A multiple medication management and control method and system based on multi-database backup synchronization

Abstract

The present invention discloses a multiple medication control method and system based on multi-database backup synchronization, which relates to the field of knowledge graph technology. The method converts drug clinical data into a drug knowledge graph, uses a distributed consistency algorithm to save the drug knowledge graph in a main database and a backup database, retrieves the interaction path from the drug knowledge graph for each group of drugs in a complete prescription, and if an interaction path exists, determines whether there is a pharmacological conflict for each group of drugs based on the interaction path. If a pharmacological conflict exists, each group of drugs with a pharmacological conflict is used as a conflicting drug combination, constructs an action intensity model for the conflicting drug combination based on the drug knowledge graph, inputs the prescription dosage of each drug in the conflicting drug combination, obtains an action intensity score, and issues a control warning to the patient based on the action intensity score; thereby improving the efficiency and accuracy of drug conflict warning.

Description

Technical Field

[0001] The present invention relates to the field of knowledge graph technology, and specifically to a multiple medication management method and system based on multi-database backup synchronization. Background Art

[0002] Currently, medical information technology has been fully implemented, with major hospitals widely adopting information systems such as electronic medical records and clinical decision support systems. These systems carry a wealth of patient medication data, medical order information, and drug interaction knowledge. However, due to the distributed deployment of hospital information systems and the differences in standards and interfaces between different systems, this extremely valuable medication data cannot be effectively integrated and utilized, resulting in data silos.

[0003] At the same time, existing medication control measures, such as order review and medication standardization, still rely primarily on manual experience and rule-making, and are insufficiently covered for complex multidrug use scenarios. With the increasing variety and number of medications patients are taking concurrently, and with the continuous growth of new medication knowledge and evidence, the limitations of manual control methods have become increasingly prominent, and they are no longer able to meet the needs of precise and personalized clinical medication decision-making.

[0004] Patent application publication number CN117291564A discloses a smart medication reminder system and method based on mobile internet technology. The system includes a smart medicine box that collects medication information from users and creates a medication log. A medical management platform uses a pre-set medical knowledge graph to analyze the log and sends the analysis results to medical staff for confirmation. Confirmed improper medication information is then sent to the smart terminal app. The smart terminal app issues reminders based on the improper medication information and sends the information to the reminder contact. This system can promptly identify medication conflicts, missed doses, and arbitrary increases or decreases in medication dosage for the elderly, preventing adverse effects of long-term improper medication use. However, this solution fails to further analyze conflicts arising from medication dosage.

[0005] To this end, the present invention proposes a multiple medication management method and system based on multi-database backup synchronization. Summary of the Invention

[0006] The present invention aims to solve at least one of the technical problems existing in the prior art. To this end, the present invention proposes a multiple medication management method and system based on multi-database backup synchronization, which improves the efficiency and accuracy of drug conflict warning.

[0007] To achieve the above objectives, a multi-drug management method based on multi-database backup synchronization is proposed, which includes the following steps:

[0008] Step 1: Collect drug clinical data and convert the drug clinical data into a drug knowledge graph;

[0009] Step 2: Use a distributed consistency algorithm to save the drug knowledge graph in the primary database and the backup database;

[0010] Step 3: Collect the complete prescription of each patient; for each group of drugs in the complete prescription, retrieve the interaction path from the drug knowledge graph. If an interaction path exists, execute step 4; otherwise, do nothing;

[0011] Step 4: Determine whether there is a pharmacological conflict for each group of drugs based on the interaction pathway. If there is a pharmacological conflict, proceed to step 5; otherwise, no action is taken.

[0012] Step 5: Take each group of drugs with pharmacological conflicts as a conflicting drug combination, and build an action intensity model for the conflicting drug combination based on the drug knowledge graph;

[0013] Step 6: Input the prescribed dosage of each drug in the conflicting drug combination into the effect intensity model to obtain an effect intensity score, and issue a control alert to the patient based on the effect intensity score;

[0014] Specifically, the collection of drug clinical data includes the following steps:

[0015] Step 11: Identify clinical data sources;

[0016] Step 12: List the raw data types of all required clinical data;

[0017] Step 13: Obtain structured data and unstructured data in the original data type from the clinical data source to form drug clinical data;

[0018] The conversion of drug clinical data into a drug knowledge graph comprises the following steps:

[0019] Step 21: extract drug triplets from the structured data and the unstructured data respectively, and combine all the extracted drug triplets into a drug triplet set;

[0020] Step 22: De-duplicate the drug triplet set;

[0021] Step 23: Pre-construct the domain ontology and the relationship ontology between domain ontologies;

[0022] Step 24: Take the node in each drug triple as an instance node in the drug knowledge graph, and take the relationship in each drug triple as the instance relationship between two nodes in the drug triple to construct the drug knowledge graph;

[0023] The method of using the distributed consistency algorithm to save the drug knowledge graph in the main database and the backup database is:

[0024] Deploy a distributed graph database cluster in a multi-data center architecture;

[0025] Configure a rack-aware topology replication strategy to store N replicas on each node of each data center partition; N is the preset number of replicas;

[0026] Use consistency algorithms to manage data consistency;

[0027] Connect to any database node in the distributed graph database cluster through the database client driver. Write requests to the drug knowledge graph are forwarded to the coordination node through the driver and then replicated to other replicas. The configured consistency algorithm is used between the replicas to reach eventual consensus.

[0028] The method of collecting the complete prescription of each patient is:

[0029] Collect all prescriptions issued to each patient during the same medication cycle through the electronic medical record system of each patient to form a complete prescription; the complete prescription includes the name and dosage of each drug provided to the patient;

[0030] For each group of drugs in the complete prescription, the method of retrieving the interaction path from the drug knowledge graph is as follows:

[0031] For each group of drugs in the complete prescription, retrieve the instance node corresponding to each drug name in the drug knowledge graph and use them as the first instance node and the second instance node respectively;

[0032] Extract all associated edges and semantic paths between the first instance node and the second instance node to form a drug action subgraph;

[0033] Search for interaction semantic edges in the drug action subgraph, and use a depth-first algorithm or a breadth-first algorithm to search whether there is a connecting path between the first instance node and the second instance node. If there is a connecting path, each connecting path is regarded as an interaction path;

[0034] The method for determining whether there is a pharmacological conflict in each group of drugs based on the interaction pathway is:

[0035] For each interaction path, if the relationship semantics corresponding to all edges in the interaction path are positive or neutral, it is judged that there is no pharmacological conflict; otherwise, it is judged that there is a pharmacological conflict;

[0036] The method of treating each group of drugs with pharmacological conflicts as a conflicting drug combination and constructing an action intensity model for the conflicting drug combination based on the drug knowledge graph is as follows:

[0037] Define the node attributes of the instance node corresponding to the drug name;

[0038] Define the relationship attributes of each instance relationship;

[0039] Manually select and annotate the expected confidence scores of some paths in the drug knowledge graph as seeds;

[0040] Generate the expected confidence scores for the remaining paths using path length and drug dosage characteristics through pre-set scoring rules;

[0041] Extract the one-hot encoding of each edge length, edge semantic type number, and instance relationship to form a set of path feature vectors;

[0042] Use the expected confidence score of the seed annotation path as the output and the path feature vector of the seed annotation path as the input to train the logistic regression model, then train it again based on the remaining annotation paths, and save the logistic regression model as the action intensity model;

[0043] Import the trained logistic regression model into the primary and backup databases of the distributed graph database cluster;

[0044] The prescription dosage of each drug in the conflicting drug combination is input into the effect intensity model to obtain the effect intensity score as follows:

[0045] Traverse each interaction path of the conflicting drug combination as the target evaluation path;

[0046] Obtain the path feature vector of the target evaluation path as the target path feature vector;

[0047] Input the target path feature vector into the action intensity model to obtain the expected confidence score output by the action intensity model;

[0048] A preset score threshold is set. If the expected confidence score output by any target evaluation path is greater than the preset score threshold, a control alarm will be issued to the patient.

[0049] A multiple medication management and control system based on multi-database backup and synchronization includes a map construction module, a data backup module, a path retrieval module, and a medication management and control module; wherein each module is electrically connected;

[0050] A graph construction module collects drug clinical data, converts the drug clinical data into a drug knowledge graph, and sends the drug knowledge graph to the data backup module and the path retrieval module;

[0051] The data backup module uses a distributed consistency algorithm to save the drug knowledge graph in the main database and the backup database;

[0052] The path retrieval module collects the complete prescription of each patient; for each group of drugs in the complete prescription, the interaction path is retrieved from the drug knowledge graph. If an interaction path exists, it is sent to the drug control module;

[0053] The drug control module determines whether there is a pharmacological conflict in each group of drugs based on the interaction path. If there is a pharmacological conflict, each group of drugs with pharmacological conflicts will be regarded as a conflicting drug combination. An effect intensity model will be constructed for the conflicting drug combination based on the drug clinical data. The prescription dose of each drug in the conflicting drug combination will be input into the effect intensity model to obtain the effect intensity score. Based on the effect intensity score, a control strategy will be provided to the patient.

[0054] An electronic device is proposed, comprising: a processor and a memory, wherein the memory stores a computer program that can be called by the processor;

[0055] The processor executes the above-mentioned multiple medication management and control method based on multi-database backup synchronization by calling the computer program stored in the memory.

[0056] A computer-readable storage medium is provided, on which a rewritable computer program is stored.

[0057] When the computer program is run on a computer device, the computer device is enabled to execute the above-mentioned multiple medication management and control method based on multi-database backup synchronization.

[0058] Compared with the prior art, the present invention has the following beneficial effects:

[0059] The present invention first uses multi-database backup and synchronization technology to integrate patient medication data, drug knowledge bases, etc. in various heterogeneous medical information systems to construct a unified drug knowledge graph. On top of the graph, by mining multimodal information such as path topology characteristics and drug dosage, and integrating the strength of evidence from evidence-based medicine, a scoring model that can accurately predict the risk of drug combination interactions is trained. It integrates multiple technologies such as knowledge graphs, machine learning, and natural language processing, providing a new paradigm for clinical drug management and control. Through a data-driven approach, drug combination risks are automatically mined, improving the efficiency and accuracy of early warning. BRIEF DESCRIPTION OF THE DRAWINGS

[0060] Figure 1 This is a flow chart of a method for managing and controlling multiple medications based on multi-database backup and synchronization in Example 1 of the present invention;

[0061] Figure 2 This is a module connection diagram of a multiple medication management and control system based on multi-database backup and synchronization in Example 2 of the present invention;

[0062] Figure 3 This is a schematic diagram of the structure of an electronic device in Example 3 of the present invention;

[0063] Figure 4 This is a schematic diagram of the computer-readable storage medium structure in Example 4 of the present invention. DETAILED DESCRIPTION

[0064] The technical solutions of the present invention will be clearly and completely described below in conjunction with the embodiments. Obviously, the embodiments described are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0065] Example 1

[0066] like Figure 1 As shown, a method for managing and controlling multiple medications based on multi-database backup synchronization includes the following steps:

[0067] Step 1: Collect drug clinical data and convert the drug clinical data into a drug knowledge graph;

[0068] Step 2: Use a distributed consistency algorithm to save the drug knowledge graph in the primary database and the backup database;

[0069] Step 3: Collect the complete prescription of each patient; for each group of drugs in the complete prescription, retrieve the interaction path from the drug knowledge graph. If an interaction path exists, execute step 4; otherwise, do nothing;

[0070] Step 4: Determine whether there is a pharmacological conflict for each group of drugs based on the interaction pathway. If there is a pharmacological conflict, proceed to step 5; otherwise, no action is taken.

[0071] Step 5: Take each group of drugs with pharmacological conflicts as a conflicting drug combination, and build an action intensity model for the conflicting drug combination based on the drug knowledge graph;

[0072] Step 6: Input the prescription dosage of each drug in the conflicting drug combination into the effect intensity model to obtain the effect intensity score, and issue a control alert to the patient based on the effect intensity score.

[0073] Specifically, the collection of drug clinical data includes the following steps:

[0074] Step 11: Determine the source of clinical data; the source of clinical data includes but is not limited to hospital electronic medical record systems, clinical databases, adverse drug reaction monitoring systems, medical literature databases, and clinical trial databases;

[0075] Step 12: List the raw data types of all required clinical data; the raw data types include but are not limited to: basic patient information (age, gender, genotype, etc.), diagnosis information (disease name, code, severity, etc.), medication information (drug name, dosage, route of administration, usage, etc.), examination and test results (laboratory tests, imaging, etc.), treatment information (surgery, radiotherapy, etc.), adverse reaction records (symptoms, time of occurrence, severity, etc.), and efficacy evaluation (effective / ineffective, degree of relief, etc.);

[0076] Step 13: Obtain structured data and unstructured data from the original data types from the clinical data source to form drug clinical data; the structured data includes but is not limited to medical records and doctor's order data in the EMR system, and coded data such as disease, test, and medication data derived from the clinical database; the unstructured data includes but is not limited to medical records manually written by doctors in the EMR system, full text of drug-related literature in databases such as PubMed, unstructured documents such as drug instructions and research reports, etc.

[0077] In a further embodiment of the present invention, clinical data and non-clinical data may be further pre-processed, such as de-identifying sensitive personal information, protecting patient privacy, marking data sources, ensuring traceability, etc.;

[0078] Furthermore, the conversion of drug clinical data into a drug knowledge graph includes the following steps:

[0079] Step 21: extract drug triplets from the structured data and the unstructured data respectively, and combine all the extracted drug triplets into a drug triplet set;

[0080] Specifically, the methods for extracting drug triple sets from structured data include:

[0081] Parse patient diagnostic data and extract disease entities and their ICD codes;

[0082] Parse patient medication data to extract generic names, dosages, routes of administration, etc.

[0083] Based on the standard ontology library, identify entity relationships such as disease-drug, drug-dose, etc.

[0084] Mining semantic associations such as diagnosis-treatment from clinical pathway data;

[0085] Forming a preliminary drug triplet; the drug triplet may be in the form of {disease node, relationship, drug node}, {drug node, relationship, symptom node};

[0086] Methods for extracting drug triple sets from unstructured data include:

[0087] Perform natural language preprocessing such as word segmentation and part-of-speech tagging on text data;

[0088] Use named entity recognition models to discover entities such as drugs, diseases, and symptoms in texts;

[0089] Discover synonyms and new words based on word vector semantic similarity and expand them into the standard dictionary;

[0090] Use the relation extraction model to identify the relation tuples in the text;

[0091] Aggregate into standardized drug triads;

[0092] Step 22: De-duplicate the drug triplet set;

[0093] Step 23: Pre-construct the domain ontology and the relationship ontology between the domain ontologies; the domain ontologies include but are not limited to drug names, disease names, side effects, symptoms, people, etc.; the relationship ontologies include but are not limited to the therapeutic and indication relationships between drugs and diseases, the improvement or induction relationships between drugs and symptoms, and the interaction relationships between drugs (further divided into synergy, synergy, and conflict); it is understandable that these domain ontologies and relationship ontologies belong to the common sense in the medical field, and therefore can be collected and summarized in advance;

[0094] Step 24: Take the node in each drug triple as an instance node in the drug knowledge graph, and take the relationship in each drug triple as the instance relationship between two nodes in the drug triple to construct the drug knowledge graph;

[0095] Furthermore, the method of using the distributed consistency algorithm to store the drug knowledge graph in the primary database and the backup database is:

[0096] Deploy a distributed graph database cluster using a multi-data center architecture. Specifically, the distributed graph database cluster includes but is not limited to DataStax Studio, Amazon Neptune, etc. It should be noted that in a distributed graph database cluster, each data center contains several data replicas, providing read and write services. The primary data center provides external read and write capabilities, and the backup center provides failover capabilities.

[0097] Configure a rack-aware topology replication strategy to store N replicas on each node of each data center partition; N is the preset number of replicas, preferably set to 3;

[0098] Use a consistency algorithm to manage data consistency. Specifically, the consistency algorithm uses a consensus protocol between multiple replicas to achieve real-time data synchronization and quickly restore to a consistent state in the event of a database failure.

[0099] Connect to any database node in the distributed graph database cluster through the database client driver. Write requests to the drug knowledge graph are forwarded to the coordination node through the driver and then replicated to other replicas. The configured consistency algorithm is used between the replicas to reach eventual consensus.

[0100] Furthermore, you can configure monitoring indicators in the distributed graph database cluster to track cluster health and data distribution, and enable the automatic repair function to periodically repair inconsistent partitions.

[0101] Furthermore, the method of collecting the complete prescription of each patient is:

[0102] Collect all prescriptions issued to each patient during the same medication cycle through the electronic medical record system of each patient to form a complete prescription; the complete prescription includes the name and dosage of each drug provided to the patient;

[0103] The medication cycle refers to the regular medication cycle of the drug, such as one week, because a patient may be prescribed different medications for different conditions within the same medication cycle, and thus take different medications, which may lead to drug reactions;

[0104] Furthermore, for each group of drugs in the complete prescription, the method of retrieving the interaction path from the drug knowledge graph is as follows:

[0105] For each group of drugs in the complete prescription, retrieve the instance node corresponding to each drug name in the drug knowledge graph and use them as the first instance node and the second instance node respectively;

[0106] Extract all associated edges and semantic paths between the first instance node and the second instance node to form a drug action subgraph;

[0107] Search for "interaction" semantic edges in the drug action subgraph and use a depth-first algorithm or a breadth-first algorithm to search for a connecting path between the first instance node and the second instance node. If a connecting path exists, each connecting path is considered an interaction path.

[0108] Furthermore, the method for determining whether an interaction path exists is:

[0109] If the number of interaction paths searched out is greater than 0, it is determined that an interaction path exists;

[0110] If the number of interaction paths searched out is less than or equal to 0, it is determined that there is no interaction path;

[0111] Furthermore, the method of determining whether there is a pharmacological conflict in each group of drugs based on the interaction pathway is:

[0112] For each interaction path, if the relationship semantics corresponding to all edges in the interaction path are positive or neutral, it is judged that there is no pharmacological conflict. If the relationship semantics corresponding to any edge is negative, it is judged that there is a pharmacological conflict.

[0113] Specifically, the positive effect refers to the existence of a synergistic relationship between drugs; the neutral effect refers to the absence of a mutual influence between drugs, that is, a synergistic effect; the negative effect refers to the conflicting effects between drugs;

[0114] Furthermore, the method of treating each group of drugs with pharmacological conflicts as a conflicting drug combination and constructing an action intensity model for the conflicting drug combination based on the drug knowledge graph is as follows:

[0115] Define the node attributes of the instance node corresponding to the drug name; the node attributes include but are not limited to drug name, molecular weight, ATC code, standard dosage range, etc.;

[0116] Define the relationship attributes of each instance relationship; the relationship attributes include but are not limited to the path ID and weight score; the weight score is the credibility, indicating the credibility of the evidence corresponding to the instance relationship;

[0117] Manually select and annotate the expected confidence scores of some paths in the drug knowledge graph as seeds; the expected confidence scores range from 0 to 1, with an increment of 0.1;

[0118] The expected confidence scores for the remaining paths are generated using the path length and drug dosage characteristics through a pre-set scoring rule; for example, if the path length is less than 3 and there is no overdose, the expected confidence score is 0.9; if the length is greater than 5 and the overdose exceeds 50%, the score is 0.2;

[0119] Extract the one-hot encoding of each edge length, edge semantic type number, and instance relationship to form a set of path feature vectors;

[0120] Specifically, the one-hot encoding of the instance relationship is constructed as follows:

[0121] For each instance node, the actual drug dose received by the patient is calculated and divided by the size of the standard dose range interval of the instance node to obtain the dose suitability;

[0122] Divide each instance node into several dose suitability intervals in advance, and obtain the one-hot encoding of the dose suitability of the instance node;

[0123] As an example:

[0124] There is a drug A, whose recommended dosage range is 0.2-0.6g, and the actual dosage is 0.8g. What is the dosage suitability?

[0125] The pre-set dose suitability intervals include [0-0.5), [0.5-1), [1-1.5), [1.5-2), and [2-+∞). When the dose suitability value is 2, the corresponding interval is [2-+∞), and the obtained one-hot encoding is [0, 0, 0, 0, 1]. As a result, the path feature vector contains both the path structure information and the dose interval encoding information of all participating drugs. This discretized interval encoding method can effectively retain the core semantic information of the dose feature (overdose, normal, below recommendation, etc.), while also avoiding the problem of difficult modeling of data distribution caused by directly using continuous values.

[0126] Use the expected confidence score of the seed annotation path as the output and the path feature vector of the seed annotation path as the input to train the logistic regression model, then train it again based on the remaining annotation paths, and save the logistic regression model as the action intensity model;

[0127] Import the trained logistic regression model into the primary and backup databases of the distributed graph database cluster;

[0128] Furthermore, the prescription dosage of each drug in the conflicting drug combination is input into the effect intensity model to obtain the effect intensity score in the following manner:

[0129] Traverse each interaction path of the conflicting drug combination as the target evaluation path;

[0130] Obtain the path feature vector of the target evaluation path as the target path feature vector;

[0131] Input the target path feature vector into the action intensity model to obtain the expected confidence score output by the action intensity model;

[0132] A preset score threshold is set. If the expected confidence score output by any target evaluation path is greater than the preset score threshold, a control alarm is issued to the patient.

[0133] The control alarm may be to remind the patient that there is a pharmacologically conflicting drug combination among all prescriptions in the patient's current medication cycle, and the patient needs to consult a doctor in a timely manner;

[0134] In a further embodiment of the present invention, after calculating the expected confidence score of each target path feature vector of the conflicting drug combination for each patient, the dosage combination of the two drugs and the expected confidence score can be used as attributes of the path and saved in any database node of the distributed graph database cluster through a consistency algorithm.

[0135] Example 2

[0136] like Figure 2 As shown, a multiple medication control system based on multi-database backup and synchronization includes a map construction module, a data backup module, a path retrieval module, and a medication control module; wherein each module is electrically connected;

[0137] A graph construction module collects drug clinical data, converts the drug clinical data into a drug knowledge graph, and sends the drug knowledge graph to the data backup module and the path retrieval module;

[0138] The data backup module uses a distributed consistency algorithm to save the drug knowledge graph in the main database and the backup database;

[0139] The path retrieval module collects the complete prescription of each patient; for each group of drugs in the complete prescription, the interaction path is retrieved from the drug knowledge graph. If an interaction path exists, it is sent to the drug control module;

[0140] The drug control module determines whether there is a pharmacological conflict in each group of drugs based on the interaction path. If there is a pharmacological conflict, each group of drugs with pharmacological conflicts will be regarded as a conflicting drug combination. An effect intensity model will be constructed for the conflicting drug combination based on the drug clinical data. The prescription dose of each drug in the conflicting drug combination will be input into the effect intensity model to obtain the effect intensity score. Based on the effect intensity score, a control strategy will be provided to the patient.

[0141] Example 3

[0142] Figure 3 This is a schematic diagram of the electronic device structure provided by an embodiment of the present application. Figure 3 As shown, the present application also provides an electronic device 100. The electronic device 100 may include one or more processors and one or more memories. The memories may store computer-readable code, which, when executed by the one or more processors, may execute the above-described method for managing and controlling multiple medications based on multi-database backup and synchronization.

[0143] The method or device according to the embodiment of the present application can also be used by Figure 3 The electronic device architecture shown in FIG. Figure 3As shown, the electronic device 100 may include a bus 101, one or more CPUs 102, a ROM 103, a RAM 104, a communication port 105 connected to a network, an input / output component 106, a hard disk 107, etc. The storage device in the electronic device 100, such as the ROM 103 or the hard disk 107, may store a method for managing and controlling multiple medications based on multi-database backup and synchronization provided in the present application.

[0144] Furthermore, the electronic device 100 may further include a user interface 108. Figure 3 The architecture shown is only exemplary and can be omitted according to actual needs when implementing different devices. Figure 3 One or more components of an electronic device are shown.

[0145] Example 4

[0146] Figure 4 This is a schematic diagram of the computer-readable storage medium structure provided by an embodiment of the present application. Figure 4 As shown, the present application also provides a computer-readable storage medium 200. Computer-readable instructions are stored on the computer-readable storage medium 200. When the computer-readable instructions are executed by the processor, a multiple medication control method based on multi-database backup synchronization according to an embodiment of the present application described with reference to the above figures can be executed. The computer-readable storage medium 200 includes, but is not limited to, for example, volatile memory and / or non-volatile memory. Volatile memory may include, for example, random access memory (RAM) and cache memory (cache). Non-volatile memory may include, for example, read-only memory (ROM), hard disk, flash memory, etc.

[0147] In addition, according to embodiments of the present application, the processes described above with reference to the flowcharts may be implemented as computer software programs. For example, the present application provides a non-transitory machine-readable storage medium storing machine-readable instructions capable of being executed by a processor to execute instructions corresponding to the steps of the method provided in the present application. When the computer program is executed by a central processing unit (CPU), the above-mentioned functions defined in the method of the present application are performed.

[0148] The methods, apparatuses, and devices of the present application may be implemented in many ways. For example, the methods, apparatuses, and devices of the present application may be implemented by software, hardware, firmware, or any combination of software, hardware, and firmware. The above order of steps for the method is for illustration only, and the steps of the method of the present application are not limited to the order specifically described above unless otherwise specified. In addition, in some embodiments, the present application may also be implemented as programs recorded in a recording medium, which include machine-readable instructions for implementing the methods according to the present application. Therefore, the present application also covers recording media that store programs for executing the methods according to the present application.

[0149] In addition, the parts of the above technical solutions provided in the embodiments of the present application that are consistent with the implementation principles of the corresponding technical solutions in the prior art are not described in detail to avoid excessive redundancy.

[0150] The above-described specific embodiments further illustrate the objectives, technical solutions, and beneficial effects of the present invention. It should be understood that the above description is merely a specific embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention shall be included within the scope of protection of the present invention.

[0151] The above preset parameters or preset thresholds are all set by those skilled in the art according to actual conditions or obtained through large amounts of data simulation.

[0152] The above embodiments are only used to illustrate the technical method of the present invention and are not intended to limit the present invention. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical method of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical method of the present invention.

Claims

1. A method for controlling multiple medications based on multi-database backup synchronization, characterized in that: The following steps are involved: Step 1: Collect drug clinical data and convert the drug clinical data into a drug knowledge graph; Step 2: Use a distributed consistency algorithm to save the drug knowledge graph in the primary database and the backup database; Step 3: Collect each patient's complete prescription; For each group of drugs in the complete prescription, the interaction path is retrieved from the drug knowledge graph. If an interaction path exists, step 4 is executed; otherwise, no action is taken. Step 4: Determine whether there is a pharmacological conflict for each group of drugs based on the interaction pathway. If there is a pharmacological conflict, proceed to step 5; otherwise, no action is taken. Step 5: Take each group of drugs with pharmacological conflicts as a conflicting drug combination, and build an action intensity model for the conflicting drug combination based on the drug knowledge graph; Step 6: Input the prescribed dosage of each drug in the conflicting drug combination into the effect intensity model to obtain an effect intensity score, and issue a control alert to the patient based on the effect intensity score; The method of treating each group of drugs with pharmacological conflicts as a conflicting drug combination and constructing an action intensity model for the conflicting drug combination based on the drug knowledge graph is as follows: Define the node attributes of the instance node corresponding to the drug name; Define the relationship attributes of each instance relationship; Manually select and annotate the expected confidence scores of some paths in the drug knowledge graph as seeds; Generate the expected confidence scores for the remaining paths using path length and drug dosage characteristics through pre-set scoring rules; Extract the one-hot encoding of each edge length, edge semantic type number, and instance relationship to form a set of path feature vectors; Use the expected confidence score of the seed annotation path as the output and the path feature vector of the seed annotation path as the input to train the logistic regression model, then train it again based on the remaining annotation paths, and save the logistic regression model as the action intensity model; Import the trained logistic regression model into the primary and backup databases of the distributed graph database cluster.

2. A multiple medication management and control method based on multi-database backup synchronization according to claim 1, characterized in that: The collection of drug clinical data comprises the following steps: Step 11: Identify clinical data sources; Step 12: List the raw data types of all required clinical data; Step 13: Obtain structured data and unstructured data in the original data type from the clinical data source to form drug clinical data.

3. A method for controlling multiple medications based on multi-database backup synchronization according to claim 2, characterized in that: The conversion of drug clinical data into a drug knowledge graph comprises the following steps: Step 21: extract drug triplets from the structured data and the unstructured data respectively, and combine all the extracted drug triplets into a drug triplet set; Step 22: De-duplicate the drug triplet set; Step 23: Pre-construct the domain ontology and the relationship ontology between domain ontologies; Step 24: Take the node in each drug triple as an instance node in the drug knowledge graph, and take the relationship in each drug triple as the instance relationship between two nodes in the drug triple to construct the drug knowledge graph.

4. A method for controlling multiple medications based on multi-database backup synchronization according to claim 3, characterized in that: The method of using the distributed consistency algorithm to save the drug knowledge graph in the main database and the backup database is: Deploy a distributed graph database cluster in a multi-data center architecture; Configure a rack-aware topology replication strategy to store N replicas on each node of each data center partition; N is the preset number of replicas; Use consistency algorithms to manage data consistency; Connect to any database node in the distributed graph database cluster through the database client driver. Write requests for the drug knowledge graph are forwarded to the coordination node through the driver and then copied to other replicas. The configured consistency algorithm is used between the replicas to reach final consensus.

5. A method for controlling multiple medications based on multi-database backup synchronization according to claim 4, characterized in that: For each group of drugs in the complete prescription, the method of retrieving the interaction path from the drug knowledge graph is as follows: For each group of drugs in the complete prescription, retrieve the instance node corresponding to each drug name in the drug knowledge graph and use them as the first instance node and the second instance node respectively; Extract all associated edges and semantic paths between the first instance node and the second instance node to form a drug action subgraph; Search for interaction semantic edges in the drug action subgraph, and use the depth-first algorithm or breadth-first algorithm to search whether there is a connecting path between the first instance node and the second instance node. If there is a connecting path, each connecting path is regarded as an interaction path.

6. A method for controlling multiple medications based on multi-database backup synchronization according to claim 5, characterized in that: The prescription dosage of each drug in the conflicting drug combination is input into the effect intensity model to obtain the effect intensity score as follows: Traverse each interaction path of the conflicting drug combination as the target evaluation path; Obtain the path feature vector of the target evaluation path as the target path feature vector; Input the target path feature vector into the action intensity model to obtain the expected confidence score output by the action intensity model; A preset score threshold is set. If the expected confidence score output by any target evaluation path is greater than the preset score threshold, a control alarm will be issued to the patient.

7. A multiple medication management and control system based on multi-database backup synchronization, which is used to implement a multiple medication management and control method based on multi-database backup synchronization according to any one of claims 1 to 6, characterized in that: It includes a graph construction module, a data backup module, a path retrieval module, and a drug control module; wherein each module is electrically connected; A graph construction module collects drug clinical data, converts the drug clinical data into a drug knowledge graph, and sends the drug knowledge graph to the data backup module and the path retrieval module; The data backup module uses a distributed consistency algorithm to save the drug knowledge graph in the main database and the backup database; The path retrieval module collects the complete prescription of each patient; for each group of drugs in the complete prescription, the interaction path is retrieved from the drug knowledge graph. If an interaction path exists, it is sent to the drug control module; The drug control module determines whether there is a pharmacological conflict in each group of drugs based on the interaction path. If there is a pharmacological conflict, each group of drugs with pharmacological conflicts will be regarded as a conflicting drug combination. An effect intensity model will be constructed for the conflicting drug combination based on the drug clinical data. The prescription dose of each drug in the conflicting drug combination will be input into the effect intensity model to obtain the effect intensity score. Based on the effect intensity score, a control strategy will be provided to the patient.

8. An electronic device, characterized in that: include: A processor and a memory, wherein: The memory stores a computer program that can be called by the processor; The processor executes the multiple medication management method based on multi-database backup synchronization according to any one of claims 1 to 6 in the background by calling the computer program stored in the memory.

9. A computer-readable storage medium, characterized in that A rewritable computer program is stored thereon; When the computer program is run on a computer device, the computer device executes the multiple medication management and control method based on multi-database backup synchronization according to any one of claims 1 to 6 in the background.

Images