidden state; A weight matrix representing candidate hidden states; A bias term representing the candidate hidden state; Representative The hidden state of time steps;

[0035] The specific calculation formula of the mean square error loss function described in S2.3 is as follows:

[0036]

[0037] In the formula, represents the mean square error value; Represents the number of patient data; An index representing patient data; Representative The actual value of each patient's data; Representative The predicted value of the patient data.

[0038] As a further solution of the present invention, the simulation verification module evaluates the performance of the health prediction module in different scenarios in the following specific steps:

[0039] S3.1: Collect the current patient data status and the prediction results corresponding to the prediction model, and use them as the initial node. Then use the status after the subsequent prediction model simulation as the child node of the initial node, and calculate the UCB value of each group of child nodes generated;

[0040] S3.2: Use the upper confidence bound strategy to gradually select the child node with the highest UCB value until reaching a node that is not fully expanded, i.e., a leaf node, expand the selected node, add one or more new prediction scenarios to the current node state, and generate one or more child nodes, where the prediction scenario specifically refers to the performance of the prediction model under new patient data, different feature combinations, or different input data;

[0041] S3.3: Starting from the expanded new node, simulate the behavior of the prediction model, perform a full simulation starting from the current node, use random sampling or model prediction to obtain the results on the simulation path, record the performance of the model on the simulation path, and calculate the corresponding reward value for the model performance based on the simulation results;

[0042] S3.4: Backtrack from the termination node to the root node step by step, update the number of visits and cumulative rewards of each node, repeat the selection, expansion, simulation and backtracking steps until the preset iteration time is reached, and select the child node with the highest reward value from the root node step by step until the termination node is reached;

[0043] S3.5: Use the selected path as the optimal validation path for the model, analyze the performance of the model under the optimal path, evaluate the accuracy and stability of the model in different scenarios and its ability to predict patient health risks, identify the advantages and disadvantages of the model under various feature combinations, time periods or input data, and improve the prediction model based on the analysis results.

[0044] As a further solution of the present invention, the clinical decision support module provides doctors with targeted clinical suggestions in the following specific steps:

[0045] S4.1: Collect the different health states of patients during treatment and construct the corresponding state space , taking the treatment decisions or drug adjustment plans that medical staff can take as actions in the MDP to construct the action space ;

[0046] S4.2: Calculate the patient's current health status based on historical data After treatment plan Then transfer to the next health state Probability ,in Representative patient time The health status within Representative patient time The health status within ,and In the range [1, m];

[0047] S4.3: Give positive rewards if the treatment plan improves the patient's health; if it causes side effects or worsening, give negative rewards, and construct the corresponding reward function , to evaluate the improvement of the patient's health after taking each treatment plan, in the state space In , a random treatment plan is assigned to each state to initialize the strategy, which is recorded as ;

[0048] S4.4: Update the value function of each state based on the actions performed by the current policy through the Bellman equation , and use the recursive calculation method to update iteratively step by step , until the value function changes of all states are less than the preset threshold, then in each state, choose to make the value function The maximized action is used as the action of the new strategy and the original strategy is updated;

[0049] S4.5: Compare the strategies before and after the update. If the two are the same in all states, the strategy is considered to have converged. Otherwise, continue to perform the strategy evaluation and improvement steps using the new strategy until the strategy no longer changes in all states, and then provide the treatment plan in the optimal strategy to the medical staff.

[0050] As a further solution of the present invention, the specific calculation formula of the Bellman equation described in S4.4 is as follows:

[0051]

[0052] In the formula, Representation status The value under the current policy; Represents the next health status; Represents the patient's status Through strategic actions Transfer to state probability; Represents taking the current policy action Time return; represents the discount factor used to weigh current and future rewards; Representation status The value under the current policy.

[0053] As a further solution of the present invention, the specific steps of dynamically adjusting the treatment plan by the decision optimization module are as follows:

[0054] S5.1: The decision optimization module collects the health status of patients under the treatment plan in the optimal strategy in real time, initializes the pheromone and heuristic function of each treatment path in the treatment plan, and then randomly generates multiple groups of individuals, and uses the health status of the current patient as the initial node;

[0055] S5.2: According to the pheromone and heuristic function of each treatment path, the selection probability of all treatment paths is calculated in the current state, and the next action is selected by random selection according to each selection probability;

[0056] S5.3: After each group of individuals selects a path, they simulate the treatment results of the treatment path, record the effect of each path based on the simulated treatment results, and update the pheromones on the path after each group of individuals completes its path;

[0057] S5.4: Repeat the path selection and pheromone update. In each iteration, all individuals will reselect the path, update the pheromone, and adjust the priority of path selection according to the simulation results until the maximum number of iterations is reached or the pheromone concentration changes converge to the preset range;

[0058] S5.5: After completing all iterations, evaluate the pheromone concentrations of all treatment pathways, select the pathway with the highest pheromone concentration, record the corresponding treatment methods, sort out each treatment plan in the best pathway, and adjust the treatment plans generated by the clinical decision support module based on the generated groups of treatment plans. At the same time, feed back each group of plans to medical staff through the visualization module for review and selection.

[0059] Compared with the prior art, the present invention has the following beneficial effects:

[0060] 1. The present invention converts the extracted patient features into a three-dimensional format, and builds a bidirectional GRU network architecture including an input layer, a Bi-GRU layer and an output layer to construct the required prediction model, divides the processed patient features into a training set and a verification set, and transmits the patient features in the training set to the prediction model. The prediction model input layer receives the preprocessed patient features, and performs forward GRU unit calculation and backward GRU unit calculation on the patient features through the Bi-GRU layer. The output layer receives the output data of the Bi-GRU layer, and combines the hidden states of the forward GRU unit and the backward GRU unit to form the final output of the bidirectional GRU, calculates the error between the predicted value and the true value, and adjusts the model parameters through back propagation, then verifies the model, repeats training and verification until the model reaches the preset target value, collects the current state of the patient data and the prediction result corresponding to the prediction model, and uses it as the initial node, and then uses the state after the subsequent prediction model simulation as the child node of the initial node, and repeats the selection, expansion, and modeling. The simulation and backtracking steps are performed until the preset iteration time is reached, and the child nodes with the highest reward value are gradually selected from the root node downward until the termination node is reached. The selected path is used as the optimal verification path of the model, and the performance of the model under the optimal path is analyzed. The accuracy and stability of the model in different scenarios and its predictive ability for patient health risks are evaluated, and the advantages and disadvantages of the model under various feature combinations, time periods or input data are identified. The prediction model is improved based on the analysis results, and the trained model is deployed in the clinical system. The latest patient data is input into the prediction model for forward propagation calculation, and the forward propagation results are input into the output layer of the model. The output layer classifies and calculates the received data and outputs the risk probability or predicted value. It can improve the quality of patient care while effectively saving medical resources. It not only enhances the ability to process complex medical data, but also improves the scientificity and rationality of clinical decision-making, provides patients with personalized and timely health management plans, and enables hospitals to be more forward-looking and proactive in disease management, thereby improving the overall level of medical services.

[0061] 2. The present invention constructs a corresponding state space and action space by collecting different health states of patients during the treatment process and the treatment decisions or drug adjustment plans that medical staff can take. Then, based on historical data, the probability of the patient transferring from the current health state to the next health state after the treatment plan is calculated, and the health improvement of the patient after each treatment plan is evaluated. A random treatment plan is assigned to each state to initialize the strategy. The value function of each state is updated according to the action executed by the current strategy through the Bellman equation, and the recursive calculation method is used to gradually iterate and update until the value function change of all states is less than a preset threshold. Then, in each state, the action that maximizes the value function is selected as the action of the new strategy, and the original strategy is updated. The strategies before and after the update are compared. If the two are the same in all states, it is considered that the strategy has converged; otherwise, the new strategy is used to continue to execute the strategy evaluation and improvement steps until the strategy no longer changes in all states, and then the treatment plan in the optimal strategy is provided to the medical staff, and the health state of the patient under the treatment plan in the optimal strategy is collected in real time. Then, multiple groups of individuals are randomly generated, and the health state of the current patient is used as The initial node is the initial node. According to the pheromone and heuristic function of each treatment path, the selection probability of all treatment paths is calculated in the current state, and the next action is selected by random selection. After selecting the path, each group of individuals simulates the treatment results of the treatment path, records the effect of each path, updates the pheromone on the path passed, and repeats the path selection and pheromone update until the maximum number of iterations is reached or the pheromone concentration change converges to the preset range. After completing all iterations, the pheromone concentration of all treatment paths is evaluated, the path with the highest pheromone concentration is selected, the corresponding treatment means are recorded, and each treatment plan in the best path is sorted out. Based on the generated groups of treatment plans, the treatment plans generated by the clinical decision support module are adjusted. At the same time, each group of plans is fed back to medical staff through the visualization display module for viewing and selection. It can effectively deal with the uncertainty in clinical data and the differences in individual patient needs, and provide medical staff with more diverse and flexible plan suggestions when making decisions, so that the entire system is not only real-time, but also can achieve long-term effect optimization across time dimensions, which helps to improve the treatment effect and experience of patients and provides a scientific basis for the rational allocation of medical resources. BRIEF DESCRIPTION OF THE DRAWINGS

[0062] The accompanying drawings are used to provide further understanding of the present invention and constitute a part of the specification. They are used to explain the present invention together with the embodiments of the present invention and do not constitute a limitation of the present invention.

[0063] Figure 1 This is a system block diagram of a clinical data analysis system for intelligent patient monitoring and prediction proposed by the present invention. DETAILED DESCRIPTION

[0064] The technical solutions in the embodiments of the present invention will be described clearly and completely below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all the embodiments.

[0065] Example 1

[0066] Reference Figure 1 , a clinical data analysis system for intelligent patient monitoring and prediction, includes a real-time monitoring module, a data acquisition module, a knowledge base building module, an integration and extraction module, a health prediction module, a simulation verification module, a visualization display module, a clinical decision support module, an abnormal alarm module, an optimization application module and a decision optimization module.

[0067] The real-time monitoring module is used to continuously monitor the patient's physiological indicators and compare them with historical data; the data acquisition module is used to automatically collect patient data from various medical devices and external electronic medical record systems; the knowledge base building module is used to associate the patient's clinical data with medical knowledge and construct a clinical medicine map.

[0068] The integration and extraction module is used to clean and integrate patient data from different sources, analyze data features, and extract key clinical feature information.

[0069] Specifically, the patient data of each patient is detected, and the characteristic values ​​of each data record are compared to determine whether there are exactly the same records, and duplicate patient data are deleted, leaving only one copy. The missing values ​​in each group of patient data are detected. If the missing field of the patient data is less than the preset threshold, the field is directly deleted. If the missing numerical data in the patient data is less than the preset filling value, the missing data is filled by mean filling, and the patient data is arranged in ascending order. Then, the first quartile Q1, the second quartile Q2, and the third quartile Q3 of the arranged patient data are calculated, and the quartile range IQR is calculated by the formula IQR=Q3-Q1, and then the outlier range [Q1-1.5×IQR, Q3+1.5×IQR] is determined by the quartile rule. If there is patient data exceeding the outlier range, the patient data is marked as abnormal data and removed. Calculate the Z score of the remaining patient data that cannot be sorted in ascending order, where represents the standardized score of the patient data, Represents the patient data value, represents the mean of the patient data, Represents the standard deviation of the patient data. If the absolute Z score of the patient data is greater than 3, the patient is considered abnormal and eliminated. The same patient data in different data sources are checked, and their names and units are unified. The timestamp alignment strategy is used to integrate the records in different data sources by time and unify the data format to obtain the integrated patient data set. After that, each group of data in the patient data set is standardized. The linear correlation between each group of patient characteristics and the target variable set by the medical staff is calculated by the Pearson correlation coefficient, and the features with a correlation with the target higher than the preset correlation threshold are screened. PCA is used to reduce the feature dimension, and then the information gain of each group of patient characteristics is calculated. Features with information gain higher than the preset gain threshold are screened. The impact of each group of patient characteristics on the variance of the target variable is analyzed, and features with an impact on the target variance higher than the preset impact value are extracted.

[0070] The health prediction module is used to analyze the extracted features and predict the patient's health risks and development trends.

[0071] Specifically, the extracted patient features of each group are arranged in chronological order, and the patient features of each group are normalized to make the data range of each feature consistent. Then, the processed patient features are converted into a three-dimensional format, namely [number of samples, number of time steps, number of features]. A bidirectional GRU network architecture including an input layer, a Bi-GRU layer, and an output layer is constructed to construct the required prediction model. The processed patient features are divided into a training set and a validation set, and the patient features in the training set are transmitted to the prediction model. The input layer of the prediction model receives the preprocessed patient features, and performs forward GRU unit calculation and backward GRU unit calculation on the patient features through the Bi-GRU layer. The output layer receives the output data of the Bi-GRU layer, and combines the hidden states of the forward GRU unit and the backward GRU unit to form the final output of the bidirectional GRU. The error between the predicted value and the true label is calculated by the mean square error loss function. The gradient of the loss function on the network parameters of each layer of the prediction model is calculated by the back propagation algorithm, and the parameters are updated by the Adam optimizer based on the calculated gradient. The patient characteristics in the evidence set are input into the trained prediction model, and the loss value and accuracy performance indicators of the model on the validation set are calculated to evaluate its prediction effect on the patient's health risks and development trends. If the performance indicators do not reach the preset target values, the prediction model is retrained and verified until the model performance indicators reach the preset target values. The trained model is deployed in the clinical system, and the latest patient data is input into the prediction model for forward propagation calculation. The forward propagation results are input into the output layer of the model. The output layer performs classification calculation on the received data. If the target is binary classification, the sigmoid activation function is used to output the risk probability. If the target is multi-classification or regression, the softmax activation function or linear activation function is used to output the corresponding prediction value, and the prediction results of each group are interpreted to determine the risk score and make a judgment. The data is collected again at fixed intervals and the model prediction is updated. When the prediction results show that the risk value is higher than the set safety threshold, the alarm module automatically issues an alarm to the medical staff and records the corresponding patient data for further analysis.

[0072] In this embodiment, the specific calculation formula for the forward GRU unit is as follows:

[0073]

[0074] In the formula, represents the update gate, which is used to control the update amount of the previous hidden state; Represents the Sigmoid activation function, which is used to map the input value to between 0 and 1; Represents the weight matrix of the update gate; Representative The hidden state of time steps; Represents the current time step input; Represents the bias term of the update gate; Represents the reset gate, which is used to control the amount of forgetting; Represents the weight matrix of the reset gate; Represents the bias term for the reset gate; represents the candidate hidden state; A weight matrix representing candidate hidden states; A bias term representing the candidate hidden state; Representative The hidden state of time steps;

[0075] The specific calculation formula of the mean square error loss function is as follows:

[0076]

[0077] In the formula, represents the mean square error value; Represents the number of patient data; An index representing patient data; Representative The actual value of each patient's data; Representative The predicted value of the patient data.

[0078] The simulation verification module is used to evaluate the performance of the health prediction module in different scenarios.

[0079] Specifically, the state of the current patient data and the prediction results corresponding to the prediction model are collected and used as the initial node. Then, the state after the subsequent prediction model simulation is used as the child node of the initial node, and the UCB value of each group of child nodes generated is calculated. The upper confidence bound strategy is used to gradually select the child node with the highest UCB value until a node that is not fully expanded, that is, a leaf node, is reached. The selected node is expanded, and one or more new prediction scenarios are added to the current node state to generate one or more child nodes, where the prediction scenario is specifically the performance of the prediction model under new patient data, different feature combinations or different input data. Starting from the expanded new node, the behavior of the prediction model is simulated, and a complete simulation starting from the current node is performed. Use the random node. Machine sampling or model prediction is used to obtain the results on the simulation path, the performance of the model on the simulation path is recorded, and the corresponding reward value is calculated for the model performance based on the simulation results. The system traces back from the termination node to the root node step by step, updates the number of visits and cumulative rewards of each node, and repeats the selection, expansion, simulation and backtracking steps until the preset iteration time is reached. The child nodes with the highest reward value are selected from the root node step by step until the termination node is reached. The selected path is used as the best verification path for the model, and the performance of the model under the best path is analyzed. The accuracy and stability of the model in different scenarios and its ability to predict patient health risks are evaluated, the advantages and disadvantages of the model under various feature combinations, time periods or input data are identified, and the prediction model is improved based on the analysis results.

[0080] Example 2

[0081] Reference Figure 1 , a clinical data analysis system for intelligent patient monitoring and prediction, includes a real-time monitoring module, a data acquisition module, a knowledge base building module, an integration and extraction module, a health prediction module, a simulation verification module, a visualization display module, a clinical decision support module, an abnormal alarm module, an optimization application module and a decision optimization module.

[0082] The visualization display module is used to visualize the prediction results and analysis information; the clinical decision support module is used to provide targeted treatment plans for medical staff based on the prediction results.

[0083] Specifically, the different health states of patients during treatment are collected and the corresponding state space is constructed. , taking the treatment decisions or drug adjustment plans that medical staff can take as actions in the MDP to construct the action space , based on historical data, calculate the patient's current health status After treatment plan Then transfer to the next health state Probability ,in Representative patient time The health status within Representative patient time The health status within ,and In the range of [1, m], positive rewards are given for improving the health status of patients according to the treatment plan; negative rewards are given for causing side effects or deterioration, and the corresponding reward function is constructed. , to evaluate the improvement of the patient's health after taking each treatment plan, in the state space In , a random treatment plan is assigned to each state to initialize the strategy, which is recorded as , through the Bellman equation, according to the actions performed by the current strategy, update the value function of each state , and use the recursive calculation method to update iteratively step by step , until the value function changes of all states are less than the preset threshold, then in each state, choose to make the value function The action that maximizes is used as the action of the new strategy, and the original strategy is updated. The strategies before and after the update are compared. If the two are the same in all states, the strategy is considered to have converged; otherwise, the new strategy is used to continue the strategy evaluation and improvement steps until the strategy no longer changes in all states, and then the treatment plan in the optimal strategy is provided to medical staff.

[0084] In this embodiment, the specific calculation formula of the Bellman equation is as follows:

[0085]

[0086] In the formula, Representation status The value under the current policy; Represents the next health status; Represents the patient's status Through strategic actions Transfer to state probability; Represents taking the current policy action Time return; represents the discount factor used to weigh current and future rewards; Representation status The value under the current policy.

[0087] The abnormal alarm module is used to automatically issue an alarm when abnormalities occur in the monitoring data; the optimization application module is used to optimize the parameter information of each module of the system; the decision optimization module is used to recommend the most suitable personalized treatment plan based on the patient's real-time health data, and dynamically adjust the treatment plan.

[0088] Specifically, the decision optimization module collects the health status of patients under the treatment plan in the optimal strategy in real time, and initializes the pheromones and heuristic functions of each treatment path in the treatment plan. Then, it randomly generates multiple groups of individuals, and takes the health status of the current patient as the initial node. According to the pheromones and heuristic functions of each treatment path, the selection probability of all treatment paths is calculated under the current state, and the next action is selected by random selection according to each selection probability. After each group of individuals selects a path, the treatment results of the treatment path are simulated. According to the simulated treatment results, the effect of each path is recorded, and after each group of individuals completes its path, the path passed is updated. pheromone, repeat path selection and pheromone update. At the same time, in each round of iteration, all individuals will reselect paths, update pheromones, and adjust the priority of path selection according to the simulation results until the maximum number of iterations is reached or the pheromone concentration change converges to the preset range. After completing all iterations, the pheromone concentrations of all treatment paths are evaluated, the path with the highest pheromone concentration is selected, the corresponding treatment methods are recorded, and each treatment plan in the best path is sorted out. Based on the generated groups of treatment plans, the treatment plans generated by the clinical decision support module are adjusted, and each group of plans is fed back to the medical staff through the visualization display module for viewing and selection.

Claims

1. A clinical data analysis system for intelligent patient monitoring and prediction, characterized in that: It includes real-time monitoring module, data collection module, knowledge base building module, integration and extraction module, health prediction module, simulation verification module, visualization display module, clinical decision support module, abnormal alarm module, optimization application module and decision optimization module; The real-time monitoring module is used to continuously monitor the patient's physiological indicators and compare them with historical data; The data acquisition module is used to automatically collect patient data from various medical devices and external electronic medical record systems; The knowledge base building module is used to associate the patient's clinical data with medical knowledge to construct a clinical medicine map; The integration and extraction module is used to clean and integrate patient data from different sources, analyze data features, and extract key clinical feature information; The health prediction module is used to analyze the extracted features and predict the health risks and development trends of the patients; The simulation verification module is used to evaluate the performance of the health prediction module in different scenarios; The visualization display module is used to visualize the prediction results and analysis information; The clinical decision support module is used to provide medical staff with targeted treatment plans based on the prediction results; The abnormal alarm module is used to automatically issue an alarm when abnormalities occur in the monitoring data; The optimization application module is used to optimize the parameter information of each module of the system; The decision optimization module is used to recommend the most suitable personalized treatment plan based on the patient's real-time health data and dynamically adjust the treatment plan.

2. The clinical data analysis system for intelligent patient monitoring and prediction according to claim 1, characterized in that: The specific steps of extracting key clinical feature information by the integrated extraction module are as follows: S1.1: Detect the patient data of each patient in the record, and compare the characteristic values ​​of each data record to determine whether there are exactly the same records, and delete the duplicate patient data, leaving only one copy, and detect the missing values ​​in each group of patient data. If the missing field of the patient data is less than the preset threshold, the field is directly deleted. If the missing numerical data in the patient data is less than the preset filling value, the missing data is filled by mean filling; S1.2: Arrange the patient data in ascending order, calculate the first quartile Q1, the second quartile Q2, and the third quartile Q3 of the arranged patient data, and calculate the interquartile range IQR by the formula IQR=Q3-Q1, and then determine the outlier range [Q1-1.5×IQR, Q3+1.5×IQR] by the quartile rule. If there is patient data that exceeds the outlier range, mark the patient data as outlier data and remove it; S1.3: Pass Calculate the Z score of the remaining patient data that cannot be sorted in ascending order, where represents the standardized score of the patient data, Represents the patient data value, represents the mean of the patient data, Represents the standard deviation of the patient's data. If the absolute Z score of the patient's data is greater than 3, the patient is considered abnormal and excluded; S1.4: Check the same patient data in different data sources, unify their names and units, use timestamp alignment strategy, integrate records in different data sources by time, and unify the data format to obtain the integrated patient data set, and then standardize each group of data in the patient data set; S1.5: The linear correlation between the characteristics of each group of patients and the target variables set by medical staff is calculated using the Pearson correlation coefficient, and the features whose correlation with the target is higher than the preset correlation threshold are screened. PCA is used to reduce the feature dimension, and then the information gain of each group of patient characteristics is calculated. The features whose information gain is higher than the preset gain threshold are screened, and the impact of each group of patient characteristics on the variance of the target variable is analyzed, and the features whose impact on the target variance is higher than the preset impact value are extracted.

3. The clinical data analysis system for intelligent patient monitoring and prediction according to claim 2, characterized in that: The specific steps of the health prediction module for predicting the health risks and development trends of patients are as follows: S2.1: Arrange the extracted patient features of each group in chronological order, and normalize the patient features of each group to make the data range of each feature consistent. Then convert the processed patient features into a three-dimensional format, i.e., [number of samples, number of time steps, number of features], and build a bidirectional GRU network architecture consisting of an input layer, a Bi-GRU layer, and an output layer to construct the required prediction model; S2.2: Divide the processed patient features into a training set and a validation set, and transmit the patient features in the training set to the prediction model. The prediction model input layer receives the preprocessed patient features and performs forward GRU unit calculation and backward GRU unit calculation on the patient features through the Bi-GRU layer. S2.3: The output layer receives the output data of the Bi-GRU layer, and combines the hidden states of the forward GRU unit and the backward GRU unit to form the final output of the bidirectional GRU. The error between the predicted value and the true label is calculated through the mean square error loss function. The gradient of the loss function to the network parameters of each layer of the prediction model is calculated through the back propagation algorithm, and the parameters are updated through the Adam optimizer based on the calculated gradient. S2.4: Input the patient characteristics in the validation set into the trained prediction model, calculate the model's loss value and accuracy performance indicators on the validation set, and evaluate its prediction effect on the patient's health risk and development trend. If the performance indicator does not reach the preset target value, retrain and verify the prediction model until the model performance indicator reaches the preset target value; S2.5: Deploy the trained model in the clinical system, input the latest patient data into the prediction model for forward propagation calculation, input the forward propagation results into the output layer of the model, and the output layer performs classification calculation on the received data; S2.6: If the target is binary classification, use the sigmoid activation function to output the risk probability. If the target is multi-classification or regression, use the softmax activation function or linear activation function to output the corresponding prediction value, and interpret the prediction results of each group to determine the risk score and make a judgment; S2.7: Re-collect data at fixed intervals and update model predictions. When the prediction results show that the risk value is higher than the set safety threshold, the alarm module automatically sends an alarm to medical staff and records the corresponding patient data for further analysis.

4. The clinical data analysis system for intelligent patient monitoring and prediction according to claim 3, characterized in that: The specific calculation formula for the forward GRU unit calculation described in S2.2 is as follows: , In the formula, represents the update gate, which is used to control the update amount of the previous hidden state; Represents the Sigmoid activation function, which is used to map the input value to between 0 and 1; Represents the weight matrix of the update gate; Representative The hidden state of time steps; Represents the current time step input; Represents the bias term of the update gate; Represents the reset gate, which is used to control the amount of forgetting; Represents the weight matrix of the reset gate; Represents the bias term for the reset gate; represents the candidate hidden state; A weight matrix representing candidate hidden states; A bias term representing the candidate hidden state; Representative The hidden state of time steps; The specific calculation formula of the mean square error loss function described in S2.3 is as follows: , In the formula, represents the mean square error value; Represents the number of patient data; An index representing patient data; Representative The actual value of each patient's data; Representative The predicted value of the patient data.

5. The clinical data analysis system for intelligent patient monitoring and prediction according to claim 3, characterized in that: The specific steps of the simulation verification module to evaluate the performance of the health prediction module in different scenarios are as follows: S3.1: Collect the current patient data status and the prediction results corresponding to the prediction model, and use them as the initial node. Then use the status after the subsequent prediction model simulation as the child node of the initial node, and calculate the UCB value of each group of child nodes generated; S3.2: Use the upper confidence bound strategy to gradually select the child node with the highest UCB value until reaching a node that is not fully expanded, i.e., a leaf node, expand the selected node, add one or more new prediction scenarios to the current node state, and generate one or more child nodes, where the prediction scenario specifically refers to the performance of the prediction model under new patient data, different feature combinations, or different input data; S3.3: Starting from the expanded new node, simulate the behavior of the prediction model, perform a full simulation starting from the current node, use random sampling or model prediction to obtain the results on the simulation path, record the performance of the model on the simulation path, and calculate the corresponding reward value for the model performance based on the simulation results; S3.4: Backtrack from the termination node to the root node step by step, update the number of visits and cumulative rewards of each node, repeat the selection, expansion, simulation and backtracking steps until the preset iteration time is reached, and select the child node with the highest reward value from the root node step by step until the termination node is reached; S3.5: Use the selected path as the optimal validation path for the model, analyze the performance of the model under the optimal path, evaluate the accuracy and stability of the model in different scenarios and its ability to predict patient health risks, identify the advantages and disadvantages of the model under various feature combinations, time periods or input data, and improve the prediction model based on the analysis results.

6. The clinical data analysis system for intelligent patient monitoring and prediction according to claim 3, characterized in that: The clinical decision support module provides doctors with targeted clinical suggestions in the following specific steps: S4.1: Collect the different health states of patients during treatment and construct the corresponding state space , taking the treatment decisions or drug adjustment plans that medical staff can take as actions in the MDP to construct the action space ; S4.2: Calculate the patient's current health status based on historical data After treatment plan Then transfer to the next health state Probability ,in Representative patient time The health status within Representative patient time The health status within ,and In the range [1, m]; S4.3: Give positive rewards if the treatment plan improves the patient's health; if it causes side effects or worsening, give negative rewards, and construct the corresponding reward function , to evaluate the improvement of the patient's health after taking each treatment plan, in the state space In , a random treatment plan is assigned to each state to initialize the strategy, which is recorded as ; S4.4: Update the value function of each state based on the actions performed by the current policy through the Bellman equation , and use the recursive calculation method to update iteratively step by step , until the value function changes of all states are less than the preset threshold, then in each state, choose to make the value function The maximized action is used as the action of the new strategy and the original strategy is updated; S4.5: Compare the strategies before and after the update. If the two are the same in all states, the strategy is considered to have converged. Otherwise, continue to perform the strategy evaluation and improvement steps using the new strategy until the strategy no longer changes in all states, and then provide the treatment plan in the optimal strategy to the medical staff.

7. The clinical data analysis system for intelligent patient monitoring and prediction according to claim 6, characterized in that: The specific calculation formula of the Bellman equation described in S4.4 is as follows: , In the formula, Representation status The value under the current policy; Represents the next health status; Represents the patient's status Through strategic actions Transfer to state probability; Represents taking the current policy action Time return; represents the discount factor used to weigh current and future rewards; Representation status The value under the current policy.

8. The clinical data analysis system for intelligent patient monitoring and prediction according to claim 6, characterized in that: The specific steps of dynamically adjusting the treatment plan by the decision optimization module are as follows: S5.1: The decision optimization module collects the health status of patients under the treatment plan in the optimal strategy in real time, initializes the pheromone and heuristic function of each treatment path in the treatment plan, and then randomly generates multiple groups of individuals, and uses the health status of the current patient as the initial node; S5.2: According to the pheromone and heuristic function of each treatment path, the selection probability of all treatment paths is calculated in the current state, and the next action is selected by random selection according to each selection probability; S5.3: After each group of individuals selects a path, they simulate the treatment results of the treatment path, record the effect of each path based on the simulated treatment results, and update the pheromones on the path after each group of individuals completes its path; S5.4: Repeat the path selection and pheromone update. In each iteration, all individuals will reselect the path, update the pheromone, and adjust the priority of path selection according to the simulation results until the maximum number of iterations is reached or the pheromone concentration changes converge to the preset range; S5.5: After completing all iterations, evaluate the pheromone concentrations of all treatment pathways, select the pathway with the highest pheromone concentration, record the corresponding treatment methods, sort out each treatment plan in the best pathway, and adjust the treatment plans generated by the clinical decision support module based on the generated groups of treatment plans. At the same time, feed back each group of plans to medical staff through the visualization module for review and selection.

Images