Clinical nursing data management method and system for anesthesiology department

Abstract

The invention relates to the technical field related to anesthesia nursing data processing, in particular to an anesthesiology department clinical nursing data management method and system, and the method comprises the steps: collecting anesthesiology department clinical nursing data, configuring a first data mapping table and a second data mapping table through static features and dynamic features according to a platform end and an application end, and generating a nursing knowledge graph, adjacent event features are obtained, and multi-scale risk assessment is combined with dynamic threshold reminding. The technical problems that time sequence correlation and cooperative influence of all events in the anesthesia process are not fully considered, risk assessment standards cannot be dynamically adjusted for different anesthesia stages, and risk misjudgment is likely to be caused are solved, and by configuring a first data mapping table and a second data mapping table and combining time entropies and correlation rule support degrees corresponding to time intervals of adjacent events, the risk assessment standard can be dynamically adjusted according to the time entropies and the correlation rule support degrees. Layered management and accurate mapping of static and dynamic characteristics in clinical nursing data of the anesthesiology department are realized, the risk threshold is dynamically adjusted according to time sequence characteristics of different anesthesia stages, and the accuracy and timeliness of risk assessment are remarkably improved.

Description

Technical Field

[0001] The present invention relates to the technical field related to anesthesia nursing data processing, and in particular to a method and system for managing clinical nursing data in anesthesia department. Background Art

[0002] The clinical nursing of the Department of Anesthesiology generates multi-source heterogeneous data such as basic patient information, anesthesia plan data, and anesthesia monitoring data. These data contain key information such as the patient's personalized anesthesia response patterns and risk evolution characteristics. However, the anesthesia process is highly dynamic and highly time-dependent. From anesthesia induction, maintenance to awakening, slight data fluctuations may cause risks.

[0003] Currently, the clinical nursing data management of the Department of Anesthesiology lacks a unified mapping and fusion mechanism, and the problem of data silos is prominent. In addition, risk assessment mostly relies on the judgment of a single indicator threshold, without considering the temporal correlation and synergistic impact between anesthetic events. Moreover, the risk threshold cannot be dynamically adjusted with the anesthesia stage, which can easily lead to misjudgment or delayed response. It is difficult to adapt to individual differences in patients and changes in clinical scenarios, which restricts the intelligent development and risk prevention and control capabilities of clinical nursing in the Department of Anesthesiology.

[0004] In summary, the existing technology has technical problems such as not fully considering the temporal correlation and synergistic influence of various events during the anesthesia process, being unable to dynamically adjust the risk assessment standards for different anesthesia stages, and easily leading to risk misjudgment. Summary of the Invention

[0005] This application provides a method and system for clinical nursing data management in the Department of Anesthesiology, aiming to solve the technical problems in the existing technology that the temporal correlation and synergistic influence of various events during the anesthesia process are not fully considered, the risk assessment standards cannot be dynamically adjusted for different anesthesia stages, and risk misjudgment is easily caused.

[0006] In view of the above problems, the technical solution to implement this application is: On the one hand, the present application provides a method for managing clinical nursing data of anesthesiology, wherein the method includes: collecting clinical nursing data of anesthesiology including basic patient information, anesthesia plan data, and anesthesia monitoring data; configuring a first data mapping table with static features and dynamic features according to the edge computing node corresponding to the platform end; configuring a second data mapping table with static features and dynamic features according to the visualization unit corresponding to the application end; generating a nursing knowledge graph based on the clinical nursing data of anesthesiology in combination with the first data mapping table and the second data mapping table; at the same time, obtaining the time entropy and association rule support corresponding to the time intervals of adjacent events under the definition of the nursing knowledge graph through a sliding time window; performing partial least squares discriminant analysis based on the time entropy and association rule support corresponding to the time intervals of adjacent events under the definition of the nursing knowledge graph, taking the dose change rate as the input feature, performing multi-scale assessment of potential risks, and issuing risk reminders in combination with dynamic risk thresholds.

[0007] Preferably, time series feature analysis is performed using event tags, which include anesthesia induction stage, anesthesia maintenance stage, and awakening stage; and dynamic risk thresholds are configured based on the time series relationships corresponding to the event tags.

[0008] Preferably, an attention mechanism is used to process outliers in the clinical nursing data of the Department of Anesthesiology, and a federated learning alignment operation is performed on heterogeneous data with inconsistent device identifiers; through the first data mapping table and the second data mapping table, the clinical nursing data of the Department of Anesthesiology is mapped into a high-dimensional vector to generate the nursing knowledge graph.

[0009] Preferably, an active learning-driven dynamic graph update mechanism is set up to screen high-conflict nodes; a graph neural network is used to obtain a conflict-related feature set through the temporal association of the nursing knowledge graph; at the high-conflict nodes, the edge weights and node attributes of the nursing knowledge graph are dynamically adjusted in combination with the conflict-related feature set.

[0010] Preferably, the node with the highest uncertainty in the nursing knowledge graph is selected as the sample to be labeled through Bayesian posterior probability; the samples to be labeled are grouped, and the group type corresponding to the group is associated with the event label; and the edge weights and node attributes of the nursing knowledge graph are updated using local parameters through the grouped samples to be labeled.

[0011] Preferably, a multi-head attention mechanism is used to perform feature fusion on the time entropy corresponding to the time interval and the support of the association rules, and configure the temporal dependency vector of the key risk event; through the temporal dependency vector of the key risk event, a path search is performed in the nursing knowledge graph to formulate a risk propagation path; when there is a sudden drop node of the anesthesia depth index in the risk propagation path, a dynamic compensation mechanism is triggered.

[0012] Preferably, static features are mapped to the local storage area of the edge computing node; based on the dynamic features, a hierarchical index is established according to the time series, and the data update frequency and data transmission strategy are configured; based on the data transmission strategy, a data synchronization protocol is established between the edge computing node and the cloud server, and two-way differential synchronization is used to transmit only the changed data blocks, while recording the data version number.

[0013] Preferably, a sliding time window is defined to traverse the event nodes in the nursing knowledge graph; the time intervals between adjacent events defined by the nursing knowledge graph are determined using a time series to obtain the time entropy corresponding to the time intervals; the event co-occurrence probability is quantified through the association rule support of adjacent events defined by the nursing knowledge graph to form an event association strength matrix.

[0014] Preferably, the structured data in the nursing knowledge graph is spatiotemporally aligned with the image data and audio data during the anesthesia process, and a generative adversarial network is used to formulate a virtual risk scenario; based on the virtual risk scenario, training data is collected, and the training data is fed back to optimize the nursing knowledge graph.

[0015] On the other hand, the present application provides an anesthesia clinical nursing data management system, wherein the system includes: a data collection module: collecting anesthesia clinical nursing data including patient basic information, anesthesia plan data, and anesthesia monitoring data; a data mapping module: configuring a first data mapping table with static features and dynamic features according to the edge computing node corresponding to the platform end; configuring a second data mapping table with static features and dynamic features according to the visualization unit corresponding to the application end; a nursing knowledge graph generation module: generating a nursing knowledge graph based on the anesthesia clinical nursing data in combination with the first data mapping table and the second data mapping table; a sliding analysis module: at the same time, obtaining the time entropy and association rule support corresponding to the time intervals of adjacent events defined by the nursing knowledge graph through a sliding time window; a risk reminder module: performing partial least squares discriminant analysis based on the time entropy and association rule support corresponding to the time intervals of adjacent events defined by the nursing knowledge graph, taking the dose change rate as the input feature, performing multi-scale assessment of potential risks, and providing risk reminders in combination with dynamic risk thresholds.

[0016] In summary, one or more technical solutions provided in this application, by configuring the first and second data mapping tables, combined with the time entropy and association rule support corresponding to the time intervals of adjacent events, perform multi-scale assessment of potential anesthesia risks, realize hierarchical management and precise mapping of static and dynamic features in clinical nursing data of the anesthesia department, dynamically adjust the risk threshold according to the temporal characteristics of different anesthesia stages, and significantly improve the accuracy and timeliness of risk assessment. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] Figure 1 A flowchart of a method for managing clinical nursing data in anesthesiology is provided for this application.

[0018] Figure 2 A structural diagram of an anesthesiology clinical nursing data management system is provided for this application.

[0019] Explanation of the accompanying drawings: data collection module M100, data mapping module M200, nursing knowledge graph generation module M300, sliding analysis module M400, risk reminder module M500. DETAILED DESCRIPTION

[0020] Example 1: The present application will be described in detail below with reference to the accompanying drawings. Figure 1 As shown, the present application provides a method for managing clinical nursing data in anesthesiology, wherein the method comprises: S1: Collect clinical nursing data of the anesthesia department including basic patient information, anesthesia plan data, and anesthesia monitoring data; S2: Configure the first data mapping table with static features and dynamic features according to the edge computing node corresponding to the platform end; configure the second data mapping table with static features and dynamic features according to the visualization unit corresponding to the application end.

[0021] Specifically, the patient's basic information covers static data such as the patient's age, gender, weight, medical history, etc., which is relatively stable during the anesthesia process and is the basis for formulating anesthesia plan data; anesthesia plan data includes the selection, dosage, and administration time of anesthetic drugs; anesthesia monitoring data is dynamic data generated in real time during anesthesia, such as heart rate, blood pressure, blood oxygen saturation, anesthesia depth index, etc., which reflects the patient's physiological changes under anesthesia and has high frequency and strong time series.

[0022] Edge computing nodes refer to computing devices deployed close to data sources, such as monitors and anesthesia machines in operating rooms, which can perform preliminary processing on the collected data, reduce data transmission delays, and improve the real-time performance of data processing; visualization units are terminal devices used to view and analyze nursing data, such as anesthesia monitoring screens and mobile nursing workstations, which need to present complex data in an intuitive and easy-to-understand manner; static features and dynamic features are the two basic attributes of data. Static features, such as the patient's medical history and allergy history, remain basically unchanged during the anesthesia process; dynamic features, such as real-time physiological monitoring indicators, will continue to change as anesthesia progresses.

[0023] Implementation steps: Collecting anesthesia clinical nursing data, including basic patient information, anesthesia plan data, and anesthesia monitoring data, is the foundation of the entire data management method. Basic patient information provides the patient's individualized background for subsequent anesthesia plan formulation and risk assessment. For example, a 60-year-old male patient weighing 70 kg and with a history of hypertension will have a direct impact on the selection and dosage of anesthetic drugs.

[0024] The anesthesia plan data clarifies the specific anesthesia method and drug use plan. For example, if general anesthesia is selected, propofol, remifentanil and other drugs are used, with initial doses of X mg / kg and Y μg / kg, respectively. These data guide the implementation of anesthesia; anesthesia monitoring data is key data generated in real time during anesthesia. Taking heart rate as an example, the normal adult heart rate is 60-100 beats / minute. During the anesthesia induction phase, the patient's heart rate decreases due to the effect of anesthetic drugs. Real-time monitoring of heart rate changes can promptly detect potential circulatory system risks.

[0025] According to the edge computing node corresponding to the platform side, the first data mapping table is configured with static features and dynamic features, and the original data collected by the edge computing node is classified and processed according to static and dynamic features. For example, the patient's medical history (static features) is stored in the local storage area of the edge computing node. At the same time, a hierarchical index is established for real-time dynamic feature data such as heart rate and blood pressure according to the time series, and a data update frequency of once per minute is configured. A data transmission strategy is also formulated. Data is transmitted to the cloud server only when the data change exceeds the set threshold, reducing the data transmission volume and transmission delay.

[0026] Based on the corresponding visualization unit on the application side, a second data mapping table is configured with static and dynamic features. This allows medical staff to clearly distinguish between static and dynamic information when viewing the data, improving data readability and usability. For example, on the visualization interface, basic patient information (static features) is displayed in table format, while dynamic monitoring data such as heart rate and blood pressure are displayed in real time as a line chart. This allows medical staff to quickly obtain key information and support clinical decision-making. These steps provide a structured data foundation for subsequent risk assessment, ensuring efficient and accurate data processing.

[0027] S3: Based on the clinical nursing data of the anesthesia department, combined with the first data mapping table and the second data mapping table, a nursing knowledge graph is generated; S4: At the same time, through a sliding time window, the time entropy and association rule support corresponding to the time intervals of adjacent events defined by the nursing knowledge graph are obtained; S5: Partial least squares discriminant analysis is performed based on the time entropy and association rule support corresponding to the time intervals of adjacent events defined by the nursing knowledge graph, and the dose change rate is used as the input feature to perform multi-scale assessment of potential risks, and risk reminders are issued in combination with dynamic risk thresholds.

[0028] Specifically, the nursing knowledge graph is a data model based on a graph structure, which is used to represent entities (such as patients, drugs, physiological indicators, etc.) and their relationships in the clinical nursing data of the anesthesia department. It can integrate scattered data into a knowledge network with semantic associations, facilitating in-depth mining of the information contained in the data; the time entropy corresponding to the time interval is a concept in information theory, which is used here to quantify the uncertainty of the time interval between adjacent events in the nursing knowledge graph. The more random the time interval, the higher the entropy value, indicating that the temporal regularity of the event is weaker.

[0029] The support of association rules is used to measure the frequency of co-occurrence between different events. The higher the support, the greater the probability of two events occurring at the same time and the stronger the correlation between the events. Partial least squares discriminant analysis can classify and predict potential risks by finding the optimal linear combination between independent variables (such as dose change rate) and dependent variables (such as risk level) when processing multivariate data. It is particularly suitable for complex data scenarios with multicollinearity between variables. Dynamic risk threshold refers to a risk judgment standard that is automatically adjusted according to the different stages of the anesthesia process (such as anesthesia induction, maintenance, and awakening stages) and the patient's real-time physiological state. Compared with fixed thresholds, it can better adapt to the dynamic changes of the anesthesia process and improve the accuracy and timeliness of risk warnings.

[0030] Execution steps: The process of generating a nursing knowledge graph based on the clinical nursing data of the Department of Anesthesiology, combined with the first data mapping table and the second data mapping table, is a process of deep fusion and semantic association of static and dynamic data that have been classified and processed. For example, in the first data mapping table, the patient's medical history (static features) and real-time heart rate data (dynamic features) are initially processed by the edge computing node and are associated in the nursing knowledge graph as the edge between the "patient node" and the "heart rate monitoring node". The weight of the edge can be dynamically adjusted according to the relevance of the data.

[0031] At the same time, the second data mapping table combines the visualization requirements of the application side with the data characteristics, so that the knowledge graph not only stores the original data, but also stores the intermediate results prepared for visualization presentation, such as converting the patient's blood pressure change trend data into a format suitable for displaying in a line chart on the monitoring screen, and then plays a key role in converting the original data into a knowledge structure with semantic association and operability, providing a basis for subsequent in-depth analysis and risk assessment.

[0032] The time entropy and association rule support corresponding to the time intervals of adjacent events within the nursing knowledge graph are obtained through a sliding time window. The sliding time window is a data processing technology that can traverse the event sequence in the knowledge graph in sequence according to the set time length and step size, thereby capturing the time interval changes and association relationships of events at different time scales. For example, when the sliding time window is set to 5 minutes and the step size is 1 minute, when analyzing the event sequence in the anesthesia induction stage, the time intervals and their change patterns between adjacent events (such as drug injection events and patient physiological indicator change events) within each 5-minute window can be obtained, and the time entropy corresponding to the time interval can be calculated to evaluate the uncertainty of the event occurrence time; at the same time, the number of co-occurrences between events is counted, and the association rule support is calculated to quantify the association strength between events to form an event association strength matrix.

[0033] For example, within 5 minutes after the injection of a certain anesthetic drug, the support degree of the association rule between the patient's heart rate decrease event and the blood pressure decrease event reaches 0.8 (that is, in 80% of cases, these two events will occur simultaneously), which shows that there is a strong correlation between the two events. The dose change rate is used as the input feature, and the potential risk is evaluated at a multi-scale through partial least squares discriminant analysis. The dose change rate refers to the change amplitude of the anesthetic drug dose per unit time, which is an important indicator reflecting the depth of anesthesia and the patient's physiological response. Furthermore, in the partial least squares discriminant analysis model, the dose change rate is used as the independent variable, and the risk level (such as low risk, medium risk, and high risk) is used as the dependent variable. By training the model with historical data, a mapping relationship between the dose change rate and the risk level can be established.

[0034] Risk reminders are provided in combination with dynamic risk thresholds, which are adjusted according to different anesthesia stages and the patient's real-time status. Furthermore, during the anesthesia induction stage, the patient's physiological system is more sensitive to drugs, and the dynamic risk threshold is relatively low. When the dose change rate exceeds the threshold of the anesthesia induction stage, a risk reminder is immediately issued to remind medical staff to pay attention to changes in the patient's physiological indicators, thereby achieving accurate and dynamic assessment and timely warning of potential risks during anesthesia, significantly improving the accuracy and timeliness of risk assessments.

[0035] Furthermore, in combination with dynamic risk thresholds for risk alerts, this application method includes: Time series feature analysis is performed using event tags, which include anesthesia induction stage, anesthesia maintenance stage, and awakening stage; dynamic risk thresholds are configured based on the time series relationships corresponding to the event tags.

[0036] Specifically, an event label is an identifier used to mark and classify different periods during the anesthesia process, dividing the anesthesia process into anesthesia induction, anesthesia maintenance, and awakening stages. Each stage has its own unique physiological and pharmacological characteristics. Among them, the anesthesia induction stage is the period from the patient's start of receiving anesthetic drugs to entering an unconscious state. During this stage, the patient's physiological indicators change rapidly and the sensitivity to drugs is high; the anesthesia maintenance stage is the period during which the patient maintains a stable anesthesia state during surgery. During this stage, continuous monitoring and adjustment of drug dosage are required to maintain the patient's anesthesia depth; the awakening stage is the period when the patient regains consciousness from the anesthesia state. During this stage, the patient's recovery needs to be closely observed and possible complications need to be dealt with in a timely manner.

[0037] Time series feature analysis is used to study the distribution and change patterns of these event labels in the time series, revealing the time series characteristics of different anesthesia stages; the dynamic risk threshold can automatically adjust the risk judgment criteria according to the time series relationship. Unlike the fixed risk threshold, the dynamic risk threshold can be adaptively adjusted according to the different anesthesia stages and the patient's real-time status, reflecting the risk level more accurately.

[0038] Execution steps: The process of analyzing temporal features with event labels is to divide the anesthesia process into stages such as induction, maintenance, and awakening, and to perform time series statistics and analysis on the events of each stage. For example, by labeling each event with a corresponding stage label in the nursing knowledge graph, the frequency of occurrence, duration, and temporal intervals between events in each stage can be counted. For example, the average temporal interval between the event labels "drug injection" and "heart rate drop" in the anesthesia induction stage is 3 minutes, while in the maintenance stage this interval is extended to 5 minutes, indicating that there are significant differences in the temporal characteristics of events in different stages.

[0039] Through time series feature analysis, we can understand the evolution of risks in each stage and provide data support for the configuration of dynamic risk thresholds. The process of configuring dynamic risk thresholds through the time series relationship corresponding to event labels is to set differentiated risk judgment standards according to different time series features. For example, during the anesthesia induction stage, due to the large fluctuations in the patient's physiological indicators, for the risk event of heart rate drop, the dynamic risk threshold can be set to trigger a risk reminder when the heart rate is lower than 50 beats / minute and lasts for more than 2 minutes; in the maintenance stage, physiological indicators are relatively stable, and a risk reminder will be triggered only when the heart rate is lower than 55 beats / minute and lasts for more than 3 minutes. The dynamic adjustment mechanism makes risk assessment more in line with actual clinical scenarios, effectively reduces the false alarm rate and missed alarm rate, and improves the timeliness and accuracy of risk warnings. In the above steps, the results of time series feature analysis are closely linked to risk assessment, realizing intelligent dynamic adjustment of risk assessment standards, and significantly improving the risk prevention and control capabilities of clinical nursing in the Department of Anesthesiology.

[0040] Furthermore, based on the clinical nursing data of the Department of Anesthesiology, combined with the first data mapping table and the second data mapping table, a nursing knowledge graph is generated. The method of this application also includes: An attention mechanism is used to process outliers in the anesthesiology clinical nursing data, and a federated learning alignment operation is performed on heterogeneous data with inconsistent device identifiers; the anesthesiology clinical nursing data is mapped into a high-dimensional vector through the first data mapping table and the second data mapping table to generate the nursing knowledge graph.

[0041] Specifically, the attention mechanism is an algorithm that simulates the human attention allocation pattern, which can automatically focus on key features when processing data and ignore unimportant information; in the clinical nursing data of the Department of Anesthesiology, outliers are caused by equipment errors, individual differences among patients or sudden conditions, such as instantaneous erroneous readings in heart rate monitoring (such as a normal heart rate of 90 beats / minute suddenly jumping to 150 beats / minute).

[0042] Heterogeneous data with inconsistent device identifiers refers to data from different devices or systems. Due to differences in device models and data formats, the data cannot be directly compared or merged. For example, the blood pressure data from monitor A is in mmHg, while that from monitor B is in kPa. Federated learning is a distributed machine learning method that allows multiple devices or systems to jointly train models by exchanging encrypted model parameters without sharing the original data, thereby protecting data privacy and security. High-dimensional vectors are a multidimensional data representation that can map complex data features into a high-dimensional space, making the intrinsic structure and relationships of the data easier to analyze and mine.

[0043] Execution steps: The process of using the attention mechanism to handle outliers in the clinical nursing data of the Department of Anesthesiology is to automatically identify and focus on key feature points in the data through the algorithm, thereby effectively detecting outliers. For example, when processing heart rate monitoring data, the attention mechanism can automatically focus on the trend and pattern of heart rate changes. When an instantaneous reading that deviates significantly from the trend occurs (such as the normal trend is 70-80 beats / minute, and 150 beats / minute suddenly appears), it will be marked as an outlier.

[0044] At the same time, the federated learning alignment operation is performed on heterogeneous data with inconsistent device identifiers. Under the premise of protecting data privacy, the data of different devices are feature aligned and model trained through the federated learning algorithm. For example, for the blood pressure data of monitor A (unit: mmHg) and monitor B (unit: kPa), the federated learning model can convert the data into a unified standard unit (such as mmHg) locally, train the model parameters locally, and then upload only the encrypted model parameters to the central server for aggregation to obtain a unified blood pressure analysis model applicable to all devices.

[0045] The process of mapping the clinical nursing data of the Department of Anesthesiology into a high-dimensional vector through the first data mapping table and the second data mapping table is to classify and organize the preprocessed and aligned data according to static and dynamic features, and convert them into a high-dimensional vector representation. For example, the patient's static features such as age and gender and dynamic features such as heart rate and blood pressure are mapped to different dimensions respectively to form a high-dimensional vector that comprehensively reflects the patient's status. If the static features include age (1 dimension) and gender (1 dimension), and the dynamic features include heart rate (1 dimension), blood pressure (2 dimensions, systolic pressure and diastolic pressure), and blood oxygen saturation (1 dimension), the high-dimensional vector formed has 6 dimensions.

[0046] The process of generating a nursing knowledge graph is to use these high-dimensional vectors as nodes and construct a graph structure by analyzing the associations between nodes (such as causal relationships and temporal relationships). For example, the patient's high-dimensional vector nodes are connected with drug dosage nodes and physiological indicator change nodes. The weights of the edges can be calculated based on the correlation. In the above steps, the accuracy and consistency of the data are improved through the attention mechanism and federated learning, and the analyzability and interpretability of the data are enhanced through high-dimensional vector mapping and knowledge graph generation.

[0047] Furthermore, the present application method also includes: An active learning-driven dynamic graph update mechanism is set up to screen high-conflict nodes; a graph neural network is used to obtain a conflict-related feature set through the temporal association of the nursing knowledge graph; at the high-conflict nodes, the edge weights and node attributes of the nursing knowledge graph are dynamically adjusted in combination with the conflict-related feature set.

[0048] Specifically, active learning is a machine learning strategy that uses the model to actively screen out the most valuable data for annotation and learning to improve the model's performance. In the nursing knowledge graph scenario, the active learning-driven dynamic graph update mechanism will actively detect and screen out nodes that are highly conflicting with the existing knowledge graph structure. High-conflict nodes refer to nodes that have inconsistent or contradictory relationships with other nodes in the knowledge graph. For example, a node shows that the patient's blood pressure suddenly rises sharply during the anesthesia maintenance stage, while other related nodes show that the blood pressure is stable. This abnormal situation may indicate data errors or potential clinical risks.

[0049] Graph neural networks can use the node and edge relationships in the graph structure to transmit information and learn features; in the nursing knowledge graph, graph neural networks capture the complex temporal associations between nodes. For example, by analyzing the temporal relationship between the anesthetic drug dosage adjustment node and the patient's physiological indicator change node, the impact of the drug on the patient is revealed; the conflict association feature set refers to the set of features related to high-conflict nodes extracted by the graph neural network. These features can reflect the nature and degree of the conflict.

[0050] Execution steps: Set up an active learning-driven dynamic graph update mechanism. The process of screening high-conflict nodes is to monitor and analyze the nursing knowledge graph in real time through active learning algorithms to identify nodes containing data errors or potential risks. For example, in the knowledge graph, a node records that the patient's blood oxygen saturation suddenly dropped to 80% during the anesthesia maintenance stage (the normal range is 95%-100%), while the adjacent nodes show that the anesthetic drug dosage and ventilator parameters have not been adjusted. This abnormal blood oxygen saturation node will be screened as a high-conflict node.

[0051] The process of using graph neural networks to obtain conflict-related feature sets through the temporal association of the nursing knowledge graph is to conduct in-depth analysis of the temporal relationship between high-conflict nodes and their surrounding nodes through graph neural networks, and extract features that can reflect the conflict. Specifically, graph neural networks can analyze the temporal association between the blood oxygen saturation node and the preceding nodes (such as drug dosage, ventilator settings) and subsequent nodes (such as heart rate changes, doctor intervention measures), and extract features such as the blood oxygen decline rate and the accompanying heart rate change pattern to form a conflict-related feature set.

[0052] At high-conflict nodes, the process of dynamically adjusting the edge weights and node attributes of the nursing knowledge graph in combination with the conflict-related feature set is to optimize the structure of the knowledge graph based on the information in the conflict-related feature set. For example, analysis found that the abnormal decrease in blood oxygen saturation was related to the low tidal volume setting of the ventilator. The system will increase the edge weight between the ventilator parameter setting node and the blood oxygen saturation node, and update the attributes of the blood oxygen saturation node at the same time, marking it as "needing attention" and possibly triggering risk reminders. In the above steps, the knowledge graph is continuously optimized, potential problems are discovered through active learning, the causes of the problems are deeply analyzed using graph neural networks, and the graph structure is dynamically adjusted, thereby improving the accuracy and reliability of the nursing knowledge graph.

[0053] Furthermore, the present application method also includes: Through the Bayesian posterior probability, the node with the highest uncertainty in the nursing knowledge graph is selected as the sample to be labeled; the samples to be labeled are grouped, and the group type corresponding to the group is associated with the event label; through the grouped samples to be labeled, the edge weights and node attributes of the nursing knowledge graph are updated using local parameters.

[0054] Specifically, Bayesian posterior probability is used to calculate the probability that a hypothesis is true given certain observational data. In the nursing knowledge graph, the nodes with the highest uncertainty refer to those whose attributes or relationships are not clear enough, have low correlation with other nodes, or have more missing data, representing potential risks or unknown factors in the data; grouping nodes in the nursing knowledge graph is to group similar or related nodes together based on the characteristics and associations of the nodes for batch processing and analysis.

[0055] Event labels refer to identifiers related to key events in the anesthesia process, such as "drug injection", "abnormal heart rate", "delayed awakening", etc. These labels help to group nodes and associate them with the actual stages and events of the anesthesia process; local parameter updates only adjust the parameters of specific nodes and edges in the knowledge graph, rather than globally updating the entire graph. This can more efficiently utilize computing resources and reduce interference with the overall structure of the graph.

[0056] Execution steps: The process of selecting the nodes with the highest uncertainty in the nursing knowledge graph as samples to be labeled through Bayesian posterior probability is to use the probability model to quantify the uncertainty of the nodes and screen out the nodes that most need further confirmation and labeling. For example, in the nursing knowledge graph, a node represents a hypotension event that occurs in the patient during the anesthesia maintenance phase, but the node has a weak correlation with other nodes (such as drug dosage and heart rate changes). The Bayesian posterior probability calculation will show that the node has a high uncertainty; selecting such a node for labeling can help clarify its true relationship with other nodes and improve the accuracy of the knowledge graph.

[0057] The samples to be labeled are grouped and classified according to the characteristics and associations of the nodes. For example, the nodes related to hypotension are divided into one group, and the nodes related to abnormal heart rate are divided into another group. The type of each group is associated with the event label. The process of updating the edge weights and node attributes of the nursing knowledge graph using local parameters based on the grouped samples to be labeled is to perform targeted optimization of the knowledge graph based on the grouped samples. For example, for the hypotension event node group, by analyzing the labeled samples, it is found that there is a strong correlation between it and the anesthetic drug dosage node, so the edge weight between the two is increased, and the attributes of the hypotension node are updated, such as adding the label "caused by anesthetic drug dosage"; in the above steps, the key nodes are screened out by Bayesian posterior probability, and the local parameters are updated after grouping, so that the knowledge graph can more realistically reflect the actual situation during anesthesia.

[0058] Furthermore, the present application method includes: Using a multi-head attention mechanism, the time entropy corresponding to the time interval and the support of the association rules are feature fused to configure the temporal dependency vector of the key risk events; through the temporal dependency vector of the key risk events, a path search is performed in the nursing knowledge graph to formulate a risk propagation path; when there is a sudden drop node of the anesthesia depth index in the risk propagation path, a dynamic compensation mechanism is triggered.

[0059] Specifically, the multi-head attention mechanism can simultaneously focus on multiple features in the data and assign different weights; in the nursing knowledge graph, the time entropy corresponding to the time interval and the support of the association rules are two key features, which respectively reflect the temporal uncertainty and association strength between events; by fusing these two features through the multi-head attention mechanism, a feature vector that comprehensively reflects the temporal dependency of events can be generated, that is, the temporal dependency vector of key risk events.

[0060] Critical risk events refer to those events that may cause serious consequences during the nursing process, such as "persistent hypotension" and "respiratory depression"; the time dependency vector is a mathematical vector that represents the dependency relationship of events in a time series; path search is an algorithm for finding specific paths in a graph structure, which is used to locate potential risk propagation paths in the nursing knowledge graph; the risk propagation path refers to the path in the knowledge graph that propagates from a risk event to other nodes through its association relationships; the dynamic compensation mechanism is a strategy that automatically adjusts system parameters to respond to sudden risks. When a serious risk is detected in the risk propagation path (such as a sudden drop in the anesthesia depth index), the dynamic compensation mechanism will be triggered.

[0061] Execution steps: The multi-head attention mechanism is used to perform feature fusion on the time entropy corresponding to the time interval and the support of the association rules. The process of configuring the temporal dependency vector of the key risk event is to comprehensively analyze the two features through the multi-head attention mechanism to generate a feature vector that can reflect the temporal dependency relationship of the events. For example, in the nursing knowledge graph, there is a strong correlation between the key risk event (such as hypotension) and the preceding event "increased drug dosage". The time entropy corresponding to the time interval is 0.8, and the support of the association rule is 0.7; the two features are fused into a temporal dependency vector through the multi-head attention mechanism, and their weights are respectively represented by the contribution of the time entropy corresponding to the time interval and the support of the association rule to the key risk event.

[0062] The process of path searching in the nursing knowledge graph through the timing dependency vectors of key risk events is to use the path search algorithm to find the risk propagation path. For example, starting from the drug dosage increase node, it is connected to the low blood pressure node through its timing dependency vector, and then connected to the heart rate decrease node through the timing dependency vector of the low blood pressure node to form a risk propagation path.

[0063] When there is a sudden drop in the anesthesia depth index in the risk transmission path, the process of triggering the dynamic compensation mechanism is a key node in monitoring the risk transmission path. Once a sudden drop in the anesthesia depth index is found (such as a sudden drop from 60 to 40, the normal range is 40-60), the system immediately triggers the compensation mechanism. The compensation mechanism will automatically adjust relevant parameters according to preset rules, such as reducing the dose of anesthetic drugs and issuing emergency reminders; in the above steps, key features are integrated through the multi-head attention mechanism, path search locates the risk transmission path, and the dynamic compensation mechanism responds to sudden risks in a timely manner, significantly improving the risk prevention and control capabilities of the nursing process.

[0064] Furthermore, according to the edge computing node corresponding to the platform side, the first data mapping table is configured with static features and dynamic features. The method of the present application includes: Static features are mapped to the local storage area of the edge computing node. Based on the dynamic features, a hierarchical index is established according to the time series, and the data update frequency and data transmission strategy are configured. Based on the data transmission strategy, a data synchronization protocol is established between the edge computing node and the cloud server, using two-way differential synchronization to only transmit the changed data blocks, and recording the data version number.

[0065] Specifically, static features refer to patient information that remains relatively unchanged during anesthesia, such as the patient's age, gender, and medical history; dynamic features refer to physiological indicators that change in real time during anesthesia, such as heart rate, blood pressure, blood oxygen saturation, etc. These data are continuously updated over time; an edge computing node is a computing device located near the data source that can perform preliminary processing on the collected data to reduce data transmission delays and improve the real-time performance of data processing.

[0066] The local storage area is the storage space on the edge computing node, which is used to temporarily store processed data; the hierarchical index is a data index structure that indexes dynamic feature data in layers according to time series for fast query and access; the data update frequency refers to the update interval of dynamic feature data on the edge computing node, for example, once per second; the data transmission strategy refers to the rules for transmitting data on the edge computing node to the cloud server, including the time, conditions and method of transmission; bidirectional differential synchronization is a data synchronization technology that only transmits the changed parts of the data to reduce the amount of data transmission; the data version number is an identifier used to record the number of times the data is updated, and the version number will increase after each data update.

[0067] Execution steps: Map static features to the local storage area of the edge computing node to ensure that the patient's static information can be quickly accessed during the anesthesia process. For example, static features such as the patient's age and medical history are stored in the local storage area of the edge computing node and can be called immediately when needed without having to obtain them from the remote server, reducing latency; based on dynamic features, establish hierarchical indexes according to time series, and configure data update frequency and data transmission strategies to ensure efficient management and timely updating of dynamic data. For example, heart rate monitoring data is updated once per second, and a hierarchical index is established according to time series, so that relevant data can be quickly located when querying heart rate changes within a specific time period.

[0068] The data transmission strategy stipulates that only when the heart rate data changes by more than 3% will the updated data block be transmitted to the cloud server, and the data version number will be recorded at the same time to ensure data consistency and transmission efficiency; based on the data transmission strategy, a data synchronization protocol is established between the edge computing node and the cloud server, and two-way differential synchronization is used to transmit only the changed data blocks, while recording the data version number; by only transmitting the changed data blocks, the data transmission volume is significantly reduced, which can save a lot of bandwidth. At the same time, recording the data version number can ensure that the data on the cloud server and the edge computing node remain synchronized to avoid data conflicts; through the above steps, the real-time and consistency of the data are ensured, providing reliable data support for subsequent risk assessment and decision-making.

[0069] Furthermore, by sliding the time window, the time entropy and association rule support corresponding to the time intervals of adjacent events defined by the nursing knowledge graph are obtained. The method of this application includes: A sliding time window is defined to traverse the event nodes in the nursing knowledge graph; the time intervals between adjacent events defined by the nursing knowledge graph are determined using a time series to obtain the time entropy corresponding to the time intervals; the event co-occurrence probability is quantified through the support of association rules of adjacent events defined by the nursing knowledge graph to form an event association strength matrix.

[0070] Specifically, the sliding time window traverses time series data by sliding a window of fixed length on the time axis; in the nursing knowledge graph, event nodes represent various events related to patient care, such as drug injections, changes in physiological indicators, etc.; the time entropy corresponding to the time interval is an indicator to measure the uncertainty of the time interval between events. The more irregular the interval, the higher the entropy value; the support of association rules is used to quantify the probability of two events occurring at the same time. The higher the support, the stronger the correlation between the events; the event correlation strength matrix is used to store the correlation strength between all event pairs for fast query and analysis.

[0071] Execution steps: define a sliding time window, traverse the event nodes in the nursing knowledge graph, and slide the window on the time axis of the knowledge graph in turn by setting a fixed time window length (such as 5 minutes) and step size (such as 1 minute) to check the event nodes in the window. For example, in the anesthesia induction stage, the sliding time window can capture the physiological indicator change event after the drug injection event; determine the time interval between adjacent events in time series, and obtain the time entropy corresponding to the time interval. In the sliding window, the time interval between drug injection and heart rate drop varies greatly in multiple events. If the calculated time entropy corresponding to the time interval is 0.9, it indicates that the time interval has a high uncertainty.

[0072] The process of quantifying the probability of event co-occurrence and forming an event association strength matrix through the association rule support of adjacent events under the definition of the nursing knowledge graph is to statistically calculate the frequency of co-occurrence of event pairs (such as drug injection and decreased heart rate) in the knowledge graph. For example, if drug injection and decreased heart rate appear together 80 times in a sequence of 100 events, the association rule support is 0.8; the support of all event pairs is organized into a matrix form, and further, the temporal characteristics of the events are captured through sliding time windows and entropy calculations. The association between events is quantified through the association rule support and strength matrix, providing important timing and association information for risk assessment.

[0073] Furthermore, the present application method includes: The structured data in the nursing knowledge graph is spatiotemporally aligned with the imaging data and audio data during the anesthesia process, and a generative adversarial network is used to formulate a virtual risk scenario. Based on the virtual risk scenario, training data is collected and fed back into the nursing knowledge graph to optimize the training data.

[0074] Specifically, structured data refers to information stored in a fixed format in the nursing knowledge graph, such as the patient's age, drug dosage, blood pressure, blood oxygen saturation, etc.; imaging data includes ultrasound, X-rays and other images during anesthesia, which usually have acquisition timestamps; audio data includes voice recordings in the operating room, etc., which usually have acquisition timestamps; spatiotemporal alignment is to match different types of medical data in time and space dimensions for comprehensive analysis; generative adversarial networks are used to generate virtual data that appears to be real, and virtual risk scenarios are data sets that simulate possible medical risk situations. Feedback optimization is to use training data to update the existing nursing knowledge graph.

[0075] Execution steps: Temporally and spatially align the structured data in the nursing knowledge graph with the image and audio data, and ensure the synchronization of multi-source data by matching timestamps and spatial coordinates. For example, align the patient's heart rate data (structured data) with the ultrasound image (image data) and voice recording (audio data) at the same moment; use a generative adversarial network to formulate virtual risk scenarios, and simulate data in risk scenarios such as hypotension by training the generator and discriminator. Furthermore, the generator is used to generate virtual nursing data to make it as close to real data as possible; the discriminator is used to distinguish between the virtual data generated by the generator and real data.

[0076] Through continuous iterative training, the virtual data generated by the generator becomes closer and closer to the real data until the discriminator cannot distinguish between real data and virtual data; possible virtual risk scenarios are defined, such as hypotension, arrhythmia, etc.; training data is collected based on virtual risk scenarios, and the training data includes virtual structured data, image data and audio data, which are used to optimize the nursing knowledge graph; the virtual data and real data are merged, the nodes and edges in the nursing knowledge graph are updated, and the training data is fed back to optimize the nursing knowledge graph, and the structure and parameters of the nursing knowledge graph are adjusted so that it can more accurately identify and predict risk events.

[0077] In summary, the beneficial effects of the embodiments of the present application are: The method adopts the method of collecting clinical nursing data of the Department of Anesthesiology including basic patient information, anesthesia plan data, and anesthesia monitoring data; configuring the first data mapping table with static features and dynamic features according to the edge computing node corresponding to the platform end; configuring the second data mapping table with static features and dynamic features according to the visualization unit corresponding to the application end; generating a nursing knowledge graph based on the clinical nursing data of the Department of Anesthesiology in combination with the first data mapping table and the second data mapping table; at the same time, obtaining the time entropy and association rule support corresponding to the time interval of adjacent events under the limitation of the nursing knowledge graph through the sliding time window; performing partial least squares discriminant analysis based on the time entropy and association rule support corresponding to the time interval of adjacent events under the limitation of the nursing knowledge graph, taking the dose change rate as the input feature, performing multi-scale assessment of potential risks, and issuing risk reminders in combination with dynamic risk thresholds. This application provides a method and system for managing clinical nursing data in the Department of Anesthesiology, which realizes a multi-scale assessment of potential anesthesia risks by configuring the first and second data mapping tables and combining the time entropy and association rule support corresponding to the time intervals of adjacent events. It realizes the hierarchical management and precise mapping of static and dynamic features in the clinical nursing data of the Department of Anesthesiology, dynamically adjusts the risk threshold according to the temporal characteristics of different anesthesia stages, and significantly improves the accuracy and timeliness of risk assessment.

[0078] Example 2, based on the same inventive concept as the method for managing clinical nursing data of anesthesiology department in the above embodiment, Figure 2 As shown, the embodiment of the present application provides an anesthesiology clinical nursing data management system, wherein the system includes: Data collection module M100: collects anesthesia clinical nursing data including patient basic information, anesthesia plan data, and anesthesia monitoring data.

[0079] Data mapping module M200: configures a first data mapping table with static features and dynamic features according to the edge computing node corresponding to the platform end; configures a second data mapping table with static features and dynamic features according to the visualization unit corresponding to the application end.

[0080] Nursing knowledge graph generation module M300: Generates a nursing knowledge graph based on the anesthesiology clinical nursing data in combination with the first data mapping table and the second data mapping table.

[0081] Sliding analysis module M400: At the same time, through the sliding time window, the time entropy and association rule support corresponding to the time intervals of adjacent events defined by the nursing knowledge graph are obtained.

[0082] Risk reminder module M500: performs partial least squares discriminant analysis based on the time entropy and association rule support corresponding to the time intervals of adjacent events defined by the nursing knowledge graph, uses the dose change rate as the input feature, conducts multi-scale assessment of potential risks, and issues risk reminders in combination with dynamic risk thresholds.

[0083] Furthermore, the risk reminder module M500 is further configured to execute the following method: Time series feature analysis is performed using event tags, which include anesthesia induction stage, anesthesia maintenance stage, and awakening stage; dynamic risk thresholds are configured based on the time series relationships corresponding to the event tags.

[0084] Furthermore, the nursing knowledge graph generation module M300 is also used to execute the following method: An attention mechanism is used to process outliers in the anesthesiology clinical nursing data, and a federated learning alignment operation is performed on heterogeneous data with inconsistent device identifiers; the anesthesiology clinical nursing data is mapped into a high-dimensional vector through the first data mapping table and the second data mapping table to generate the nursing knowledge graph.

[0085] Furthermore, the nursing knowledge graph generation module M300 is also used to execute the following method: An active learning-driven dynamic graph update mechanism is set up to screen high-conflict nodes; a graph neural network is used to obtain a conflict-related feature set through the temporal association of the nursing knowledge graph; at the high-conflict nodes, the edge weights and node attributes of the nursing knowledge graph are dynamically adjusted in combination with the conflict-related feature set.

[0086] Furthermore, the nursing knowledge graph generation module M300 is also used to execute the following method: Through the Bayesian posterior probability, the node with the highest uncertainty in the nursing knowledge graph is selected as the sample to be labeled; the samples to be labeled are grouped, and the group type corresponding to the group is associated with the event label; through the grouped samples to be labeled, the edge weights and node attributes of the nursing knowledge graph are updated using local parameters.

[0087] Furthermore, the nursing knowledge graph generation module M300 is also used to execute the following method: Using a multi-head attention mechanism, the time entropy corresponding to the time interval and the support of the association rules are feature fused to configure the temporal dependency vector of the key risk events; through the temporal dependency vector of the key risk events, a path search is performed in the nursing knowledge graph to formulate a risk propagation path; when there is a sudden drop node of the anesthesia depth index in the risk propagation path, a dynamic compensation mechanism is triggered.

[0088] Furthermore, the data mapping module M200 is used to perform the following method: Static features are mapped to the local storage area of the edge computing node. Based on the dynamic features, a hierarchical index is established according to the time series, and the data update frequency and data transmission strategy are configured. Based on the data transmission strategy, a data synchronization protocol is established between the edge computing node and the cloud server, using two-way differential synchronization to only transmit the changed data blocks, and recording the data version number.

[0089] Furthermore, the sliding analysis module M400 is further configured to execute the following method: A sliding time window is defined to traverse the event nodes in the nursing knowledge graph; the time intervals between adjacent events defined by the nursing knowledge graph are determined using a time series to obtain the time entropy corresponding to the time intervals; the event co-occurrence probability is quantified through the support of association rules of adjacent events defined by the nursing knowledge graph to form an event association strength matrix.

[0090] Furthermore, the sliding analysis module M400 is further configured to execute the following method: The structured data in the nursing knowledge graph is spatiotemporally aligned with the imaging data and audio data during the anesthesia process, and a generative adversarial network is used to formulate a virtual risk scenario. Based on the virtual risk scenario, training data is collected and fed back into the nursing knowledge graph to optimize the training data.

[0091] In summary, any step can be stored as a computer instruction or program in an unlimited computer memory and can be called and recognized by an unlimited computer processor, without any unnecessary restrictions.

[0092] Furthermore, the above technical solution only reflects the preferred technical solution of the technical solution of the embodiment of the present application. Some changes that may be made to certain parts thereof by technical personnel in this technical field all reflect the novel principles of the embodiment of the present application. Obviously, technical personnel in this field can make various changes and modifications to the present application without departing from the scope of the present application.

Claims

1. A method for managing clinical nursing data in anesthesiology department, characterized in that: The method comprises: Collect clinical nursing data of the Department of Anesthesiology, including basic patient information, anesthesia plan data, and anesthesia monitoring data; According to the edge computing node corresponding to the platform side, a first data mapping table is configured with static features and dynamic features; according to the visualization unit corresponding to the application side, a second data mapping table is configured with static features and dynamic features; Generate a nursing knowledge graph based on the anesthesiology clinical nursing data in combination with the first data mapping table and the second data mapping table; At the same time, by sliding the time window, the time entropy and association rule support corresponding to the time intervals of adjacent events defined by the nursing knowledge graph are obtained; Partial least squares discriminant analysis is performed based on the time entropy and association rule support corresponding to the time intervals of adjacent events defined by the nursing knowledge graph. The dose change rate is used as an input feature to conduct a multi-scale assessment of potential risks, and risk reminders are issued in combination with dynamic risk thresholds.

2. A method for managing clinical nursing data of anesthesiology department according to claim 1, characterized in that: The method for providing risk reminders in combination with dynamic risk thresholds includes: Time series feature analysis is performed using event labels, including anesthesia induction stage, anesthesia maintenance stage, and awakening stage; The dynamic risk threshold is configured according to the temporal relationship corresponding to the event labels.

3. A method for managing clinical nursing data in anesthesiology department according to claim 2, characterized in that: Based on the clinical nursing data of the anesthesiology department, combined with the first data mapping table and the second data mapping table, a nursing knowledge graph is generated, and the method further includes: An attention mechanism is used to process outliers in the anesthesiology clinical care data and perform federated learning alignment on heterogeneous data with inconsistent device identifiers. The anesthesiology clinical nursing data is mapped into a high-dimensional vector through the first data mapping table and the second data mapping table to generate the nursing knowledge graph.

4. A method for managing clinical nursing data of anesthesiology department according to claim 3, characterized in that: Set up an active learning-driven graph dynamic update mechanism to filter out high-conflict nodes; Utilizing a graph neural network, obtaining a conflict association feature set through temporal association of the nursing knowledge graph; At the high-conflict node, the edge weights and node attributes of the nursing knowledge graph are dynamically adjusted in combination with the conflict-related feature set.

5. A method for managing clinical nursing data of anesthesiology department according to claim 4, characterized in that: The method further comprises: Using Bayesian posterior probability, select the node with the highest uncertainty in the nursing knowledge graph as the sample to be labeled; Grouping the samples to be labeled, and associating the group types corresponding to the groups with the event labels; The edge weights and node attributes of the nursing knowledge graph are updated using local parameters through the grouped samples to be labeled.

6. A method for managing clinical nursing data in anesthesiology department according to claim 5, characterized in that: The method comprises: Using a multi-head attention mechanism, we fuse the temporal entropy corresponding to the time interval with the support of the association rules to configure the temporal dependency vector of key risk events. Through the time sequence dependency vector of the key risk events, a path search is performed in the nursing knowledge graph to formulate a risk propagation path; When there is a sudden drop node of the anesthesia depth index in the risk propagation path, a dynamic compensation mechanism is triggered.

7. A method for managing clinical nursing data of anesthesiology department according to claim 3, characterized in that: According to the edge computing node corresponding to the platform end, a first data mapping table is configured with static features and dynamic features, and the method includes: Mapping static features to the local storage area of the edge computing node; Based on dynamic characteristics, hierarchical indexes are established according to time series, and data update frequency and data transmission strategy are configured; Based on the data transmission strategy, a data synchronization protocol is established between the edge computing node and the cloud server, and two-way differential synchronization is used to transmit only the changed data blocks, while recording the data version number.

8. A method for managing clinical nursing data in anesthesiology department according to claim 7, characterized in that: Obtaining time entropy and association rule support corresponding to time intervals of adjacent events defined by the nursing knowledge graph through a sliding time window, the method comprising: Define a sliding time window and traverse the event nodes in the nursing knowledge graph; Determine the time interval between adjacent events defined by the nursing knowledge graph using a time series, and obtain the time entropy corresponding to the time interval; The event co-occurrence probability is quantified through the association rule support of adjacent events defined by the nursing knowledge graph to form an event association strength matrix.

9. A method for managing clinical nursing data in anesthesiology department according to claim 8, characterized in that: The method comprises: The structured data in the nursing knowledge graph is spatiotemporally aligned with the image data and audio data during the anesthesia process, and a generative adversarial network is used to formulate a virtual risk scenario; Based on the virtual risk scenario, training data is collected, and the training data is fed back to optimize the nursing knowledge graph.

10. An anesthesiology clinical nursing data management system, characterized in that: A method for managing clinical nursing data in anesthesiology according to any one of claims 1 to 9, the system comprising: Data collection module: collects anesthesia clinical nursing data including basic patient information, anesthesia plan data, and anesthesia monitoring data; Data mapping module: configures a first data mapping table based on static features and dynamic features according to the edge computing node corresponding to the platform end; configures a second data mapping table based on static features and dynamic features according to the visualization unit corresponding to the application end; Nursing knowledge graph generation module: generates a nursing knowledge graph based on the anesthesiology clinical nursing data, combined with the first data mapping table and the second data mapping table; Sliding analysis module: At the same time, through the sliding time window, the time entropy and association rule support corresponding to the time intervals of adjacent events defined by the nursing knowledge graph are obtained; Risk reminder module: Partial least squares discriminant analysis is performed based on the time entropy and association rule support corresponding to the time intervals of adjacent events defined by the nursing knowledge graph. The dose change rate is used as the input feature to conduct a multi-scale assessment of potential risks, and risk reminders are issued in combination with dynamic risk thresholds.

Images