Medication education push method and system based on actual needs of patients

Abstract

The present invention discloses a method and system for pushing medication education based on actual needs of patients, the method comprising: constructing a medication education knowledge base; associating the HIS system and the LIS system to obtain information about patients and their medicines; analyzing the information about patients and their medicines to obtain the actual medication education needs of patients; according to the actual medication education needs of patients, searching for the corresponding medication education information in the medication education knowledge base and accurately pushing it. By using the present invention, the instructions for drugs can be made easy to understand while accurately identifying the actual medication needs of patients and pushing the corresponding medication education. As a method and system for pushing medication education based on actual needs of patients, the present invention can be widely used in the field of medical technology.

Description

Technical Field

[0001] The present invention relates to the field of medical technology, and in particular to a method and system for pushing medication education based on actual needs of patients. Background Art

[0002] As medical technology matures, there is an increasing need for automated analysis and processing of medical data in medical teaching and the construction of medical electronic consultation platforms to meet different analysis needs. However, the current medical treatment system mainly relies on oral medication education from pharmacists or blindly pushes large amounts of information to patients, requiring patients to screen the information themselves. Information is not pushed based on their actual needs. Existing drug instructions are too professional for patients to understand, and pharmacists currently mainly explain medication to patients based on their personal understanding of drug knowledge. There is no uniform standard for the content of the instructions, especially for young pharmacists who lack work experience. They may miss key points or even make mistakes, causing safety hazards. Summary of the invention

[0003] In order to solve the above technical problems, the purpose of the present invention is to provide a method and system for pushing medication education based on the actual needs of patients, which can make the instructions of the drug easy to understand while accurately identifying the actual medication needs of patients and pushing corresponding medication education.

[0004] The first technical solution adopted by the present invention is: a method for pushing medication education based on actual needs of patients, comprising the following steps:

[0005] Build a knowledge base for medication education;

[0006] Link the HIS system and LIS system to obtain information about patients and their medicines;

[0007] Analyze the patient and drug information to obtain the patient's actual medication education needs;

[0008] According to the patient's actual medication education needs, the corresponding medication education information in the medication education knowledge base is searched and pushed.

[0009] Furthermore, the step of constructing a medication education knowledge base specifically includes:

[0010] Obtain drug information and corresponding pharmacist medication education;

[0011] Standardize the medication education for pharmacists and obtain the medication education rules;

[0012] Build a knowledge graph based on drug information and medication education rules;

[0013] Based on the knowledge graph, a medication education knowledge base is generated.

[0014] Furthermore, the step of constructing a knowledge graph based on drug information and medication education rules specifically includes:

[0015] Determine the corresponding medication education rules according to the field to which the drug information belongs;

[0016] Perform knowledge fusion processing according to the corresponding medication education rules to obtain candidate medication education;

[0017] The candidate medication education is fused through the knowledge fusion method to obtain structured medication education data;

[0018] Conduct knowledge reasoning on structured medication education data and construct a knowledge graph corresponding to the field of drug information.

[0019] Furthermore, the step of associating the HIS system with the LIS system to obtain information about the patient and his / her medicines specifically includes:

[0020] Obtain patient information from the HIS system and LIS system, including basic personal information, disease information, test information, doctor's advice information, and medication information;

[0021] Compare the semantic and grammatical similarity of the patient information obtained from the HIS system and the patient information obtained from the LIS system, obtain the similarity comparison value and make a judgment;

[0022] If the similarity comparison value is greater than or equal to the preset threshold, the patient information of the HIS system and the LIS system are associated through the metric learning method, and the patient and drug information is output.

[0023] Furthermore, the step of analyzing the patient and drug information to obtain the patient's actual medication education needs specifically includes:

[0024] Perform key point detection operations on patient and drug information to obtain key point location information;

[0025] Based on the key point location information, the patient and drug information is intercepted and processed to obtain the patient prescription key information image;

[0026] The optical character recognition technology is used to perform text recognition and extraction on the patient prescription key information image to obtain the initial patient prescription key information;

[0027] Based on the pre-trained semantic recognition model, semantic recognition and extraction are performed on the initial patient prescription key information to obtain the patient prescription key information, which is the patient's actual medication education needs.

[0028] Furthermore, before the step of performing semantic recognition extraction on the initial patient prescription key information based on the pre-trained semantic recognition model to obtain the patient prescription key information, the patient prescription key information is the patient's actual medication education needs, the following steps are also included:

[0029] Obtain multiple initial patient prescription key information;

[0030] Constructing a semantic recognition model, wherein the semantic recognition model comprises an encoding layer, a self-attention layer, a recognition classification layer, and an output layer;

[0031] Based on the coding layer, the key information of the initial patient prescription is encoded to obtain a semantic vector of the key information of the initial patient prescription;

[0032] Determine the weight vector of the initial patient prescription key information semantic vector based on the self-attention layer;

[0033] The weight vector of the initial patient prescription key information semantic vector is weighted according to the recognition and classification layer to obtain a trained semantic recognition model.

[0034] Furthermore, the step of searching for corresponding medication education information in the medication education knowledge base and accurately pushing the information based on the patient's actual medication education needs specifically includes:

[0035] Extract the name of the link call based on the patient's actual medication education needs;

[0036] Search the medication education knowledge base according to the name of the link call and output the corresponding medication education information;

[0037] The medication education information is packaged and combined with the patient's actual medication education needs and pushed.

[0038] The second technical solution adopted by the present invention is: a medication education push system based on the actual needs of patients, including:

[0039] A building module for constructing a medication education knowledge base;

[0040] The acquisition module is used to link the HIS system and the LIS system to obtain the patient and drug information;

[0041] The analysis module is used to analyze the patient and drug information to obtain the patient's actual medication education needs;

[0042] The push module is used to find the corresponding medication education information in the medication education knowledge base and push it accurately according to the patient's actual medication education needs.

[0043] The beneficial effects of the method and system of the present invention are as follows: the present invention helps patients to quickly and accurately extract medication education information by standardizing the professional terms used by professional pharmacists to interpret instructions and constructing a standard medication education knowledge base, and further retrieves patient and drug information by associating the HIS system and the LIS system, and analyzes the patient information based on the medication prescription information, dividing the patient information into eight categories and twenty subcategories of individualized characteristics of patients, such as gender, age, liver and kidney function, and physiological period, so as to accurately identify the patient's actual medication needs and realize accurate push of medication education information. BRIEF DESCRIPTION OF THE DRAWINGS

[0044] Figure 1 It is a flowchart of the steps of the medication education push method based on the actual needs of patients of the present invention;

[0045] Figure 2 It is a structural block diagram of the medication education push system based on the actual needs of patients of the present invention;

[0046] Figure 3 It is a schematic diagram of the result of outputting the medication education knowledge base of the present invention;

[0047] Figure 4 is a schematic diagram of the classification standard of the present invention based on the actual individual characteristics of the patient;

[0048] Figure 5 is a schematic diagram of classification results obtained by the semantic recognition process of the present invention;

[0049] Figure 6 is a schematic diagram of the result of outputting patient and drug information of the present invention;

[0050] Figure 7 It is a schematic diagram of the result of the individualized medication education explanation ultimately output by the present invention. DETAILED DESCRIPTION

[0051] The present invention is further described in detail below in conjunction with the accompanying drawings and specific embodiments. The step numbers in the following embodiments are only provided for the convenience of explanation and description, and the order between the steps is not limited in any way. The execution order of each step in the embodiment can be adaptively adjusted according to the understanding of those skilled in the art.

[0052] Reference Figure 1 The present invention provides a method for pushing medication education based on actual needs of patients, the method comprising the following steps:

[0053] S1. Build a knowledge base for medication education;

[0054] S11. Obtain drug information and corresponding pharmacist medication education;

[0055] S12. Standardize the medication education for pharmacists and obtain medication education rules;

[0056] Specifically, refer to Figure 3 The drug information includes medication time, medication requirements, treatment methods for missed medications, interval usage of multiple drugs, common or serious adverse drug reactions, patient precautions for driving or high-risk work, precautions for elderly patients, precautions for pregnant women, precautions for breastfeeding, precautions for children, precautions for patients with liver dysfunction, precautions for patients with renal dysfunction, storage methods, and lifestyle and dietary recommendations, etc. Experienced professional pharmacists interpret the instructions for the drugs and convert them into language that ordinary people can understand for output. For example, the pharmacist interprets the adverse reaction "ataxia" as "motor coordination disorders, such as unstable gait, and be careful to prevent falls", etc.

[0057] S13. Construct a knowledge graph based on drug information and medication education rules;

[0058] S131, determining the corresponding medication education rules according to the field to which the drug information belongs;

[0059] S132, performing knowledge fusion processing according to corresponding medication education rules to obtain candidate medication education;

[0060] S133, fusing candidate medication education data through a knowledge fusion method to obtain structured medication education data;

[0061] S134. Perform knowledge reasoning on structured medication education data and construct a knowledge graph corresponding to the field of drug information.

[0062] S14. Generate a medication education knowledge base based on the knowledge graph.

[0063] Specifically, the construction of knowledge graph is the premise for the application and development of knowledge graph, which involves entity extraction and the establishment of entities and relationships between entities. At the same time, it is also necessary to organize and store the extracted entities and relationships well so that they can be quickly accessed and operated. The knowledge graph construction process can usually be divided into two steps: knowledge graph ontology layer construction and entity layer learning. The ontology layer construction usually includes term extraction, synonym extraction, concept extraction, classification relationship extraction, axiom and rule learning; entity layer learning includes entity learning, entity data filling, entity alignment and entity linking, etc.; and in terms of the construction of knowledge graph for drug information, that is, first build different ontology libraries for different data sources and data types, and then map each ontology library into a global ontology library, and then perform entity alignment and entity linking on the global ontology library of each source as a knowledge base, so as to enrich, improve and expand the constructed multi-data fusion drug information knowledge graph; the knowledge graph construction method usually consists of two methods: top-down and bottom-up; the top-down method refers to first building a knowledge graph The bottom-up approach starts from the entity layer, uses certain technical means to summarize and organize entities, align entities, and link entities, and extracts new patterns with high confidence. After manual review, structured medication education data is generated and added to the knowledge graph. Due to the professionalism of knowledge in the medical field, it is not appropriate to automatically construct entities and relationships through big data analysis and clustering. Therefore, a top-down construction model will be selected here. According to the guidance of experts in the medical field, the ontology and association patterns in the drug knowledge base are extracted to construct a drug information knowledge graph. The field is to search in a smaller range according to the field of drugs, that is, if the patient has a cold and fever, the corresponding standard text description of the drug will be searched in the range of cold medicines. If the patient has stomach pain symptoms, the corresponding standard text description of the drug will be searched in the field of stomach medicines.

[0064] S2, link the HIS system and LIS system to obtain information about patients and their medicines;

[0065] Specifically, the medication education system is linked to the HIS system and the LIS system. The individualized patient information in the HIS, LIS, and medication education push systems is linked with the patient's unique ID, and the drug information in the patient's prescription is linked with the medication education knowledge base with the drug's unique ID; the medication education push system is connected to the HIS system to read the simplified medication instructions of the HIS system; the patient's medication prescription is linked with the drug ID, and the corresponding drug education rules are extracted from the medication education knowledge base; the patient's HIS and LIS related data are extracted to determine the patient's personalized medication education precise push content.

[0066] S21, obtaining patient information from the HIS system and the LIS system, wherein the patient information includes basic personal information, disease information, test information, doctor's advice information, and medication information;

[0067] S22, presetting a similarity threshold for determining the patient information associated with the HIS system and the LIS system;

[0068] S23, performing semantic and grammatical similarity comparison based on the patient information obtained from the HIS system and the LIS system, obtaining a similarity comparison value and making a judgment;

[0069] Specifically, the patient and drug information includes basic personal information, disease information, doctor's order information and medication information. It is set that there are first associated data and second associated data. The first associated data comes from the HIS system (hospital information system), and the second associated data comes from the LIS system (laboratory information management system). There are attributes in the first associated data and the second associated data: patient age, patient gender, and patient ID. When the first associated data and the second associated data are compared, the patient age and patient gender are successfully matched, and the patient ID is the same, so the item can be successfully matched. Therefore, the associated attribute of the first associated data and the second associated data is the patient ID, which is further preset for judging The similarity threshold of the patient information with the associated HIS system and LIS system can be set manually, and 80% is generally selected. When the first associated data and the second associated data are mutually associated in terms of patient ID, it can be further determined whether other data information is also correspondingly associated, and the degree of association is calculated. If the degree of association is greater than or equal to the preset threshold, it is determined that the acquired system data is for the same patient. This can not only improve the accuracy of data association but also avoid the problem of association failure due to untimely synchronization of patient information in the two systems. If the patient ID in the second associated data is different from that in the first associated data, the first associated data and the second associated data have no associated attributes; and there is no connection between the first associated data and the second associated data.

[0070] S24. If it is determined that the similarity comparison value is greater than or equal to the similarity threshold, the patient information of the HIS system and the LIS system are associated through a metric learning method, and the patient and drug information is output.

[0071] Specifically, by using the model obtained by metric learning to screen the data to be associated with associated attributes, the similarity between the data to be associated to be compared can be determined based on the semantics and syntax of the attributes of the data to be associated, and a second judgment can be made on the data to be associated, thereby obtaining a more accurate comparison result. The data to be associated that meet the requirements of the model are then determined as the data to be associated with the same attributes. By comparing the similarities, a rapid screening process can be achieved, which is beneficial to improving the processing efficiency of data association. Screening through the model can then improve the accuracy of the data association relationship.

[0072] S3. Analyze the patient and drug information to obtain the patient's actual medication needs;

[0073] Specifically, the key point detection operation can be implemented using the existing deep neural network (such as SSD, YOLO), and the patient's medication prescription can be detected and intercepted from the image; and the key area corresponding to the keyword is predicted on the intercepted patient medication prescription image. Here, the detection method based on CenterNet is used to locate the center point of the keyword area such as "name", "gender", "age", "precautions", "dietary advice", "drug instructions", and "adverse reaction instructions";

[0074] S31, performing key point detection operation processing on the patient and drug information to obtain key point position information;

[0075] S32, intercepting and processing the patient and drug information based on the key point position information to obtain a patient prescription key information image;

[0076] S33, performing text recognition and extraction on the patient prescription key information image by optical character recognition technology to obtain initial patient prescription key information;

[0077] Specifically, the interception operation and the correction operation can be to customize the standard placement position of the patient's medication prescription after predicting the key points, and then obtain the one-to-one corresponding key points, and calculate the radiation matrix according to the corresponding relationship of the key points, and then multiply the intercepted patient medication prescription image by the radiation matrix to obtain the corrected standard image, that is, filter irrelevant information, such as the prescribing doctor information, etc., and intercept to obtain the above key information; the text recognition operation refers to the existing general optical character recognition technology, and the technology based on it can be OCR text recognition software, so as to realize the software that directly converts the text content on the picture and photo into editable text;

[0078] S34. Perform semantic recognition and extraction on the initial patient prescription key information based on the pre-trained semantic recognition model to obtain the patient prescription key information, which is the patient's actual medication education needs.

[0079] Reference Figure 4 and Figure 5 The specific classification results are as follows:

[0080] Ordinary patients: Prerequisites: male or non-pregnant if pregnancy is indicated and non-lactating if breastfeeding is indicated, 14 years old < age < 60 years old, liver and kidney function problems are not indicated in the test index or diagnosis. Medication education push content: concise version of medication instructions, medication time, medication requirements, treatment methods for missed medications, interval use methods for multiple medications, common or serious adverse drug reactions, patient precautions such as driving or high-risk work, storage methods, and lifestyle and dietary recommendations;

[0081] Patients with liver and kidney dysfunction: Prerequisites: male or non-pregnant if pregnancy is indicated and non-lactating if breastfeeding is indicated, 14 years old < age < 60 years old, liver and kidney function problems are indicated in the test index or diagnosis. Medication education push content: concise version of medication instructions, medication time, medication requirements, treatment methods for missed medications, interval use methods for multiple medications, common or serious adverse drug reactions, precautions for patients such as driving or high-risk work, precautions for patients with abnormal liver function taking this medicine, precautions for patients with abnormal kidney function taking this medicine, storage methods, and lifestyle and dietary recommendations;

[0082] Patients with renal insufficiency: Prerequisites: male or non-pregnant if pregnancy is indicated and non-lactating if breastfeeding is indicated, 14 years old < age < 60 years old, renal function problems are indicated in the test index or diagnosis, and medication education push content: concise version of medication instructions, medication time, medication requirements, treatment methods for missed medications, interval use methods for multiple medications, common or serious adverse drug reactions, patient precautions such as driving or high-risk work, precautions for patients with abnormal renal function taking this medicine, storage methods, and lifestyle and dietary recommendations;

[0083] Patients with liver dysfunction: Prerequisites: male or non-pregnant if pregnancy is indicated and non-lactating if breastfeeding is indicated, 14 years old < age < 60 years old, liver function problems are indicated in the test index or diagnosis, medication education push content: concise version of medication instructions, medication time, medication requirements, treatment methods for missed medications, interval use methods for multiple medications, common or serious adverse drug reactions, patient precautions such as driving or high-risk work, precautions for patients with abnormal liver function taking this medicine, storage methods, and lifestyle and diet;

[0084] Pediatric patients: Prerequisites: Children: age less than or equal to 14 years old, no liver or kidney function problems are indicated in the test index or diagnosis. Medication education push content: concise version of medication instructions, medication time, medication requirements, treatment methods for missed medications, interval use methods for multiple medications, common or serious adverse drug reactions, storage, lifestyle and diet recommendations, and precautions for pediatric patients taking this medicine;

[0085] Children with impaired liver and kidney function: Prerequisites: Children: age less than or equal to 14 years old, with liver and kidney function problems indicated in the test index or diagnosis. Medication education push content: concise version of medication instructions, medication time, medication requirements, treatment methods for missed medications, interval use methods for multiple medications, common or serious adverse drug reactions, storage methods, lifestyle and dietary recommendations, precautions for children taking this medicine, precautions for patients with impaired liver function taking this medicine, precautions for patients with impaired kidney function taking this medicine;

[0086] Children with renal insufficiency: Prerequisites: Children: age less than or equal to 14 years old, with renal function problems indicated in the test index or diagnosis. Medication education push content: concise version of medication instructions, medication time, medication requirements, treatment methods for missed medications, interval use methods for multiple medications, common or serious adverse drug reactions, storage methods, lifestyle and dietary recommendations, precautions for children taking this medicine, precautions for patients with renal insufficiency taking this medicine;

[0087] Children with liver dysfunction: Prerequisites: Children: age less than or equal to 14 years old, with liver function problems indicated in the test index or diagnosis. Medication education push content: concise version of medication instructions, medication time, medication requirements, treatment methods for missed medications, interval use methods for multiple medications, common or serious adverse drug reactions, storage methods, lifestyle and dietary recommendations, precautions for children taking this medicine, precautions for patients with liver dysfunction taking this medicine;

[0088] Elderly patients: Prerequisites: age greater than or equal to 60 years old, no liver or kidney function problems are indicated in the test index or diagnosis. Medication education push content: concise version of medication instructions, medication time, medication requirements, treatment methods for missed medications, interval use methods for multiple medications, common or serious adverse drug reactions, patient precautions such as driving or high-risk work, storage methods, lifestyle and dietary recommendations, and precautions for elderly patients taking this medicine;

[0089] Elderly patients with impaired liver and kidney function: Prerequisites: age greater than or equal to 60 years old, liver and kidney function problems are indicated in the test index or diagnosis. Medication education push content: concise version of medication instructions, medication time, medication requirements, methods for handling missed medications, methods for using multiple medications at intervals, common or serious adverse drug reactions, patient precautions such as driving or high-risk work, storage methods, lifestyle and dietary recommendations, precautions for elderly patients taking this medicine, precautions for patients with impaired liver function taking this medicine, precautions for patients with impaired kidney function taking this medicine;

[0090] Elderly patients with renal insufficiency: Prerequisites: age greater than or equal to 60 years old, with renal function problems indicated in the test index or diagnosis. Medication education push content: concise version of medication instructions, medication time, medication requirements, treatment methods for missed medications, interval use methods for multiple medications, common or serious adverse drug reactions, patient precautions such as driving or high-risk work, storage methods, lifestyle and dietary recommendations, precautions for elderly patients taking this medicine, precautions for patients with renal insufficiency taking this medicine;

[0091] Elderly patients with impaired liver function: Prerequisites: age greater than or equal to 60 years old, liver function problems indicated in the test index or diagnosis. Medication education push content: concise version of medication instructions, medication time, medication requirements, treatment methods for missed medications, interval use methods for multiple medications, common or serious adverse drug reactions, patient precautions such as driving or high-risk work, storage methods, lifestyle and diet recommendations, precautions for elderly patients taking this medicine, precautions for patients with impaired liver function taking this medicine;

[0092] Pregnant patients: Prerequisites: Female, pregnancy marked as pregnancy, no liver or kidney function problems indicated in the test index or diagnosis. Medication education push content: concise version of medication instructions, medication time, medication requirements, treatment methods for missed medications, interval use methods for multiple medications, common or serious adverse drug reactions, patient precautions such as driving or high-risk work, storage methods, lifestyle and dietary recommendations, and precautions for pregnant women taking this medicine;

[0093] Pregnant patients with impaired liver and kidney function: Prerequisites: Female, pregnancy marked as pregnancy, liver and kidney function problems indicated in the test index or diagnosis. Medication education push content: concise version of medication instructions, medication time, medication requirements, treatment methods for missed medications, interval use methods for multiple medications, common or serious adverse drug reactions, patient precautions such as driving or high-risk work, storage methods, lifestyle and dietary recommendations, precautions for pregnant women taking this medicine, precautions for patients with impaired liver function taking this medicine, precautions for patients with impaired kidney function taking this medicine;

[0094] Pregnant patients with renal insufficiency: Prerequisites: female, pregnancy marked as pregnancy, test index or diagnosis indicating renal function problems. Medication education push content: concise version of medication instructions, medication time, medication requirements, treatment methods for missed medications, interval use of multiple medications, common or serious adverse drug reactions, patient precautions such as driving or high-risk work, storage methods, lifestyle and diet recommendations, precautions for pregnant women taking this medicine, precautions for patients with renal insufficiency taking this medicine;

[0095] Pregnant patients with impaired liver function: Prerequisites: Female, pregnancy marked as pregnant, liver function problems indicated in the test index or diagnosis. Medication education push content: concise version of medication instructions, medication time, medication requirements, treatment methods for missed medications, interval use methods for multiple medications, common or serious adverse drug reactions, patient precautions such as driving or high-risk work, storage methods, lifestyle and dietary recommendations, precautions for pregnant women taking this medicine, precautions for patients with impaired liver function taking this medicine;

[0096] Lactating patients: Prerequisites: Female, breastfeeding is marked as breastfeeding, and liver and kidney function problems are not indicated in the test index or diagnosis. Medication education push content: concise version of medication instructions, medication time, medication requirements, treatment methods for missed medications, interval use methods for multiple medications, common or serious adverse drug reactions, patient precautions such as driving or high-risk work, storage methods, lifestyle and dietary recommendations, precautions for lactating patients taking this medicine;

[0097] Lactating patients with impaired liver and kidney function: Prerequisites: Female, breastfeeding is marked as breastfeeding, liver and kidney function problems are indicated in the test index or diagnosis. Medication education push content: concise version of medication instructions, medication time, medication requirements, treatment methods for missed medications, interval use of multiple medications, common or serious adverse drug reactions, patient precautions such as driving or high-risk work, storage methods, lifestyle and dietary recommendations, precautions for lactating patients taking this medicine, precautions for patients with impaired liver function taking this medicine, precautions for patients with impaired kidney function taking this medicine;

[0098] Lactating patients with renal insufficiency: Prerequisites: female, breastfeeding is marked as breastfeeding, and renal function problems are indicated in the test index or diagnosis. Medication education push content: concise version of medication instructions, medication time, medication requirements, treatment methods for missed medications, interval use methods for multiple medications, common or serious adverse drug reactions, patient precautions such as driving or high-risk work, storage methods, lifestyle and dietary recommendations, precautions for lactating patients taking this medicine, precautions for patients with renal insufficiency taking this medicine;

[0099] Lactating patients with liver dysfunction: Prerequisites: Female, breastfeeding is marked as breastfeeding, liver function problems are indicated in the test index or diagnosis. Medication education push content: concise version of medication instructions, medication time, medication requirements, treatment methods for missed medications, interval use methods for multiple medications, common or serious adverse drug reactions, patient precautions such as driving or high-risk work, storage methods, lifestyle and dietary recommendations, precautions for lactating patients taking this medicine, and precautions for patients with liver dysfunction taking this medicine.

[0100] The training process of the semantic recognition model in step S34 is specifically as follows:

[0101] S341. Obtaining multiple initial patient prescription key information;

[0102] Specifically, the semantic analysis model is a pre-trained deep recognition network model, which can obtain the true meaning of the target vocabulary by analyzing the associated text content. First, the medical record data to be processed can be obtained, wherein the medical record data is electronic data. Optionally, the medical record data can be obtained by converting paper medical records into electronic data through text recognition or other methods. Optionally, the medical record data can also be obtained by exporting the electronic medical record data directly entered into the hospital medical record electronic platform during the doctor's consultation. There is no specific limitation.

[0103] S342, constructing a semantic recognition model, wherein the semantic recognition model comprises an encoding layer, a self-attention layer, a recognition classification layer, and an output layer;

[0104] S343, encoding the initial patient prescription key information based on the encoding layer to obtain a semantic vector of the initial patient prescription key information;

[0105] Specifically, the relationship between medical record data and certain individual characteristics of patients can be analyzed. Specifically, at least two individual characteristics of a patient can be preset to construct a preset individual feature set. For example, the individual characteristics of a patient (gender, age, disease condition, etc.) can be constructed as a preset individual feature set. The BERT model can first convert text data into vector representation and initialize the input vector. Then, the BERT model can encode the input vector through the self-attention mechanism and feedforward network, and output a text semantic vector, that is, the initial patient prescription key information semantic vector;

[0106] S344, determining a weight vector of an initial patient prescription key information semantic vector based on the self-attention layer;

[0107] S345. Perform weight calculation on the weight vector of the initial patient prescription key information semantic vector according to the recognition and classification layer to obtain a trained semantic recognition model.

[0108] Specifically, the essence of the attention mechanism is the mapping of key-value pairs from the query statement to the target statement. By allocating limited attention to different feature vectors, key information that contributes greatly to auxiliary diagnosis can be quickly screened out. Since there are a large number of long sequences in the medical record data, semantic attenuation is prone to occur. The self-attention mechanism can directly calculate the dependency between two characters at any position, breaking the sentence length limit. Therefore, when calculating the attention score, the same self-attention of the query statement and the target statement is considered. The multi-head self-attention mechanism is based on the self-attention mechanism and is the result of multiple self-attention calculations. The BERT model can learn semantic features from different semantic spaces, so that the case semantic vector obtained by fusion of the multi-head self-attention mechanism is input into the softmax classifier. The classification categories set in the softmax classifier are the preset diseases, so that the softmax classifier can output the weight vector corresponding to the medical record data, and each component in the weight vector is the weight of each possibility.

[0109] S4. According to the patient’s actual medication needs, search for the corresponding medication education information in the medication education knowledge base and push it accurately.

[0110] S41. extract the name of the link call according to the patient's actual medication education needs;

[0111] S42, searching in the medication education knowledge base according to the name of the link call, and outputting the corresponding medication education information;

[0112] S43: Packaging and combining the medication education information with the patient's actual medication education needs and pushing them

[0113] When the application calls patient information, the attribute information of the call link, i.e., the corresponding name, is extracted. The call link is the path for finding the patient's prescription. A graphic code or other information is generated based on the attribute information. The graphic code is used to trace the application call. When the application calls user information, the application's call link contains attribute information related to the application, such as patient prescription, patient ID, query time, etc. The system extracts this attribute information and processes it through code or software to process the attribute information into a graphic code. In this embodiment, the graphic code is a QR code, and it can also be a bar code in practice. There is no specific limitation. Further, the graphic code is used to trace the application call. ; When an application calls patient information, these graphic codes carry the attribute information of the application's call link. The patient can understand his or her medication education information by scanning the graphic code. The specific search process extracts the application name from the application call link, and then searches the system for application usage status data corresponding to the application name. The application usage status data includes: the number of personal information calls recently, the corresponding medication education information is the status record when the patient agrees to the query, the patient's name, and the patient's age. The application usage status data is then combined to obtain the authorization application information, that is, the patient's name, ID, and medication education information data are matched in this system and pushed to the patient's receiving end.

[0114] For the above content, refer to Figure 6 and Figure 7 For example, a young man took Zolpidem tablets for insomnia. When he took the medicine, the system pushed medication education content. The medication education content was based on the individual situation of the patient and did not push content that the patient did not need, such as "Precautions for children, precautions for elderly patients, precautions for pregnant women, precautions for breastfeeding patients, precautions for abnormal liver function, and precautions for abnormal kidney function." Because the patient only took one drug, the "method of interval use of multiple drugs" was not pushed.

[0115] Reference Figure 2 , a medication education push system based on the actual needs of patients, including:

[0116] A building module for constructing a medication education knowledge base;

[0117] The acquisition module is used to link the HIS system and the LIS system to obtain the patient and drug information;

[0118] The analysis module is used to analyze the patient and drug information to obtain the patient's actual medication education needs;

[0119] The push module is used to find the corresponding medication education information in the medication education knowledge base and push it accurately according to the patient's actual medication education needs.

[0120] The contents of the above method embodiments are all applicable to the present system embodiments. The functions specifically implemented by the present system embodiments are the same as those of the above method embodiments, and the beneficial effects achieved are also the same as those achieved by the above method embodiments.

[0121] The above is a specific description of the preferred implementation of the present invention, but the invention is not limited to the embodiments. Those skilled in the art may make various equivalent modifications or substitutions without violating the spirit of the present invention. These equivalent modifications or substitutions are all included in the scope defined by the claims of this application.

Claims

1. A medication education push method based on the actual needs of patients, characterized by: The following steps are involved: Build a knowledge base for medication education; Link the HIS system and LIS system to obtain information about patients and their medicines; Analyze the patient and drug information to obtain the patient's actual medication education needs; According to the actual medication education needs of patients, search for the corresponding medication education information in the medication education knowledge base and push it; The step of constructing a medication education knowledge base specifically includes: Obtain drug information and corresponding pharmacist medication education; Standardize the medication education for pharmacists and obtain the medication education rules; Build a knowledge graph based on drug information and medication education rules; Generate a medication education knowledge base based on the knowledge graph; The step of constructing a knowledge graph based on drug information and medication education rules specifically includes: Determine the corresponding medication education rules according to the field to which the drug information belongs; Perform knowledge fusion processing according to the corresponding medication education rules to obtain candidate medication education; The candidate medication education is fused through the knowledge fusion method to obtain structured medication education data; Perform knowledge reasoning on structured medication education data and construct a knowledge graph for the corresponding field of drug information; The step of associating the HIS system with the LIS system to obtain the patient and drug information specifically includes: Obtain patient information from the HIS system and LIS system, including basic personal information, disease information, test information, doctor's advice information, and medication information; Compare the semantic and grammatical similarity of the patient information obtained from the HIS system and the patient information obtained from the LIS system, obtain the similarity comparison value and make a judgment; If the similarity comparison value is greater than or equal to the preset threshold, the patient information of the HIS system and the LIS system are associated through the metric learning method, and the patient and drug information are output; It is assumed that there are first associated data and second associated data, the first associated data comes from the HIS system, and the second associated data comes from the LIS system. When the first associated data and the second associated data are associated with each other by patient ID, it is further determined whether other data information is also correspondingly associated.

2. The method for pushing medication education based on actual needs of patients according to claim 1, characterized in that: The step of analyzing the patient and drug information to obtain the patient's actual medication education needs specifically includes: Perform key point detection operations on patient and drug information to obtain key point location information; Based on the key point location information, the patient and drug information is intercepted and processed to obtain the patient prescription key information image; The optical character recognition technology is used to perform text recognition and extraction on the patient prescription key information image to obtain the initial patient prescription key information; Based on the pre-trained semantic recognition model, semantic recognition and extraction of key information of the initial patient prescription are performed to obtain the patient's actual medication education needs.

3. The method for pushing medication education based on actual needs of patients according to claim 2 is characterized in that: The training steps of the semantic recognition model include the following steps: Obtain multiple initial patient prescription key information; Constructing a semantic recognition model, wherein the semantic recognition model comprises an encoding layer, a self-attention layer, a recognition classification layer, and an output layer; Based on the coding layer, the key information of the initial patient prescription is encoded to obtain a semantic vector of the key information of the initial patient prescription; Determine the weight vector of the initial patient prescription key information semantic vector based on the self-attention layer; The weight vector of the initial patient prescription key information semantic vector is weighted according to the recognition and classification layer to obtain a trained semantic recognition model.

4. The method for pushing medication education based on actual needs of patients according to claim 3 is characterized in that: The step of searching for and pushing corresponding medication education information in the medication education knowledge base according to the actual medication education needs of the patient specifically includes: Extract the name of the link call based on the patient's actual medication education needs; Search the medication education knowledge base according to the name of the link call and output the corresponding medication education information; The medication education information is packaged and combined with the patient's actual medication education needs and pushed.

5. The medication education push system based on the actual needs of patients is characterized by: Used to execute the medication education push method as claimed in claim 1, Includes the following modules: A building block for constructing a medication education knowledge base; The acquisition module is used to link the HIS system and the LIS system to obtain the patient and drug information; The analysis module is used to analyze the patient and drug information to obtain the patient's actual medication education needs; The push module is used to find the corresponding medication education information in the medication education knowledge base and push it accurately according to the patient's actual medication education needs.

Images