Operation anesthesia information intelligent management system

Abstract

The invention discloses a surgical anesthesia information intelligent management system, and relates to the technical field of medical information, and the system comprises a medical data collection module which is used for collecting first metadata in real time; the intelligent analysis module analyzes the first metadata by using a deep learning algorithm to obtain a first analysis result; the knowledge base management module is used for maintaining a surgical anesthesia case knowledge base; the association verification module verifies and optimizes the first analysis result in combination with historical cases to obtain a second analysis result; the monitoring early warning module is used for generating early warning information based on the second analysis result; the information sharing module visually displays the early warning information and updates the electronic medical record; and the system fault tolerance and backup module ensures system availability and data integrity. The invention further discloses a corresponding management method. Through cooperative work of multiple modules, intelligent monitoring, timely early warning and information sharing in the operation anesthesia process are achieved, and operation safety and medical efficiency are improved.

Description

Technical Field

[0001] The present invention relates to the field of medical information technology, and in particular to an intelligent management system for surgical anesthesia information. Background Art

[0002] The surgical anesthesia process is a key link in medical activities, which requires real-time monitoring and precise control of the patient's vital signs and anesthetic dosage. Traditional surgical anesthesia information management mainly relies on manual records and experience judgments, which has problems such as incomplete information collection, low analysis efficiency, and inaccurate complication prediction, making it difficult to support intelligent management of the entire surgical anesthesia process.

[0003] Especially in the anesthesia management of complex surgeries or high-risk patients, it is difficult to timely detect and prevent potential complications by manual labor alone, which increases the medical safety risks. Therefore, how to use information technology to achieve comprehensive monitoring and intelligent analysis of the surgical anesthesia process and improve the ability to predict and warn of complications has become a technical problem that urgently needs to be solved. Summary of the invention

[0004] In view of the problems existing in traditional surgical anesthesia information management, the present invention is proposed.

[0005] Therefore, the problem to be solved by the present invention is how to achieve intelligent management of the surgical anesthesia process and improve monitoring and early warning capabilities, thereby improving surgical safety and success rate.

[0006] In order to solve the above technical problems, the present invention provides the following technical solutions: In the first aspect, an embodiment of the present invention provides an intelligent management system for surgical anesthesia information, which includes a medical data acquisition module, which is used to collect first metadata in real time and transmit the first metadata to a central database; an intelligent analysis module, which is connected to the central database, uses a deep learning algorithm to analyze the first metadata, obtains a first analysis result, and transmits the first analysis result to an associated verification module; a knowledge base management module, which is used to establish and maintain a surgical anesthesia case knowledge base, and provides historical case data support for the associated verification module; the associated verification module, which is connected to the central database and the surgical anesthesia case knowledge base, verifies and optimizes the first analysis result in combination with historical cases, obtains a second analysis result, and updates the surgical anesthesia case knowledge base according to the second analysis result; a monitoring and early warning module, which is used to monitor the patient's status in real time, and when an abnormal change trend or potential complication risk is detected, issues a corresponding level of early warning information and pushes it to the doctor's terminal.

[0007] As a preferred solution of the intelligent management system for surgical anesthesia information described in the present invention, the information sharing module visualizes the second analysis results and automatically updates the electronic medical records; the system fault tolerance and backup module is used to regularly check the system operation status, perform data backup at a fixed time, and start the recovery mechanism when a system abnormality is detected; the first analysis results include abnormal change trend prediction results and potential complication assessment results.

[0008] As a preferred solution of the intelligent management system for surgical anesthesia information described in the present invention, wherein: 3. The processing flow of the intelligent analysis module includes: preprocessing the first metadata to generate a normalized input data set; extracting time series features and clinical indicator features from the input data set to construct a multidimensional feature data set; designing a deep learning model architecture based on a long short-term memory network, the deep learning model architecture includes an input layer, a multi-layer LSTM unit and a fully connected output layer; using a stratified sampling method to divide the multidimensional feature data set into a training set, a validation set and a test set; using the training set to iteratively train the deep learning model, and at the same time using the validation set to evaluate the model performance and adjust the hyperparameters; when the target performance indicator on the validation set has not improved within a preset number of consecutive training cycles, triggering the early stopping mechanism; using the test set to perform a comprehensive performance evaluation on the trained deep learning model; configuring a dual output unit structure in the output layer of the deep learning model.

[0009] As a preferred solution of the intelligent management system for surgical anesthesia information described in the present invention, the processing flow of the associated verification module includes: retrieving historical case data related to the first analysis result from the surgical anesthesia case knowledge base to generate a similar case set; performing feature extraction and cluster analysis on the similar case set to form multiple case clusters; calculating the similarity between the first analysis result and each case cluster, and selecting the case cluster with the highest similarity as the reference cluster; constructing a time series comparison network based on the attention mechanism, performing a time series comparison between the first analysis result and the historical cases in the reference cluster, and generating a time series difference feature; based on the time series difference feature, correcting and optimizing the first analysis result to obtain a second analysis result; comparing the second analysis result with the actual clinical result, and calculating the accuracy and credibility indicators; dynamically updating the surgical anesthesia case knowledge base according to the accuracy and credibility indicators.

[0010] As a preferred solution of the intelligent management system for surgical anesthesia information described in the present invention, wherein: 5. The association verification module also includes setting an anomaly detection threshold. If the similarity between a case and all existing case clusters is lower than the anomaly detection threshold, it will be marked as a potential outlier case, and medical experts will review it to decide whether to retain it in the knowledge base.

[0011] As a preferred solution of the intelligent management system for surgical anesthesia information described in the present invention, the processing flow of the monitoring and early warning module includes: receiving the second analysis result, the second analysis result is the optimized abnormal change trend prediction result and the potential complication assessment result; constructing a multidimensional early warning matrix, the multidimensional early warning matrix includes the abnormal degree dimension and the time urgency dimension; mapping the abnormal change trend prediction result to the abnormal degree dimension of the multidimensional early warning matrix, and mapping the potential complication assessment result to the time urgency dimension of the multidimensional early warning matrix; based on the multidimensional early warning matrix, calculating the comprehensive risk score; determining the early warning level according to the comprehensive risk score, the early warning level includes low-level warning, medium-level warning and high-level warning; generating early warning information according to the early warning level and the second analysis result, and pushing the early warning information to the doctor terminal.

[0012] As a preferred solution of the intelligent management system for surgical anesthesia information described in the present invention, the processing flow of the information sharing module includes: receiving warning information from the monitoring and early warning module; designing a visualization interface based on the warning information, using a line chart to display the changing trend of key indicators, and providing text descriptions; pushing the visualized warning information to the doctor's workstation and mobile terminal in real time; automatically updating the warning information to the patient's electronic medical record, and marking the timeline of key events in the medical record.

[0013] In the second aspect, an embodiment of the present invention provides an intelligent management method for surgical anesthesia information, which includes real-time collection of first metadata of patients, and transmitting the first metadata to a central database; using a deep learning algorithm to analyze the first metadata, generate abnormal change trend prediction results and potential complication assessment results, and form a first analysis result; retrieve relevant historical cases from a surgical anesthesia case knowledge base, and match and compare them with the first analysis result; verify and optimize the first analysis result in combination with historical cases, generate a second analysis result, and update the surgical anesthesia case knowledge base; construct a multidimensional warning matrix based on the second analysis result, calculate the comprehensive risk score, and determine the warning level; generate warning information according to the warning level and the second analysis result; push the warning information to the doctor's terminal in real time, and display it in a visual manner, and update the patient's electronic medical record at the same time.

[0014] The beneficial effects of the present invention are as follows: the present invention realizes the analysis and prediction of abnormal change trends and potential complications by real-time collection of first metadata such as patient vital signs data, anesthetic drug dosage and surgical operation records, combined with deep learning algorithms and surgical anesthesia case knowledge base. The system adopts a multi-dimensional early warning matrix and a comprehensive risk scoring mechanism to improve the accuracy of early warning. Early warning information is pushed to the doctor's terminal in real time through a visual interface, and the electronic medical record is automatically updated, which improves the efficiency of information sharing. At the same time, the system implements distributed storage and multi-level caching mechanisms to enhance the reliability of data management and system performance. In addition, role-based access control and data desensitization technology are adopted to protect patient privacy while promoting information sharing. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0016] Figure 1 This is the module connection diagram of the surgical anesthesia information intelligent management system.

[0017] Figure 2 This is a processing flow chart of the associated verification module of the surgical anesthesia information intelligent management system.

[0018] Figure 3 This is the processing flow chart of the monitoring and early warning module of the surgical anesthesia information intelligent management system. DETAILED DESCRIPTION

[0019] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the specific implementation methods of the present invention are described in detail below in conjunction with the accompanying drawings.

[0020] In the following description, many specific details are set forth to facilitate a full understanding of the present invention, but the present invention may also be implemented in other ways different from those described herein, and those skilled in the art may make similar generalizations without violating the connotation of the present invention. Therefore, the present invention is not limited to the specific embodiments disclosed below.

[0021] Secondly, the term "one embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The term "in one embodiment" that appears in different places in this specification does not necessarily refer to the same embodiment, nor does it refer to a separate or selective embodiment that is mutually exclusive with other embodiments.

[0022] Example 1, reference Figure 1~Figure 3, which is the first embodiment of the present invention, and provides an intelligent management system for surgical anesthesia information. The module connection diagram is as follows Figure 1 As shown, including: The medical data collection module is used to collect the first metadata in real time and transmit the first metadata to the central database.

[0023] The intelligent analysis module is connected to the central database, uses a deep learning algorithm to analyze the first metadata, obtains a first analysis result, and transmits the first analysis result to the associated verification module.

[0024] The knowledge base management module is used to establish and maintain the surgical anesthesia case knowledge base and provide historical case data support for the associated verification module; the knowledge base management module is regularly updated based on new cases and feedback from the associated verification module to ensure the timeliness and accuracy of the knowledge base.

[0025] The associated verification module connects the central database and the surgical anesthesia case knowledge base, verifies and optimizes the first analysis result in combination with historical cases, obtains the second analysis result, and updates the surgical anesthesia case knowledge base according to the second analysis result.

[0026] The monitoring and early warning module is used to monitor the patient's status in real time. When abnormal change trends or potential complication risks are detected, early warning information of the corresponding level will be issued and pushed to the doctor's terminal.

[0027] The information sharing module visualizes the warning information and automatically updates the electronic medical records.

[0028] The system fault tolerance and backup module is used to regularly check the system operation status, perform data backup at a fixed time, and start the recovery mechanism when a system anomaly is detected to ensure the continuous availability of the system and data integrity.

[0029] Among them, the first metadata includes patient vital signs data, anesthetic drug dosage, surgical operation records, etc. The patient vital signs data includes heart rate, blood pressure, respiratory rate, body temperature, blood oxygen saturation, etc.

[0030] It should be noted that the central database is the core component of the system of the present invention, which adopts a distributed architecture design and is used to efficiently store and manage the first metadata, analysis results and knowledge base information. The database supports multiple data types, including structured patient vital signs data and surgical anesthesia indicators, as well as semi-structured analysis results and early warning information. In order to adapt to the characteristics of medical data, the central database implements the fast access function of time series data, supports high-concurrency real-time data writing and complex historical data query. In terms of security, the central database adopts multi-layer encryption and access control mechanisms to ensure the security of sensitive medical data.

[0031] Specifically, the processing flow of the intelligent analysis module includes: preprocessing the first metadata, including outlier detection and processing, missing data interpolation, data normalization, and generating a standardized input data set; extracting time series features and clinical indicator features from the input data set based on the evidence-based medicine knowledge system and statistical methods, and constructing a multidimensional feature data set; designing and implementing a deep learning model architecture based on a long short-term memory network (LSTM), including an input layer, multi-layer LSTM units, and a fully connected output layer; using a stratified sampling method to divide the multidimensional feature data set into a training set, a validation set, and a test set, with the ratio following the industry standard configuration of 60%, 20%, and 20%; using a batch stochastic gradient descent algorithm combined with an adaptive learning rate optimization strategy, and using the training set to iteratively train the deep learning model , and at the same time, the model performance is evaluated and the hyperparameters are adjusted through the validation set; when the target performance indicator on the validation set has not improved within a preset number of consecutive training cycles, the early stopping mechanism is triggered to prevent overfitting; the test set is used to conduct a comprehensive performance evaluation of the trained deep learning model to ensure that the model has good generalization ability and clinical application value; a dual output unit structure is configured in the output layer of the deep learning model, where the first output unit is responsible for predicting the abnormal change trend of the patient's physiological indicators, and the second output unit is used to assess the risk of potential complications; the real-time first metadata processed by feature engineering is input into the trained deep learning model, and the first analysis result is obtained through forward propagation calculation, where the first analysis result includes the abnormal change trend prediction result and the potential complication assessment result.

[0032] The design and implementation process of the deep learning model is as follows: a multi-layer bidirectional LSTM network is constructed as the main architecture to capture the long-term and short-term temporal dependencies of the patient's physiological indicators; a self-attention mechanism is added on top of the LSTM layer to dynamically weight the importance of features at different time steps; residual connections are introduced to alleviate the gradient vanishing problem and improve the stability of model training; a multi-task learning framework is designed, including a shared feature extraction layer and two task-specific output branches; the first output branch adopts a regression task design and uses a mean square error loss function to predict the abnormal change trend of the patient's physiological indicators and output continuous trend prediction values; the second output branch is constructed as a multi-class classification task and uses a cross-entropy loss function to assess the risk of potential complications and output the probability distribution of various complications.

[0033] Specifically, the processing flow chart of the associated verification module is as follows Figure 2As shown, it includes: retrieving historical case data related to the first analysis result from the surgical anesthesia case knowledge base to generate a similar case set; performing feature extraction and cluster analysis on the similar case set to form multiple case clusters; calculating the similarity between the first analysis result and each case cluster, and selecting the case cluster with the highest similarity as the reference cluster; constructing a time series comparison network based on the attention mechanism, performing time series comparison between the first analysis result and the historical cases in the reference cluster, and generating time series difference features; based on the time series difference features, correcting and optimizing the first analysis result to obtain a second analysis result; comparing the second analysis result with the actual clinical result, and calculating the accuracy and credibility indicators; dynamically updating the surgical anesthesia case knowledge base according to the accuracy and credibility indicators, including adding new valid cases and marking low-relevance cases.

[0034] In addition, the association verification module also sets an anomaly detection threshold. When the similarity between a case and all existing case clusters is lower than the threshold, it will be marked as a potential outlier case, and medical experts will review it to decide whether to retain it in the knowledge base.

[0035] Preferably, the association verification module improves the accuracy of case similarity calculation and the flexibility of time series comparison by introducing a dynamically weighted multimodal fusion mechanism and an adaptive time window adjustment strategy. Among them, the multimodal fusion mechanism comprehensively considers the multidimensional characteristics of the patient's physiological indicators, medication information, surgical type, etc., and uses the attention mechanism to adjust the weight of each modality to adapt to the characteristics of different types of surgery; the adaptive time window adjustment strategy adjusts the time span of the time series comparison according to the surgical stage and anesthesia depth, which helps to capture relevant time periods for comparison. In addition, the module also integrates a knowledge representation method based on a graph structure, and uses the associated information in the surgical anesthesia case knowledge base for similarity analysis, providing additional reference for handling complex cases. These improvements have improved the performance of association verification to a certain extent, especially when dealing with unconventional surgeries or high-risk patients.

[0036] Optionally, the surgical anesthesia case knowledge base adopts a multi-level hybrid storage architecture, combining the advantages of relational databases and graph databases. Relational databases are used to store structured patient information and standardized anesthesia records, while graph databases are used to represent complex medical entity relationships and knowledge graphs. This architecture implements standardized data model design and supports entity-relationship mapping. The knowledge base has an incremental learning mechanism and can be dynamically updated based on new cases and clinical feedback. In addition, it provides an efficient multimodal query interface, supports semantic query and rule reasoning, and implements data version control and traceability mechanisms.

[0037] Furthermore, the processing flow chart of the monitoring and early warning module is as follows: Figure 3As shown, it includes: receiving a second analysis result, which is an optimized abnormal change trend prediction result and a potential complication assessment result; constructing a multidimensional early warning matrix, including an abnormality degree dimension and a time urgency dimension; mapping the abnormal change trend prediction result to the abnormality degree dimension of the multidimensional early warning matrix, and mapping the potential complication assessment result to the time urgency dimension of the multidimensional early warning matrix; based on the multidimensional early warning matrix, calculating a comprehensive risk score, the comprehensive risk score reflects the overall abnormality degree of the patient's condition; determining a warning level according to the comprehensive risk score, including a low-level warning, a medium-level warning and a high-level warning; generating warning information according to the warning level and the second analysis result, including a warning level, a description of the abnormal change trend and a description of the potential complication risk; pushing the warning information to the doctor's terminal, and displaying the warning information of different warning levels on the doctor's terminal with different colors and priorities.

[0038] In addition, the monitoring and warning module also provides a user interface that allows authorized medical personnel to adjust the warning threshold within a predetermined range to optimize the warning sensitivity.

[0039] In this embodiment, the abnormal change trend prediction result in the first analysis result is mapped to the abnormal degree dimension of the multi-dimensional warning matrix. Usually, this mapping can be done by numericalizing the abnormal change trend, for example, dividing the abnormal degree into different levels: 0 (no abnormality), 1 (mild abnormality), 2 (moderate abnormality) 3 (highly abnormal). This definition can be determined based on specific clinical criteria or statistical analysis of historical data. Map the potential complication assessment results to the time urgency dimension of the multidimensional early warning matrix. Similarly, time urgency can be divided into multiple levels: 0 (no risk), 1 (mild urgency), 2 (moderate urgency) 3 (High Urgency). This step reflects the urgency of how quickly intervention is needed based on the predicted risk of complications.

[0040] Moreover, the calculation process of the comprehensive risk score maps the second analysis results to a multi-dimensional warning matrix, combines the two dimensions of abnormality and time urgency, and uses weighted summation or matrix search to obtain a score that reflects the overall abnormality of the patient's status. This score is ultimately used to determine the warning level and generate warning information to help doctors make timely clinical decisions.

[0041] Preferably, the monitoring and early warning module achieves more accurate abnormality detection and early warning by introducing dynamic time window analysis and multi-parameter joint evaluation strategies. Among them, the dynamic time window analysis adaptively adjusts the observation window size according to the changing characteristics of different physiological parameters, improving the ability to identify acute changes and slow trends; the multi-parameter joint evaluation strategy comprehensively considers the coordinated changes of multiple related indicators, reduces false alarms caused by fluctuations in a single parameter, and improves sensitivity to complex pathological conditions. These improvements help to reduce the incidence of false positive warnings while maintaining high sensitivity, providing more reliable decision support for medical staff.

[0042] Furthermore, the processing flow of the information sharing module is as follows: receiving warning information from the monitoring and early warning module; designing a visualization interface based on the warning information, including red, yellow and green color codes corresponding to low-level warnings, medium-level warnings and high-level warnings respectively, using a line chart to display the changing trend of key indicators, and providing text descriptions; pushing the visualized warning information to the doctor's workstation and mobile terminal in real time; automatically updating the warning information to the patient's electronic medical record, and marking the timeline of key events in the medical record; generating a system performance report once a month, including indicators such as warning accuracy, doctor scoring statistics and system response time, for continuous improvement of the warning algorithm and display method.

[0043] In addition, the information sharing module also provides a simple five-star rating interface, allowing doctors to evaluate the accuracy and timeliness of the warning.

[0044] Preferably, the information sharing module realizes a safer and more efficient information sharing mechanism by implementing role-based access control (RBAC) and data desensitization technology. Among them, RBAC dynamically adjusts information access rights according to user roles (such as attending physicians, anesthesiologists, and nurses) to ensure that sensitive information is only visible to specific roles; data desensitization technology automatically hides or blurs the patient's identity information when sharing data, protecting patient privacy without affecting medical decisions. These measures help to improve the efficiency of information sharing while enhancing the system's data security and patient privacy protection capabilities, providing more reliable technical support for cross-departmental and cross-institutional medical collaboration.

[0045] Furthermore, the processing flow of the system fault tolerance and backup module is as follows: check the operating status of each functional module every 15 minutes; perform incremental data backup every hour, and perform full data backup every morning; when data abnormalities are detected, automatically switch to the most recent valid backup data; generate a system operation report once a week, including module operating status, data backup records and system availability data, for review by system administrators.

[0046] Preferably, the system fault tolerance and backup modules improve the reliability and performance of the system by implementing distributed storage and multi-level caching mechanisms. Among them, distributed storage adopts data sharding and replication strategies to store data in multiple physical nodes in a dispersed manner, reducing the risk of single point failure; the multi-level caching mechanism establishes a data access hierarchy between memory, solid-state drives, and mechanical hard drives, optimizing the reading speed of hot data. The application of these technologies improves the stability of the system in high-concurrency scenarios. At the same time, through intelligent data scheduling, the overall response speed of the system is improved while ensuring data security, providing medical personnel with a smoother user experience.

[0047] Furthermore, the present embodiment also provides an intelligent management method for surgical anesthesia information, including real-time collection of first metadata such as patients' vital signs data, anesthetic drug dosage, surgical operation records, etc., and transmitting the data to a central database; using a deep learning algorithm to analyze the first metadata, generate abnormal change trend prediction results and potential complication assessment results, and form a first analysis result; retrieve relevant historical cases from a surgical anesthesia case knowledge base, and match and compare them with the first analysis result; verify and optimize the first analysis result in combination with historical cases, generate a second analysis result, and update the knowledge base; construct a multidimensional warning matrix based on the second analysis result, calculate a comprehensive risk score, and determine a warning level; generate warning information according to the warning level and the second analysis result, including a warning level, a description of abnormal change trend, and a description of potential complication risk; push the warning information to the doctor's terminal in real time, and display it in a visual manner, and update the patient's electronic medical record at the same time.

[0048] In summary, the present invention realizes the analysis and prediction of abnormal change trends and potential complications by real-time collection of first metadata such as patient vital signs data, anesthetic drug dosage and surgical operation records, combined with deep learning algorithms and surgical anesthesia case knowledge base. The system adopts a multi-dimensional early warning matrix and a comprehensive risk scoring mechanism to improve the accuracy of early warning. The early warning information is pushed to the doctor's terminal in real time through a visual interface, and the electronic medical record is automatically updated, which improves the efficiency of information sharing. At the same time, the system implements distributed storage and multi-level caching mechanisms to enhance the reliability of data management and system performance. In addition, role-based access control and data desensitization technology are adopted to protect patient privacy while promoting information sharing.

[0049] Example 2, reference Figure 1~Figure 3 , which is the second embodiment of the present invention, provides an intelligent management system for surgical anesthesia information. In order to verify the beneficial effects of the present invention, scientific demonstration is carried out through economic benefit calculation and simulation experiments.

[0050] In order to verify the performance of the surgical anesthesia information intelligent management system of the present invention in practical applications, the research team conducted a 6-month simulation experiment in the anesthesia department of a tertiary hospital. The experiment involved 500 elective surgical cases, covering multiple departments such as general surgery, cardiothoracic surgery, and neurosurgery. The hardware environment used in the experiment included 10 high-performance servers (each equipped with a 64-core CPU, 256GB RAM, and 4TB SSD storage), and 50 doctor workstation terminals. The software environment adopts a Docker-based microservice architecture, uses PostgreSQL as a relational database, Neo4j as a graph database, and deploys Elasticsearch for efficient full-text retrieval.

[0051] Before the experiment began, the research team initialized the system configuration. The sampling frequency of the medical data acquisition module was set to once every 5 seconds to ensure that subtle changes in the patient's vital signs could be captured. The deep learning model of the intelligent analysis module adopted a 5-layer bidirectional LSTM network structure, each layer containing 128 neurons, and the dropout rate was set to 0.3 to prevent overfitting. The knowledge base management module initially imported 100,000 surgical anesthesia case data from the hospital in the past 5 years as the basic knowledge base. The anomaly detection threshold of the association verification module was set to 0.75, that is, when the similarity of a case with all existing case clusters is less than 0.75, it is marked as a potential outlier case. The comprehensive risk score threshold of the monitoring and early warning module is set as follows: less than 3 points is a low-level warning, 3-7 points is a medium-level warning, and 7 points or more is a high-level warning.

[0052] During the experiment, the system collected the patient's heart rate, blood pressure, respiratory rate, body temperature, blood oxygen saturation and other vital signs data, as well as the use of anesthetic drugs and surgical operation records in real time. The intelligent analysis module processed and analyzed this data in real time, generating analysis results every 3 seconds on average. The association verification module completed the matching and optimization process with historical cases within 30 seconds. When the system detects an abnormal situation, it can generate an early warning message within 5 seconds and push it to the terminal device of the relevant doctor.

[0053] Experimental results show that the system performs well in anomaly detection and early warning. In 500 surgeries, the system successfully warned of 37 potential complications, including 10 hypotension events, 8 arrhythmias, 6 respiratory depression, 5 malignant hyperthermia, and 8 other types of complications. Doctors' feedback shows that the system's early warning information provides valuable reference for clinical decision-making in 85% of cases. The system's average response time is 1.2 seconds, which meets the needs of real-time monitoring. The false positive rate is controlled below 5%, which is significantly lower than the 10-15% of traditional monitoring systems.

[0054] In order to more intuitively demonstrate the advantages of the present invention over the prior art, the research team compiled the following comparison table: Table 1 Comparison index table between the system of the present invention and the traditional monitoring system It can be clearly seen from the table that the surgical anesthesia information intelligent management system of the present invention is superior to the traditional monitoring system in multiple key performance indicators. In particular, the system shows significant advantages in terms of early warning accuracy, false positive rate control and complication prediction advance. These improvements not only improve patient safety, but also greatly reduce the workload of medical staff, allowing them to focus more on important clinical decisions.

[0055] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention, which should all be included in the scope of the claims of the present invention.

Claims

1. An intelligent management system for surgical anesthesia information, characterized in that: include, A medical data collection module, used for collecting first metadata in real time and transmitting the first metadata to a central database; an intelligent analysis module, connected to the central database, analyzing the first metadata using a deep learning algorithm to obtain a first analysis result, and transmitting the first analysis result to an association verification module; The knowledge base management module is used to establish and maintain the surgical anesthesia case knowledge base and provide historical case data support for the associated verification module; an association verification module, connecting the central database and the surgical anesthesia case knowledge base, verifying and optimizing the first analysis result in combination with the historical cases to obtain a second analysis result, and updating the surgical anesthesia case knowledge base according to the second analysis result; The monitoring and early warning module is used to monitor the patient's status in real time. When abnormal change trends or potential complication risks are detected, early warning information of the corresponding level will be issued and pushed to the doctor's terminal.

2. The surgical anesthesia information intelligent management system according to claim 1, characterized in that: Also includes, An information sharing module that visualizes the second analysis result and automatically updates the electronic medical record; System fault tolerance and backup module, which is used to regularly check the system operation status, perform data backup at regular intervals, and start the recovery mechanism when a system anomaly is detected; The first analysis results include abnormal change trend prediction results and potential complication assessment results.

3. The surgical anesthesia information intelligent management system according to claim 1, characterized in that: The processing flow of the intelligent analysis module includes: Preprocessing the first metadata to generate a normalized input data set; Extracting time series features and clinical indicator features from the input data set to construct a multidimensional feature data set; Designing a deep learning model architecture based on a long short-term memory network, wherein the deep learning model architecture includes an input layer, multi-layer LSTM units, and a fully connected output layer; Using a stratified sampling method to divide the multidimensional feature data set into a training set, a validation set, and a test set; Iteratively train the deep learning model using the training set, and perform model performance evaluation and hyperparameter adjustment using the validation set; When the target performance indicator on the validation set does not improve within a preset number of consecutive training cycles, an early stopping mechanism is triggered; Use the test set to conduct a comprehensive performance evaluation on the trained deep learning model; Configure a dual output unit structure in the output layer of the deep learning model.

4. The surgical anesthesia information intelligent management system according to claim 1, characterized in that: The processing flow of the association verification module includes: Retrieving historical case data related to the first analysis result from the surgical anesthesia case knowledge base to generate a similar case set; Performing feature extraction and cluster analysis on the similar case set to form multiple case clusters; Calculate the similarity between the first analysis result and each case cluster, and select the case cluster with the highest similarity as the reference cluster; Constructing a time series comparison network based on an attention mechanism, performing a time series comparison between the first analysis result and the historical cases in the reference cluster, and generating time series difference features; Based on the time series difference feature, the first analysis result is corrected and optimized to obtain a second analysis result; Comparing the second analysis result with the actual clinical result, and calculating the accuracy and credibility index; The surgical anesthesia case knowledge base is dynamically updated according to the accuracy and credibility indicators.

5. The surgical anesthesia information intelligent management system according to claim 1, characterized in that: The association verification module also includes: An anomaly detection threshold is set. If the similarity between a case and all existing case clusters is lower than the anomaly detection threshold, it will be marked as a potential outlier case, and medical experts will review it to decide whether to keep it in the knowledge base.

6. The surgical anesthesia information intelligent management system according to claim 1, characterized in that: The processing flow of the monitoring and early warning module includes: receiving a second analysis result, wherein the second analysis result is an optimized abnormal change trend prediction result and a potential complication assessment result; Constructing a multi-dimensional early warning matrix, wherein the multi-dimensional early warning matrix includes an abnormality degree dimension and a time urgency dimension; Mapping the abnormal change trend prediction result to the abnormal degree dimension of the multidimensional early warning matrix, and mapping the potential complication assessment result to the time urgency dimension of the multidimensional early warning matrix; Based on the multi-dimensional early warning matrix, calculating a comprehensive risk score; Determine a warning level according to the comprehensive risk score, wherein the warning level includes a low-level warning, a medium-level warning, and a high-level warning; Generate warning information according to the warning level and the second analysis result, and push the warning information to the doctor terminal.

7. The surgical anesthesia information intelligent management system according to claim 2, characterized in that: The processing flow of the information sharing module includes: Receive warning information from the monitoring and warning module; Design a visual interface based on the warning information, use a line chart to show the trend of key indicators, and provide text descriptions; Push the visualized warning information to the doctor's workstation and mobile terminal in real time; Automatically update early warning information to the patient's electronic medical record and mark the timeline of key events in the medical record.

8. A method for intelligent management of surgical anesthesia information, based on the intelligent management system for surgical anesthesia information according to any one of claims 1 to 7, characterized in that: Also includes, Collecting the first metadata of the patient in real time and transmitting the first metadata to a central database; Analyzing the first metadata using a deep learning algorithm to generate abnormal change trend prediction results and potential complication assessment results, thereby forming a first analysis result; Retrieve relevant historical cases from the surgical anesthesia case knowledge base and match and compare them with the first analysis results; Verify and optimize the first analysis result in combination with historical cases, generate a second analysis result, and update the surgical anesthesia case knowledge base; Based on the second analysis results, a multi-dimensional early warning matrix is ​​constructed, the comprehensive risk score is calculated, and the early warning level is determined; Generate warning information according to the warning level and the second analysis result; The warning information is pushed to the doctor's terminal in real time and displayed in a visual way, while the patient's electronic medical record is updated.

Images