Nursing decision support system and nursing decision support method

Abstract

The invention discloses a nursing decision support system and a nursing decision support method, and relates to the field of health care, in the system, a data acquisition module is used for acquiring multi-source data in real time; the data processing module is used for integrating multi-source data by adopting a natural language processing technology and a data mining technology to obtain a unified data set; the graph construction module is used for taking each type of data in the unified data set as a knowledge node, analyzing association relationships among the knowledge nodes by adopting a data mining technology, endowing each association relationship with a weight value, and constructing a knowledge graph according to the knowledge nodes, the association relationships and the weight values; the dynamic updating module is used for regularly updating the knowledge graph according to the currently acquired multi-source data; the decision support module is used for calling the current knowledge graph to generate nursing measures of the patient, comprehensive, accurate and real-time nursing decision support can be provided, and the nursing experience and nursing quality of the patient are improved.

Description

Technical Field

[0001] The present application relates to the field of health care, and in particular to a nursing decision support system and a nursing decision support method. Background Art

[0002] As the global population ages, the elderly patient population is expanding rapidly. How to effectively respond to the nursing needs of this special group has become one of the important research directions of nursing science. However, the relevant nursing decision support systems face the following challenges in responding to the complex nursing needs of elderly patients: (1) heterogeneity and integration difficulty of multi-source data; (2) lack of personalized nursing support; and (3) delayed updating of nursing information.

[0003] Taking elderly HIV (Human Immunodeficiency Virus) / AIDS (Acquired Immune Deficiency Syndrome) patients as an example, specifically, for challenge (1), the nursing needs of elderly HIV / AIDS patients involve a variety of health data, such as clinical medical records, biomarkers, genotype information, living habits, and environmental data. These data often come from different data sources with different formats, contents, and accuracy. Due to the lack of effective data standardization and integration methods, the relevant nursing decision support system cannot effectively integrate these heterogeneous data and apply them to nursing decisions, resulting in the lack of personalization and comprehensiveness of nursing plans.

[0004] Taking elderly HIV / AIDS patients as an example, specifically, for challenge (2), the relevant nursing decision support systems mostly adopt fixed nursing processes, lack consideration of individual differences, and are difficult to adapt to the complex health conditions of elderly HIV / AIDS patients. For example, many nursing decision support systems can only provide experience-based nursing advice, but cannot generate dynamic and personalized nursing plans based on individual characteristics such as the patient's genotype, biomarkers, and lifestyle habits, resulting in low patient care quality and serious waste of nursing resources.

[0005] Taking elderly HIV / AIDS patients as an example, specifically, for challenge (3), since the condition of elderly HIV / AIDS patients changes rapidly, their nursing needs also change accordingly. However, the relevant nursing decision support system usually lacks the ability to dynamically update the patient's health information, and the information in the system may lag behind the patient's actual health status. This lag seriously affects the real-time and accuracy of nursing decisions, resulting in great uncertainty and misjudgment risks for nursing staff in the decision-making process.

[0006] In summary, how to realize nursing decision support comprehensively, accurately and in real time has become an urgent problem to be solved. Summary of the invention

[0007] The purpose of this application is to provide a nursing decision support system and a nursing decision support method, which can provide comprehensive, accurate and real-time nursing decision support and improve the patient's nursing experience and nursing quality.

[0008] To achieve the above objectives, this application provides the following solutions.

[0009] In a first aspect, the present application provides a nursing decision support system, comprising: a data acquisition module, a data processing module, a graph construction module, a dynamic update module and a decision support module.

[0010] The data acquisition module is used to acquire multi-source data in real time; the multi-source data includes: the patient's clinical data, genotype data, biomarkers, living habits data and environmental data; the data processing module is used to integrate the multi-source data using natural language processing technology and data mining technology to obtain a unified data set; the graph construction module is used to treat each type of data in the unified data set as a knowledge node, use data mining technology to analyze the association relationship between each knowledge node, and assign a weight value to each association relationship, and construct a knowledge graph based on the knowledge nodes, the association relationship and the weight value; the dynamic update module is used to regularly update the knowledge graph based on the currently acquired multi-source data; the decision support module is used to call the current knowledge graph to generate patient care measures.

[0011] Optionally, the nursing decision support system further includes: a data transmission module; the data transmission module is used to transmit the multi-source data acquired in real time by the data acquisition module to the data processing module by using encryption technology and fault-tolerant mechanism.

[0012] Optionally, the nursing decision support system further includes: a graph database; the graph database is used to store the multi-source data, the unified data set and the knowledge graph.

[0013] Optionally, the data acquisition module includes: a first acquisition unit, a second acquisition unit and a third acquisition unit.

[0014] The first acquisition unit is used to call the API interface of the hospital information system through the HTTPS protocol, and use the API interface to obtain the patient's clinical data; the second acquisition unit is used to obtain the patient's genotype data and biomarkers through the laboratory information management system; the third acquisition unit is used to use the smart health monitoring device worn by the patient to obtain the patient's living habit data and environmental data through the Internet of Things protocol.

[0015] Optionally, the data processing module includes: a data standardization processing unit, an information extraction unit and a data integration unit.

[0016] The data standardization processing unit is used to perform data cleaning and format conversion on the multi-source data to obtain standardized data; the information extraction unit is used to use natural language processing technology to perform text preprocessing, entity recognition, entity relationship extraction and structured data conversion on the standardized data in sequence to obtain structured information data; the data integration unit is used to use data mining technology to perform data aggregation, data association analysis, data dimensionality reduction, feature selection and data merging on the structured information data in sequence to obtain a unified data set.

[0017] Optionally, the graph construction module includes: a knowledge node determination unit, an association relationship analysis unit, a weight calculation unit and a knowledge graph construction unit.

[0018] The knowledge node determination unit is used to determine a knowledge node for each type of data in the unified data set; the association relationship analysis unit is used to adopt an association rule mining algorithm to mine potential association relationships between the knowledge nodes; the weight calculation unit is used to assign a weight value to each association relationship using the support and confidence in the association rule mining algorithm; the knowledge graph construction unit is used to connect similar knowledge nodes through relationship edges according to the association relationships, and increase the corresponding weight values ​​to obtain a knowledge graph.

[0019] Optionally, the dynamic update module includes: a new node adding unit and a node relationship updating unit.

[0020] The new node adding unit is used to use the Cypher query language to check whether there is a new knowledge node in the currently acquired multi-source data. If so, the new knowledge node is added to the knowledge graph to obtain the knowledge graph after the node is added; the node relationship updating unit is used to use the KNN algorithm to calculate the similarity between the new knowledge node and the existing knowledge nodes in the knowledge graph, determine the association relationship between the new knowledge node and the existing knowledge node based on the similarity to obtain a new association relationship, and assign a weight value to the new association relationship to obtain a new weight value, and use the new association relationship and the new weight value to update the knowledge graph after the node is added to obtain an updated knowledge graph.

[0021] Optionally, the decision support module includes: a nursing measure generation unit; the nursing measure generation unit is used to call the current knowledge graph and use an inference engine, a rule engine or a machine learning model to generate nursing measures for patients.

[0022] Optionally, the data integration unit includes: a data aggregation subunit, a correlation analysis subunit, a data dimension reduction and feature selection subunit and a data merging subunit.

[0023] The data aggregation subunit is used to aggregate data from different sources in the structured information data using the K-means clustering algorithm to obtain aggregated data; the association analysis subunit is used to analyze the association relationship between the aggregated data using the Apriori algorithm to obtain analysis data; the data dimension reduction and feature selection subunit is used to perform dimension reduction and feature selection on the analysis data using the principal component analysis method to obtain data after dimension reduction selection; the data merging subunit is used to merge the data after dimension reduction selection to obtain a unified data set.

[0024] In a second aspect, the present application provides a nursing decision support method, comprising the following steps.

[0025] Acquire multi-source data in real time; the multi-source data includes: patient's clinical data, genotype data, biomarkers, lifestyle data and environmental data.

[0026] The multi-source data are integrated using natural language processing technology and data mining technology to obtain a unified data set.

[0027] Each type of data in the unified data set is taken as a knowledge node, and the association relationship between each knowledge node is analyzed using data mining technology. A weight value is assigned to each association relationship, and a knowledge graph is constructed based on the knowledge nodes, the association relationship and the weight value; the knowledge graph is updated regularly based on the multi-source data currently acquired.

[0028] Call the current knowledge graph to generate nursing measures for patients.

[0029] According to the specific embodiments provided by the present application, the present application has the following technical effects: the present application provides a nursing decision support system and a nursing decision support method. The data processing module uses natural language processing technology and data mining technology to integrate multi-source data to obtain a unified data set. Not only is the data comprehensive, but it also effectively reduces the repetitive work and time consumption in the data integration process, and significantly improves the data processing efficiency; the graph construction module constructs a knowledge graph based on the unified data set, and the decision support module calls the current knowledge graph to generate nursing measures for patients, and generates nursing measures based on the knowledge graph, providing high-precision nursing decision support; the dynamic update module can periodically update the knowledge graph according to the currently acquired multi-source data, ensuring the real-time generation of nursing measures. Therefore, the present application can provide comprehensive, accurate, and real-time nursing decision support, improve the patient's nursing experience and nursing quality, and achieve better nursing results. BRIEF DESCRIPTION OF THE DRAWINGS

[0030] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0031] Figure 1 A structural diagram of a nursing decision support system provided in one embodiment of the present application.

[0032] Figure 2 A flowchart of a nursing decision support method provided in another embodiment of the present application. DETAILED DESCRIPTION

[0033] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0034] In order to make the above-mentioned objects, features and advantages of the present application more obvious and easy to understand, the present application is further described in detail below with reference to the accompanying drawings and specific implementation methods.

[0035] At present, the relevant nursing decision support systems lack effective means to integrate multi-source heterogeneous data, resulting in scattered nursing data and difficulty in using them, making data integration difficult; the relevant nursing decision support systems fail to provide personalized nursing support based on individual differences, resulting in poor patient care outcomes and insufficient personalization of nursing decisions; the relevant nursing decision support systems do not update patient information in a timely manner, affecting the real-time and accuracy of nursing decisions, and their nursing information lags behind.

[0036] Based on the above problems, the embodiments of this application mainly solve them through three aspects. The first is data integration and standardization: using natural language processing and data mining technology to standardize and integrate heterogeneous data from different sources to form a unified data set. The second is the construction and application of knowledge graphs: based on the scientific and precise health model of nursing, a knowledge graph that reflects the overall health of the patient is constructed, and the knowledge graph is used to provide personalized support for nursing decisions. The third is to adopt a dynamic update mechanism: through the dynamic update mechanism of the knowledge graph, the latest patient data and medical research results are integrated in real time to ensure the real-time and accuracy of nursing decisions. This application aims to improve the quality of care and living standards of elderly patients by integrating multi-source heterogeneous data, constructing a dynamic knowledge graph and using it for nursing decision support.

[0037] In an exemplary embodiment, Figure 1 As shown, a nursing decision support system is provided, comprising: a data acquisition module, a data processing module, a graph construction module, a dynamic update module and a decision support module which are connected in sequence.

[0038] The data acquisition module is used to acquire multi-source data in real time; the multi-source data includes: the patient's clinical data, genotype data, biomarkers, living habits data and environmental data; the data processing module is used to integrate the multi-source data using natural language processing technology and data mining technology to obtain a unified data set; the graph construction module is used to treat each type of data in the unified data set as a knowledge node, use data mining technology to analyze the association relationship between each knowledge node, and assign a weight value to each association relationship, and construct a knowledge graph based on the knowledge nodes, the association relationship and the weight value; the dynamic update module is used to regularly update the knowledge graph based on the currently acquired multi-source data; the decision support module is used to call the current knowledge graph to generate patient care measures.

[0039] In another exemplary embodiment of the present application, please refer to Figure 1 The nursing decision support system also includes: a data transmission module; the data transmission module is used to transmit the multi-source data acquired by the data acquisition module in real time to the data processing module using encryption technology and fault-tolerant mechanism.

[0040] In another exemplary embodiment of the present application, the data acquisition module includes: a first acquisition unit, a second acquisition unit and a third acquisition unit. The first acquisition unit is used to call the Application Programming Interface (API) interface of the Hospital Information System (HIS) through the Hypertext Transfer Protocol Secure (HTTPS) protocol, and use the API interface to obtain the patient's clinical data; the second acquisition unit is used to obtain the patient's genotype data and biomarkers through the Laboratory Information Management System (LIMS); the third acquisition unit is used to use the smart health monitoring device worn by the patient to obtain the patient's living habit data and environmental data through the Internet of Things protocol.

[0041] In practical applications, taking elderly HIV / AIDS patients as an example, the data acquisition module collects multi-source data of elderly HIV / AIDS patients from multiple data sources (including hospital information systems, gene detection equipment, health monitoring systems, etc.), and collects data from different sources to the data center through interfaces. The data acquisition module has built-in API call components, data conversion components and data storage components, which respectively realize API interface calls, data flow transmission and standardization, and multi-source data standardization and integration.

[0042] Among them, API interface call: The data collection module establishes API interface calls with data sources such as hospital information systems, laboratory information management systems, genetic testing equipment, and health monitoring equipment to regularly obtain relevant health data of elderly HIV / AIDS patients, that is, to obtain multi-source data.

[0043] Transmission and standardization of data flow: The data acquisition module first obtains data from the data source through a RESTful API or a Web service interface based on the Simple Object Access Protocol (SOAP), and transmits data in JSON or XML format to ensure the reliability and security of data transmission.

[0044] Standardization and integration of multi-source data: After data collection, the raw data is standardized using data conversion components. Data from different sources, such as clinical data, genotype data, biomarkers, lifestyle data, and environmental data, often have different formats and standards. Therefore, the data collection module includes a standardized data conversion tool that uses the ETL (Extract Transform Load) process to extract and convert data into a standardized format for subsequent integration and analysis.

[0045] Among them, the first acquisition unit obtains clinical data such as patient visit records, medication status, and examination results through the hospital information system. The first acquisition unit calls the API interface of the hospital information system through the HTTPS protocol to ensure the security of data during transmission.

[0046] The second acquisition unit obtains the patient's biomarker and genotype data through the laboratory information management system. The second acquisition unit calls the interface of the laboratory equipment, uses the LIS (Laboratory Information System) connection to achieve data interaction, and transmits the data to the data center.

[0047] Among them, the third acquisition unit interacts with the smart health monitoring devices worn by the patient (such as smart bracelets, health trackers, etc.) and obtains the patient's living habit data (such as daily activities, eating habits, etc.) and environmental data (such as temperature, humidity, etc.) through the Internet of Things (IoT) protocol (such as MQTT protocol). These data are transmitted to the central computer via Bluetooth or Wi-Fi, and then transmitted to the data center through the acquisition module interface.

[0048] The data transmission module realizes data transmission and storage. Specifically, communication protocols and encryption measures: During the data transmission process, the data transmission module adopts SSL / TLS (Secure Sockets Layer / Transport LayerSecurity) encryption technology to ensure the security of data transmission and achieve privacy protection. The data is transmitted to the data center through a dedicated API interface and stored in a distributed database for subsequent data analysis and knowledge graph construction. Fault tolerance and retry mechanism for data collection: During the data collection process, if the data source is unavailable, the fault tolerance mechanism embedded in the data transmission module will retry the data collection and record the collection log to ensure the integrity and availability of the data.

[0049] The data acquisition module and data transmission module in this embodiment use technical means such as API calls (RESTful API and SOAP), data transmission protocols (HTTPS, MQTT) transmission, standardized conversion (ETL process) and encrypted transmission (SSL / TLS), combined with specific interfaces of data sources, to ensure efficient collection and transmission of multi-source data. The data acquisition module also has a timed task scheduling function, which can automatically obtain data according to a preset time period, ensuring that the data center has the latest patient health data and supports the real-time and accuracy of nursing decisions.

[0050] In another exemplary embodiment of the present application, the data processing module includes: a data standardization processing unit, an information extraction unit and a data integration unit.

[0051] The data standardization processing unit is used to perform data cleaning and format conversion on the multi-source data to obtain standardized data.

[0052] In practical applications, the standardization process of the data standardization processing unit includes: data extraction (Extract), data transformation (Transform) and data loading (Load). Data extraction: Through the data acquisition module, the data from each data source is extracted into the data standardization processing unit according to a specific time period. Data transformation: Use data standardization tools (such as Talend or Apache Nifi) to clean and convert the data format, including format unification, data type conversion (such as date format unification, numerical type unification) and unit conversion. Use mapping rules to map data field names from different systems to unified field names, such as mapping patient ID fields from different hospitals to a unified "Patient_ID"; apply data validation rules, such as null value checks and format validity checks, to ensure data consistency and accuracy. Data loading: Load the processed standardized data into a unified data storage for subsequent analysis and map construction.

[0053] The data standardization processing unit mainly uses ETL tools, such as Apache NiFi, to perform data standardization and conversion, ensuring that multi-source data meets unified standards after entering the data center, so as to facilitate efficient data integration and analysis.

[0054] The information extraction unit is used to use natural language processing (NLP) technology to perform text preprocessing, entity recognition (NER), entity relationship extraction and structured data conversion on the standardized data in sequence to obtain structured information data. The information extraction unit can also extract information from text data (such as patient medical history) and convert it into structured data for subsequent use.

[0055] Among them, text preprocessing: use NLP toolkits (such as NLTK or spaCy) to preprocess text data such as patient medical history, including tokenization, stopword removal, part-of-speech tagging, and stemming.

[0056] Entity recognition: Use the BERT model (Bidirectional Encoder Representations from Transformers) or spaCy's NER model to identify important entities in the patient's medical history text (such as disease names, symptoms, drug names, family medical history, etc.). These key medical information can be extracted through the BERT model or NER model.

[0057] Entity Relation Extraction: Use dependency parsing technology to identify the relationship between different entities in the text, such as the causal relationship between "patients have diabetes" and "taking certain medications". This process uses training models in deep learning frameworks (such as PyTorch or TensorFlow) for semantic parsing.

[0058] Structured data conversion: The identified entities and their relationships are converted into structured data and stored in JSON or relational databases. Through the structured format, text data can be effectively integrated into the nodes of the knowledge graph, facilitating subsequent query and analysis.

[0059] An example of a processing unit for data standardization: extract "Patient X suffers from hypertension and has been taking medication Y since 2018" from the patient's medical history, identify "hypertension" (disease) and "medication Y" (drug name) through the NER model, and then use the relationship extraction technology to determine the relationship between "patient X" and "hypertension" and "medication Y", and finally convert this information into structured data nodes and store them in the database.

[0060] The data integration unit is used to use data mining technology to perform data aggregation, data association analysis, data dimension reduction, feature selection and data merging on the structured information data in sequence to obtain a unified data set. The data integration unit includes: a data aggregation subunit, a correlation analysis subunit, a data dimension reduction and feature selection subunit and a data merging subunit.

[0061] The data aggregation subunit is used to aggregate data from different sources in the structured information data using a data mining algorithm (such as a K-means clustering algorithm) to find similar patient characteristics and health patterns to obtain aggregated data. For example, based on the patient's living habit data and clinical data, the patients are grouped to identify groups with similar health conditions.

[0062] The association analysis subunit is used to analyze the association relationship between the aggregated data using the Apriori algorithm to obtain analysis data. Specifically, the Apriori algorithm is used to perform association analysis to mine potential association rules from the patient's genotype data, biomarkers, and clinical data. For example, a high correlation between certain genotypes and specific symptoms (such as decreased immunity) is found. This analysis can help identify potential health risks and provide targeted care measures.

[0063] The data dimension reduction and feature selection subunit is used to perform dimension reduction and feature selection on the analysis data using principal component analysis (PCA) to obtain data after dimension reduction and selection. For high-dimensional data, using principal component analysis to reduce dimension can reduce the complexity of the data, improve the efficiency of integration and the accuracy of subsequent analysis.

[0064] The data merging subunit is used to merge the data after dimensionality reduction selection to obtain a unified data set. Specifically, after various types of data (including genotype data, biomarkers, lifestyle data, environmental data, etc.) are aggregated, associated and reduced, the data are merged to generate a unified data set. The data set is stored in the form of a graph database (such as Neo4j) for subsequent knowledge graph construction.

[0065] An example of a data integration unit: Through data mining technology, the lifestyle data and clinical data of different patients are aggregated and analyzed, and it is found that a certain type of patients are more likely to have a decline in immunity due to similar lifestyle habits (such as lack of exercise). This information is integrated and converted into nodes and relationships through data mining technology, and stored in the knowledge graph for use in nursing decision-making. The data integration unit can ensure the consistency and integrity of the data, which is convenient for the subsequent construction of the knowledge graph.

[0066] In another exemplary embodiment of the present application, the graph construction module includes: a knowledge node determination unit, an association relationship analysis unit, a weight calculation unit and a knowledge graph construction unit.

[0067] The knowledge node determination unit is used to determine a knowledge node for each type of data in the unified data set; the association relationship analysis unit is used to use an association rule mining algorithm (such as an Apriori algorithm) to mine the potential association relationships between the knowledge nodes; the weight calculation unit is used to use the support and confidence in the association rule mining algorithm to assign a weight value to each association relationship; the knowledge graph construction unit is used to connect similar knowledge nodes through relationship edges (Edges) according to the association relationship, and increase the corresponding weight value to obtain a knowledge graph.

[0068] In practical applications, the graph construction module defines various information (such as clinical data, genotype data, environmental data, etc.) as knowledge nodes based on the patient's health data. Each node represents specific health information, such as medical history, gene mutation, living habits, etc. Through data mining technology, the logical relationship between each knowledge node is analyzed and established. For example, the logical relationship between certain gene mutations and disease risks is established, and the logical relationship can be explicitly modeled as a node relationship.

[0069] The graph construction module mainly realizes relationship analysis and identification, relationship modeling and relationship weight calculation.

[0070] Relationship analysis and identification: Use data mining technology to analyze the association and causal relationship between knowledge nodes. Specifically, use association rule mining algorithms (such as the Apriori algorithm) to find potential associations between different nodes. For example, the association between genotype data and symptoms in clinical data can be found through association rule mining. For example, some gene mutations are significantly associated with reduced immunity.

[0071] Cluster analysis (such as K-means algorithm) is used to group patients' health data and identify groups of patients with similar health characteristics. For example, a group of patients may have similar genetic characteristics and lifestyle habits, and cluster analysis can be used to aggregate these patients into a node group.

[0072] Relationship modeling: Through the data after cluster analysis, similar knowledge nodes are connected through relationship edges. These relationships represent similarities or associations. For example, the association between lifestyle data and clinical manifestations can be clearly expressed through relationship modeling.

[0073] Dependency analysis: Use dependency extraction algorithms to identify the causal relationship between different information in text data. For example, "a patient's specific disease caused by a gene mutation" can establish a clear causal relationship through dependency analysis and form a connection in the knowledge graph.

[0074] Relationship weight calculation: When establishing node relationships, each relationship is assigned a weight to indicate the importance of the relationship. The weight calculation is based on the frequency and support of data mining. For example, the support and confidence values ​​in the Apriori algorithm are used to assign different weights to relationships to distinguish between strong and weak associations.

[0075] The graph construction module of this embodiment analyzes and models the logical relationship between knowledge nodes through data mining technology to form a knowledge graph that reflects the overall health status of the patient. This knowledge graph can not only provide personalized nursing decision support for nursing staff, but also intuitively display the patient's health information and their interrelationships through visualization tools, helping nursing staff to better understand and deal with complex nursing needs.

[0076] The data processing module provides a reliable data foundation for the definition of knowledge nodes. Only after the data is standardized and integrated can the accuracy and consistency of the knowledge nodes be ensured. The graph construction module relies on the data set of the data processing module to define knowledge nodes and build relationships. These data sets are analyzed through data mining technology to find out the relationship between data, thereby realizing the construction of the knowledge graph. Therefore, data collection and integration are prerequisites for the definition and establishment of the knowledge graph, and the two together constitute the basic framework of the entire nursing decision support system.

[0077] In another exemplary embodiment of the present application, please refer to Figure 1 , the nursing decision support system also includes: a graph database; the graph database is used to store the multi-source data, the unified data set and the knowledge graph.

[0078] Specifically, the graph database uses Neo4j as a storage tool to save the defined knowledge nodes and the relationships between them as a structured knowledge graph. Each node and edge (relationship) is stored in the graph database with a unique identifier for subsequent fast query and retrieval.

[0079] The graph database in this embodiment can ensure that the knowledge graph maintains an efficient data structure during storage and supports fast query and update.

[0080] In another exemplary embodiment of the present application, the dynamic update module includes: a new node adding unit and a node relationship updating unit.

[0081] The new node adding unit is used to use the Cypher query language to check whether there is a new knowledge node in the currently acquired multi-source data. If so, the new knowledge node is added to the knowledge graph to obtain the knowledge graph after the node is added; the node relationship updating unit is used to use the (K-Nearest Neighbor, KNN) algorithm to calculate the similarity between the new knowledge node and the existing knowledge nodes in the knowledge graph, determine the association relationship between the new knowledge node and the existing knowledge node based on the similarity to obtain a new association relationship, and assign a weight value to the new association relationship to obtain a new weight value, and use the new association relationship and new weight value to update the knowledge graph after the node is added to obtain an updated knowledge graph.

[0082] The dynamic update mechanism of the knowledge graph in the dynamic update module is further introduced below.

[0083] (1) Regularly collect patients’ latest health data and research results.

[0084] The dynamic update mechanism relies on the data collection channels established in the previous data collection module and data processing module. These channels regularly collect the latest health data of patients and newly released related medical research results from sources such as hospital information systems, genetic testing equipment, and smart health monitoring equipment.

[0085] The early data collection and integration built an integrated platform for multi-source data, while the dynamic update mechanism uses the same data collection channel to ensure that new data can be automatically injected into the knowledge graph to keep it updated.

[0086] Specifically, by setting up a scheduled task scheduler (such as Cron Jobs) to regularly call the API interface, the latest data can be obtained from different sources. The update principle is: use a message queue (such as Apache Kafka) to process data collection and update requests to ensure the order and consistency of data updates. Whenever new data is collected, it is pushed to a message queue, and the dynamic update module listens to the queue to obtain the latest data in real time.

[0087] (2) Data integration and adjustment and optimization of knowledge graph.

[0088] Data integration: The collected data is first standardized by the data integration component in the dynamic update module to ensure that the data is consistent with the data structure of the existing knowledge graph. Incremental update of nodes and relationships: After the new data is standardized, the system will use the preset algorithm to incrementally update the nodes and relationships in the knowledge graph. The specific incremental update process is as follows.

[0089] Adding new nodes: Using the Cypher query language in the graph database, automatically check whether there are new knowledge nodes based on the new data. If so, add the new node to the graph through the Cypher statement. For example, for a newly identified gene mutation, automatically create a new node and assign it a unique identifier.

[0090] Update of the relationship between nodes: ① Similarity matching algorithm: For newly added nodes, the KNN algorithm is used to determine the similarity between them and existing nodes. If it is found that the new node is related to the existing node, these nodes are connected by establishing an edge. ② Relationship weight adjustment: Use an algorithm based on association rules (such as the Apriori algorithm) to analyze new data and update the relationship weights between nodes. The confidence and support of the association rules are used to calculate the weight of the relationship in order to provide a more accurate reference for nursing decisions.

[0091] The above incremental updates rely on the graph database and use the Cypher query language to incrementally add or update nodes and their relationships. During this process, the dynamic update module uses the Breadth First Search (BFS) algorithm to traverse the existing knowledge graph and find a suitable location for adding new nodes, thereby ensuring the orderliness and logical consistency of the data.

[0092] The above dynamic weight adjustment recalculates the weights of node relationships to reflect the relationship changes brought about by new data, and optimizes them through association rule mining to ensure that the relationship edge weights in the knowledge graph can dynamically adapt to new information.

[0093] The dynamic update module of this embodiment, through regular collection and incremental updates, enables the knowledge graph to timely reflect the patient's latest health status and the latest results of medical research, ensuring the real-time and accuracy of nursing decision support. The construction of the knowledge graph includes two stages: initial construction and subsequent dynamic update. In the initial construction stage, data is collected, integrated and modeled to form a complete knowledge graph; in the dynamic update stage, the knowledge graph after the initial construction is regularly maintained and optimized to keep it updated at any time and adapt to changing clinical needs. The dynamic update mechanism adopted in the dynamic update stage is not to construct a new knowledge graph, but to continuously expand and improve the initial knowledge graph in an incremental manner to ensure that it continues to reflect the patient's latest condition.

[0094] In another exemplary embodiment of the present application, the decision support module includes: a nursing measure generation unit; the nursing measure generation unit is used to call the current knowledge graph and use an inference engine, a rule engine or a machine learning model to generate nursing measures for patients.

[0095] Specifically, the decision support module first implements the call of the knowledge graph, and then provides nursing recommendations based on the knowledge graph by means of information retrieval and matching, decision generation and visual display according to the decision support principles (such as data processing and reasoning engine, rule engine and machine learning model).

[0096] Knowledge graph call: During clinical work, nursing staff access the decision support module through computer equipment and call the knowledge graph to obtain the patient's comprehensive health information, including the patient's clinical data, genotype data, biomarkers, lifestyle data, and environmental data.

[0097] Data processing and reasoning engine: The decision support module contains a data processing and reasoning engine, which is responsible for extracting relevant information from the knowledge graph and analyzing it. The core function of this engine is to match the patient's health information with the nodes and relationships in the knowledge graph and identify the nursing measures related to the patient's health status.

[0098] Rule engine: The decision support module integrates a rule engine (such as Drools) to generate personalized nursing recommendations based on preset nursing decision rules and data relationships in the knowledge graph. The rule engine can parse and apply clinical guidelines and best practices to provide scientific basis for nursing staff.

[0099] Machine Learning Models: Incorporating machine learning models, such as decision trees or random forests, to analyze historical data and identify relationships between nursing interventions and patient outcomes. This enables the decision support module to predict the most effective nursing intervention based on historical data and patient characteristics.

[0100] Information retrieval and matching: The decision support module retrieves node information related to the current patient's condition by querying the knowledge graph, and finds matching nursing measures based on the patient's health status, historical medical records, treatment process, etc. For example, the decision support module can identify the patient's disease type and biomarkers, and recommend the most appropriate intervention plan in combination with relevant nursing measure nodes.

[0101] Decision generation: Based on the retrieved knowledge graph information and the analysis results of the rule engine, the system generates personalized nursing recommendations for patients. These recommendations include but are not limited to disease management plans, mental health interventions, nursing precautions, etc., ensuring that nursing staff can obtain specific and executable recommendations.

[0102] Visual display: All nursing recommendations are presented to nursing staff through a user-friendly interface, which is easy to understand and operate. The decision support module uses visualization technology (such as charts and dashboards) to display the patient's health status and recommended nursing measures, helping nursing staff to conduct a comprehensive analysis.

[0103] The decision support module of this embodiment can provide personalized and efficient decision-making. By combining the comprehensive data of the knowledge graph and intelligent algorithms, it can provide personalized and scientific nursing suggestions to nursing staff, significantly improve the accuracy and efficiency of nursing decisions, and ensure that patients receive the best nursing experience in clinical practice.

[0104] The decision support module of this embodiment can also provide support for clinical decision-making. It can not only assist nursing staff in referring to relevant clinical practices and experiences when formulating nursing plans, but also help nursing staff reduce subjective factors in decision-making and improve nursing quality through data-driven analysis.

[0105] This application realizes multi-source data integration and knowledge extraction, dynamic and personalized nursing support, and real-time updating and decision optimization.

[0106] Multi-source data integration and knowledge extraction: Data standardization and integration are achieved for multi-source data. Data from different sources are converted into structured and actionable information by using data mining and natural language processing technologies, and then integrated into the knowledge graph. Multi-source data integration and knowledge extraction can integrate and model different types of health data in a unified way, thus forming a comprehensive knowledge base that reflects the overall health status of patients.

[0107] Dynamic and personalized nursing support: The current nursing decision support system lacks personalized support, and elderly HIV / AIDS patients often have complex health backgrounds and individual differences. This application provides support to caregivers in a dynamic and personalized manner by constructing a knowledge graph that includes the patient's clinical data, genotype data, lifestyle data, and environmental data. By mapping nursing needs to the knowledge graph, caregivers can obtain decision-making recommendations for the specific health needs of individual patients, thereby achieving precision care and reducing misjudgments caused by standardized nursing processes.

[0108] Real-time update and decision optimization: In order to cope with the rapid changes in the condition of elderly HIV / AIDS patients, a dynamic update mechanism for the knowledge graph is proposed. By regularly collecting the latest patient health data and related clinical research results, the knowledge graph can be automatically optimized and updated to ensure that nursing decisions are always based on the latest patient status and the latest medical progress. This dynamic update mechanism can significantly improve the real-time and effectiveness of nursing and avoid the problem of lagging nursing decisions.

[0109] The advantages of this application mainly include the following aspects.

[0110] 1) Achieve efficient data integration and analysis through data collection and standardization.

[0111] Specifically, by using natural language processing and data mining technology, multi-source data is standardized and integrated, significantly improving the efficiency of data processing. Compared with the common manual data integration method, this application can quickly unify data from different sources into a standard data format, effectively reducing the duplication of work and time consumption in the data integration process. Therefore, this application greatly improves the efficiency of data collection and processing, ensuring that nursing staff can obtain updated nursing information in real time.

[0112] 2) Accurate nursing decision support can be achieved through the definition of knowledge nodes and relationships and personalized nursing decision support.

[0113] Specifically, by constructing a knowledge graph, the health information of elderly HIV / AIDS patients is converted into structured knowledge nodes and their relationships, forming a multi-dimensional individual health model. In the nursing decision support module, nursing staff can directly call the knowledge graph to make personalized nursing decisions. Compared with the decision-making method based on a fixed nursing process, the application of the knowledge graph can comprehensively consider the patient's multi-dimensional health information, thereby generating a nursing plan that is more in line with the patient's individual characteristics. Therefore, this application provides high-precision nursing decision support, significantly reduces the uncertainty and error rate in nursing decisions, and improves the nursing effect.

[0114] 3) The knowledge graph built based on the dynamic update mechanism has real-time and dynamic update capabilities.

[0115] Specifically, through the dynamic update mechanism of the knowledge graph, the graph can be automatically updated according to the latest clinical data, patient feedback and research results, so that the nursing plan can be adjusted as the patient's health status changes and new technologies develop. Related nursing decision-making systems usually lack this flexibility and real-time performance. Nursing staff need to manually update patient information regularly, and this manual update method is time-consuming and error-prone. The dynamic update mechanism of this application effectively solves this problem, ensures the real-time applicability of nursing decisions, and ensures that patients can obtain the latest and most appropriate nursing plans.

[0116] 4) Through the promotion and application expansion of knowledge graphs, it has personalization and scalability.

[0117] Specifically, the knowledge graph design of this application is highly personalized and scalable, and can respond to the specific nursing needs of different patients through the definition of different nodes and relationships. In addition, the application of knowledge graphs is not only applicable to the care of elderly HIV / AIDS patients, but can also be extended to the management of other chronic diseases, such as diabetes, chronic obstructive pulmonary disease, etc. Therefore, this application can not only meet the nursing needs of specific groups, but also promote its application to a wider range of medical scenarios through simple node and relationship adjustments, significantly enhancing the application value and breadth of this application in the field of nursing.

[0118] 5) Optimal allocation of nursing resources can be achieved through personalized nursing decision support and dynamic update mechanism of knowledge graph.

[0119] Specifically, the knowledge graph is used to optimize the allocation of nursing resources and ensure that nursing measures are implemented accurately and effectively. For example, by analyzing the patient's medical history and risk factors, the system can predict possible future health changes in the patient and provide advice to nursing staff in advance, thereby optimizing the allocation and utilization of nursing resources. Compared with related nursing models, the knowledge graph of this application can provide nursing staff with personalized, pre-planned nursing plans, which helps to improve nursing efficiency and reduce manpower and resource waste.

[0120] This application is based on the unique nursing needs of elderly HIV / AIDS patients and aims to improve the accuracy and real-time nature of nursing science. Through the proposed multi-source data integration, knowledge graph construction, and dynamic update mechanism, an innovative nursing support tool is provided for nursing staff, which significantly improves the efficiency and accuracy of nursing. In addition, this method has broad potential for promotion and application, and can provide similar support and decision-making tools in other chronic disease management and nursing fields, promoting the application of precision nursing in clinical practice.

[0121] The uniqueness of the present application is further illustrated below by comparing several alternative solutions with the present application.

[0122] Alternative 1: Traditional nursing decision support system.

[0123] Traditional nursing decision support systems are mostly based on hospital information systems or electronic health records, providing recommendations to patients through established nursing guidelines and fixed nursing processes. These systems usually rely on nursing staff to manually enter data and make nursing decisions based on predefined nursing templates.

[0124] Traditional nursing decision systems can provide nursing support to some extent, especially when the patient's condition is simple and does not require highly personalized care. These systems can provide guidance to caregivers through empirical rules and standard nursing procedures, which is feasible. However, traditional systems usually only process limited clinical data, lack the ability to integrate multi-source heterogeneous data such as genotype data, biomarkers, and lifestyle data, cannot fully reflect the patient's health status, and lack multi-source data integration. Nursing information usually needs to be updated manually, and the update is delayed, making it difficult to keep up with changes in the patient's health status, and lacks the ability to update in real time and dynamically. It is impossible to provide personalized care plans based on complex patient characteristics, resulting in poor accuracy and applicability of care, and limited personalized support.

[0125] Alternative 2: Rule-based nursing expert system.

[0126] Rule-based nursing expert systems use a set of fixed rules to make nursing decisions. These rules are usually summarized by domain experts based on nursing practice and stored in the system to match the patient's specific situation and generate recommendations.

[0127] Rule-based nursing expert systems can provide rapid nursing decisions in specific situations, especially for certain standardized nursing needs. However, the rules of such systems are pre-set and have limited adaptability. Rule-based nursing expert systems mainly rely on patients' medical history and laboratory data, and cannot make full use of multi-source heterogeneous data including genotype data, lifestyle data, environmental data, etc., and the data types are limited. Once the rules are set, it is difficult to change with the progress of clinical research or individual differences of patients. They lack the ability to dynamically adjust and update, and cannot meet the needs of rapidly changing conditions of elderly HIV / AIDS patients. Therefore, their rules are solidified. When faced with multi-dimensional data, the number of rules grows exponentially, and it is difficult to effectively handle complex data dependencies. Therefore, its processing complexity is limited.

[0128] Alternative solution three: Artificial intelligence-assisted nursing decision-making system.

[0129] AI-assisted nursing decision systems usually use machine learning algorithms to learn patterns from a large amount of historical data to predict and assist nursing decisions. These systems are able to process relatively complex data and make a certain degree of prediction of patient health status. Machine learning algorithms can achieve personalized nursing support to a certain extent, and analyze and predict patient data, which is feasible. However, machine learning models, especially deep learning models, often lack good interpretability. For nursing staff, it is difficult to intuitively understand and interpret the nursing recommendations output by the model, which affects the trust in patients and decision implementation, so data interpretation is difficult. Although machine learning models can extract features from data, they lack tools such as knowledge graphs that can reflect the structured relationship between data, making it difficult to provide nursing staff with a clear and comprehensive view of health information. Therefore, the system lacks a structured knowledge graph. Although machine learning models can be updated through regular retraining, the entire process often requires a large amount of data and computing resources, and cannot achieve real-time dynamic updates like knowledge graphs, so their dynamic updates are insufficient.

[0130] Compared with the above alternatives, the unique advantages of this application lie in the following aspects.

[0131] Multi-source data integration and knowledge graph construction: Through natural language processing and data mining technology, it is possible to efficiently integrate multi-source data (clinical data, genotype data, lifestyle data, environmental information, etc.) of elderly HIV / AIDS patients and construct a structured knowledge graph to provide all-round support for nursing decisions.

[0132] Dynamic update and real-time decision-making: Through the dynamic update mechanism, the knowledge graph can be adjusted in real time based on the patient's latest health status and clinical research results to ensure the accuracy and effectiveness of nursing decisions.

[0133] Personalization and visualization support: The knowledge graph provides an intuitive visualization of the patient's health information, allowing caregivers to better understand the patient's overall picture and develop personalized care plans accordingly.

[0134] From the above analysis, it can be seen that although there are some alternatives, such as traditional nursing decision support systems, rule-based nursing expert systems, and artificial intelligence-assisted decision-making systems, these solutions are insufficient in terms of multi-source data integration, personalized nursing support, and real-time dynamic updates, and it is difficult to achieve the overall effect of this application. In contrast, this application has significant advantages in improving the accuracy, real-time and applicability of nursing decisions by combining the nursing science precision health model with knowledge graph technology. Therefore, the solution provided by this application is irreplaceable, especially in terms of the nursing needs of elderly HIV / AIDS patients, showing unique technical advantages.

[0135] Based on the same inventive concept, the embodiment of the present application also provides a nursing decision support method implemented based on the nursing decision support system involved above. The implementation scheme for solving the problem provided by the method is similar to the implementation scheme recorded in the above system, so the specific limitations in one or more nursing decision support method embodiments provided below can refer to the limitations on the nursing decision support system above, and will not be repeated here.

[0136] In an exemplary embodiment, Figure 2 As shown, a nursing decision support method is provided, comprising the following steps.

[0137] Step 201, acquiring multi-source data in real time; the multi-source data includes: clinical data, genotype data, biomarkers, lifestyle data and environmental data of the patient.

[0138] Step 202: Integrate the multi-source data using natural language processing technology and data mining technology to obtain a unified data set.

[0139] Step 203, treat each type of data in the unified data set as a knowledge node, use data mining technology to analyze the association relationship between the knowledge nodes, assign a weight value to each association relationship, and construct a knowledge graph based on the knowledge nodes, the association relationship and the weight value.

[0140] The knowledge graph is updated periodically based on the multi-source data currently acquired.

[0141] Step 204, calling the current knowledge graph to generate nursing measures for the patient.

[0142] Taking elderly HIV / AIDS patients as an example, this application can realize the construction of elderly HIV / AIDS knowledge graph based on the nursing science precision health model. Specifically, it integrates multi-source heterogeneous data and builds a comprehensive knowledge graph on this basis to assist personalized nursing decisions for elderly HIV / AIDS patients. This application aims to standardize and integrate multi-source data such as clinical data, genotype data, biomarkers, lifestyle data and environmental data of elderly HIV / AIDS patients by using advanced data analysis technologies (such as data mining and natural language processing) to generate a structured knowledge graph. This knowledge graph serves as a core decision-making tool in nursing science, providing real-time nursing decision support through a dynamic update mechanism to ensure personalized and precise nursing plans.

[0143] Among them, multi-source data collection and integration: collect information on elderly HIV / AIDS patients from multiple data sources, and standardize and unify these data.

[0144] Knowledge node and relationship modeling: Through data mining and natural language processing technology, define the nodes in the knowledge graph and the relationships between them, and build a multi-dimensional knowledge graph of elderly HIV / AIDS patients.

[0145] Generation and storage of knowledge graphs: Use graph databases to save the constructed nodes and relationships into a knowledge graph structure so that it can be queried and used by nursing staff.

[0146] Dynamic update of knowledge graph: Develop a dynamic update mechanism to continuously optimize and update the knowledge graph based on the latest clinical data, research results and real-time feedback from patients to ensure its adaptability and long-term use value.

[0147] Nursing decision support: Use knowledge graphs to provide personalized nursing recommendations to nursing staff, optimize nursing decisions, and improve the accuracy of nursing and the quality of life of patients.

[0148] Through the implementation of this application method, it is possible to achieve a comprehensive analysis and accurate response to the nursing needs of elderly HIV / AIDS patients, optimize the decision-making process of nursing staff, and improve patients' nursing experience and quality of care.

[0149] The technical features of the above embodiments may be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0150] This article uses specific examples to illustrate the principles and implementation methods of this application. The description of the above embodiments is only used to help understand the method and core ideas of this application. At the same time, for those skilled in the art, according to the ideas of this application, there will be changes in the specific implementation methods and application scope. In summary, the content of this specification should not be understood as limiting this application.

Claims

1. A nursing decision support system, characterized in that: The nursing decision support system includes: a data acquisition module, a data processing module, a graph construction module, a dynamic update module and a decision support module; The data acquisition module is used to acquire multi-source data in real time; the multi-source data includes: the patient's clinical data, genotype data, biomarkers, living habits data and environmental data; the data processing module is used to integrate the multi-source data using natural language processing technology and data mining technology to obtain a unified data set; the graph construction module is used to treat each type of data in the unified data set as a knowledge node, use data mining technology to analyze the association relationship between each knowledge node, and assign a weight value to each association relationship, and construct a knowledge graph based on the knowledge nodes, the association relationship and the weight value; the dynamic update module is used to regularly update the knowledge graph based on the currently acquired multi-source data; the decision support module is used to call the current knowledge graph to generate patient care measures.

2. The nursing decision support system according to claim 1, characterized in that: The nursing decision support system further comprises: a data transmission module; the data transmission module is used to transmit the multi-source data acquired in real time by the data acquisition module to the data processing module by using encryption technology and fault tolerance mechanism.

3. The nursing decision support system according to claim 1, characterized in that: The nursing decision support system also includes: a graph database; the graph database is used to store the multi-source data, the unified data set and the knowledge graph.

4. The nursing decision support system according to claim 1, characterized in that: The data acquisition module includes: a first acquisition unit, a second acquisition unit and a third acquisition unit; The first acquisition unit is used to call the API interface of the hospital information system through the HTTPS protocol, and use the API interface to obtain the patient's clinical data; the second acquisition unit is used to obtain the patient's genotype data and biomarkers through the laboratory information management system; the third acquisition unit is used to use the smart health monitoring device worn by the patient to obtain the patient's living habit data and environmental data through the Internet of Things protocol.

5. The nursing decision support system according to claim 1, characterized in that: The data processing module includes: a data standardization processing unit, an information extraction unit and a data integration unit; The data standardization processing unit is used to perform data cleaning and format conversion on the multi-source data to obtain standardized data; the information extraction unit is used to use natural language processing technology to perform text preprocessing, entity recognition, entity relationship extraction and structured data conversion on the standardized data in sequence to obtain structured information data; the data integration unit is used to use data mining technology to perform data aggregation, data association analysis, data dimensionality reduction, feature selection and data merging on the structured information data in sequence to obtain a unified data set.

6. The nursing decision support system according to claim 1, characterized in that: The graph construction module includes: a knowledge node determination unit, an association relationship analysis unit, a weight calculation unit and a knowledge graph construction unit; The knowledge node determination unit is used to determine a knowledge node for each type of data in the unified data set; the association relationship analysis unit is used to adopt an association rule mining algorithm to mine potential association relationships between the knowledge nodes; the weight calculation unit is used to assign a weight value to each association relationship using the support and confidence in the association rule mining algorithm; the knowledge graph construction unit is used to connect similar knowledge nodes through relationship edges according to the association relationships, and increase the corresponding weight values ​​to obtain a knowledge graph.

7. The nursing decision support system according to claim 1, characterized in that: The dynamic update module includes: a new node adding unit and a node relationship updating unit; The new node adding unit is used to use the Cypher query language to check whether there is a new knowledge node in the currently acquired multi-source data. If so, the new knowledge node is added to the knowledge graph to obtain the knowledge graph after the node is added; the node relationship updating unit is used to use the KNN algorithm to calculate the similarity between the new knowledge node and the existing knowledge nodes in the knowledge graph, determine the association relationship between the new knowledge node and the existing knowledge node based on the similarity to obtain a new association relationship, and assign a weight value to the new association relationship to obtain a new weight value, and use the new association relationship and the new weight value to update the knowledge graph after the node is added to obtain an updated knowledge graph.

8. The nursing decision support system according to claim 1, characterized in that: The decision support module includes: a nursing measure generation unit; The nursing measure generation unit is used to call the current knowledge graph and use an inference engine, a rule engine or a machine learning model to generate nursing measures for the patient.

9. The nursing decision support system according to claim 5, characterized in that: The data integration unit includes: a data aggregation subunit, a correlation analysis subunit, a data dimension reduction and feature selection subunit and a data merging subunit; The data aggregation subunit is used to aggregate data from different sources in the structured information data using the K-means clustering algorithm to obtain aggregated data; the association analysis subunit is used to analyze the association relationship between the aggregated data using the Apriori algorithm to obtain analysis data; the data dimension reduction and feature selection subunit is used to perform dimension reduction and feature selection on the analysis data using the principal component analysis method to obtain data after dimension reduction selection; the data merging subunit is used to merge the data after dimension reduction selection to obtain a unified data set.

10. A nursing decision support method, characterized in that: The nursing decision support method includes: Acquire multi-source data in real time; the multi-source data includes: patient's clinical data, genotype data, biomarkers, lifestyle data and environmental data; Using natural language processing technology and data mining technology to integrate the multi-source data to obtain a unified data set; Each type of data in the unified data set is taken as a knowledge node, and the association relationship between each of the knowledge nodes is analyzed by using data mining technology, and a weight value is assigned to each of the association relationships, and a knowledge graph is constructed according to the knowledge nodes, the association relationships and the weight values; the knowledge graph is updated regularly according to the multi-source data currently acquired; Call the current knowledge graph to generate nursing measures for patients.

Images