CN111667890A - Medication recommendation method, device, equipment and storage medium based on artificial intelligence - Google Patents

Medication recommendation method, device, equipment and storage medium based on artificial intelligence Download PDF

Info

Publication number
CN111667890A
CN111667890A CN202010475961.3A CN202010475961A CN111667890A CN 111667890 A CN111667890 A CN 111667890A CN 202010475961 A CN202010475961 A CN 202010475961A CN 111667890 A CN111667890 A CN 111667890A
Authority
CN
China
Prior art keywords
medication
data
recommendation
recommended
term
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Pending
Application number
CN202010475961.3A
Other languages
Chinese (zh)
Inventor
赵惟
徐卓扬
左磊
孙行智
刘卓
赵婷婷
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Ping An Technology Shenzhen Co Ltd
Original Assignee
Ping An Technology Shenzhen Co Ltd
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Ping An Technology Shenzhen Co Ltd filed Critical Ping An Technology Shenzhen Co Ltd
Priority to CN202010475961.3A priority Critical patent/CN111667890A/en
Priority to PCT/CN2020/099255 priority patent/WO2021114624A1/en
Publication of CN111667890A publication Critical patent/CN111667890A/en
Pending legal-status Critical Current

Links

Images

Classifications

    • GPHYSICS
    • G16INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
    • G16HHEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
    • G16H10/00ICT specially adapted for the handling or processing of patient-related medical or healthcare data
    • G16H10/20ICT specially adapted for the handling or processing of patient-related medical or healthcare data for electronic clinical trials or questionnaires
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06FELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00Pattern recognition
    • G06F18/20Analysing
    • G06F18/24Classification techniques
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06NCOMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00Computing arrangements based on biological models
    • G06N3/02Neural networks
    • GPHYSICS
    • G16INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
    • G16HHEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
    • G16H20/00ICT specially adapted for therapies or health-improving plans, e.g. for handling prescriptions, for steering therapy or for monitoring patient compliance
    • G16H20/10ICT specially adapted for therapies or health-improving plans, e.g. for handling prescriptions, for steering therapy or for monitoring patient compliance relating to drugs or medications, e.g. for ensuring correct administration to patients
    • GPHYSICS
    • G16INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
    • G16HHEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
    • G16H70/00ICT specially adapted for the handling or processing of medical references
    • G16H70/40ICT specially adapted for the handling or processing of medical references relating to drugs, e.g. their side effects or intended usage

Landscapes

  • Engineering & Computer Science (AREA)
  • Health & Medical Sciences (AREA)
  • Theoretical Computer Science (AREA)
  • General Health & Medical Sciences (AREA)
  • Data Mining & Analysis (AREA)
  • Public Health (AREA)
  • Primary Health Care (AREA)
  • Bioinformatics & Cheminformatics (AREA)
  • Medical Informatics (AREA)
  • Epidemiology (AREA)
  • Physics & Mathematics (AREA)
  • Chemical & Material Sciences (AREA)
  • General Physics & Mathematics (AREA)
  • General Engineering & Computer Science (AREA)
  • Evolutionary Computation (AREA)
  • Medicinal Chemistry (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Artificial Intelligence (AREA)
  • Bioinformatics & Computational Biology (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Evolutionary Biology (AREA)
  • Toxicology (AREA)
  • Pharmacology & Pharmacy (AREA)
  • Biomedical Technology (AREA)
  • Biophysics (AREA)
  • Computational Linguistics (AREA)
  • Molecular Biology (AREA)
  • Computing Systems (AREA)
  • Mathematical Physics (AREA)
  • Software Systems (AREA)
  • Medical Treatment And Welfare Office Work (AREA)

Abstract

The invention discloses a medication recommending method, a medication recommending device, medication recommending equipment and a storage medium based on artificial intelligence. The method comprises the following steps: acquiring a medication recommendation request, wherein the medication recommendation request comprises a disease type, current disease data and user portrait data; acquiring a target medication recommendation model according to a target analysis result of the original medication recommendation model corresponding to the disease type; analyzing and processing the current disease data by adopting a target medication recommendation model to obtain original recommended medication; carrying out characteristic analysis on the original recommended medication to obtain a characteristic analysis value corresponding to the characteristic to be analyzed; determining a recommendation tendency type based on the user portrait data, and acquiring a feature weight corresponding to a feature to be analyzed according to the recommendation tendency type; and acquiring a recommended evaluation value of the original recommended medication based on the characteristic analysis value and the characteristic weight corresponding to the characteristic to be analyzed, and determining the original recommended medication with the maximum recommended evaluation value as the target recommended medication. The method can effectively improve the effectiveness and pertinence of intelligent medication recommendation.

Description

Medication recommendation method, device, equipment and storage medium based on artificial intelligence
Technical Field
The invention relates to the technical field of artificial intelligence, in particular to a medication recommending method, a medication recommending device, medication recommending equipment and a storage medium based on artificial intelligence.
Background
A Clinical Decision Support System (CDSS System) generally refers to a computer System that can provide Support for Clinical Decision, and this System fully utilizes available and suitable computer technology, and improves and enhances Decision efficiency by a man-machine interaction manner for semi-structured or unstructured medical problems. The CDSS system is an important means for improving medical quality, i.e., the CDSS system deeply analyzes medical record data to make the most appropriate diagnosis and treatment decision, and the primary purpose of the CDSS system is to evaluate and improve medical quality, reduce medical errors, and control medical expenses.
The medication recommendation models integrated on the current CDSS system are models for recommending medication by analyzing disease data, and are mainly classified into two types, namely a knowledge-driven-based medication recommendation model and a data-driven-based medication recommendation model. The knowledge-driven based medication recommendation model is a medication recommendation model formed based on various medical guidelines and expert consensus. The medication recommendation model based on data driving is a medication recommendation model formed by integrating various machine learning, deep learning and other artificial intelligence technologies. The quality of the current medicine recommendation model is evaluated mainly by judging whether the recommended medicine is matched with the clinically-actually prescribed medicine or not as a standard, and medicine is recommended according to the medicine recommendation model determined by the standard, and due to the fact that the clinically-actually actual levels are different, the curative effect of the clinically-actually prescribed medicine is poor, the recommendation accuracy of the medicine recommendation model is reduced, namely the curative effect effectiveness of the acquired recommended medicine is low, and the popularization of the medicine recommendation model is affected.
Disclosure of Invention
The embodiment of the invention provides a medication recommendation method, a medication recommendation device, medication recommendation equipment and a storage medium based on artificial intelligence, and aims to solve the problem of low recommendation accuracy when a current medication recommendation model carries out medication recommendation.
A medication recommending method based on artificial intelligence comprises the following steps:
acquiring a medication recommendation request, wherein the medication recommendation request comprises a disease type, current disease data and user portrait data;
acquiring at least two target medication recommendation models with better target analysis results according to the target analysis results of the original medication recommendation models corresponding to the disease types;
analyzing and processing the current disease data by adopting at least two target medication recommendation models to obtain at least two original recommended medications;
performing characteristic analysis on the original recommended medication to obtain characteristic analysis values corresponding to at least two characteristics to be analyzed;
determining a recommendation tendency type based on the user portrait data, and acquiring feature weights corresponding to at least two features to be analyzed according to the recommendation tendency type;
and acquiring recommended evaluation values of at least two original recommended medicaments based on the feature analysis values and the feature weights corresponding to at least two features to be analyzed, and determining the original recommended medicament with the maximum recommended evaluation value as a target recommended medicament.
An artificial intelligence-based medication recommendation device, comprising:
the recommendation request acquisition module is used for acquiring a medication recommendation request, wherein the medication recommendation request comprises a disease type, current disease data and user portrait data;
the recommendation model determining module is used for acquiring at least two target medication recommendation models with better target analysis results according to the target analysis results of the original medication recommendation models corresponding to the disease types;
the original recommended medication obtaining module is used for analyzing and processing the current disease data by adopting at least two target medication recommending models to obtain at least two original recommended medications;
the characteristic analysis value acquisition module is used for carrying out characteristic analysis on the original recommended medication to acquire characteristic analysis values corresponding to at least two characteristics to be analyzed;
the characteristic weight acquisition module is used for determining a recommendation tendency type based on the user portrait data and acquiring characteristic weights corresponding to at least two characteristics to be analyzed according to the recommendation tendency type;
and the target recommended medication acquiring module is used for acquiring at least two recommended evaluation values of the original recommended medication based on the feature analysis values and the feature weights corresponding to the at least two features to be analyzed, and determining the original recommended medication with the maximum recommended evaluation value as the target recommended medication.
A computer device comprising a memory, a processor and a computer program stored in the memory and executable on the processor, the processor implementing the artificial intelligence based medication recommendation method when executing the computer program.
A computer-readable storage medium, storing a computer program which, when executed by a processor, implements the artificial intelligence based medication recommendation method described above.
According to the artificial intelligence-based medication recommendation method, the artificial intelligence-based medication recommendation device, the artificial intelligence-based medication recommendation equipment and the storage medium, at least two target medication recommendation models with better target analysis results are obtained according to the target analysis results of the original medication recommendation models corresponding to the disease types, and then the target medication recommendation models are utilized to analyze and process current disease data to obtain original recommended medication, so that effectiveness and pertinence of the original recommended medication can be guaranteed; and then, carrying out feature analysis on the original recommended medication, determining a feature analysis value corresponding to the feature to be analyzed, determining a feature weight corresponding to the feature to be analyzed according to the recommendation tendency type determined by analyzing the user portrait data, and carrying out weighting operation based on the feature analysis value corresponding to the feature to be analyzed and the corresponding feature weight to determine a recommendation evaluation value, so that the recommendation evaluation value can effectively reflect the effectiveness of medication recommendation and the attaching degree with the user portrait data, and the effectiveness and pertinence of intelligent medication recommendation are improved.
Drawings
In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings needed to be used in the description of the embodiments of the present invention will be briefly introduced below, and it is obvious that the drawings in the following description are only some embodiments of the present invention, and it is obvious for those skilled in the art that other drawings can be obtained according to these drawings without inventive labor.
FIG. 1 is a schematic diagram of an application environment of an artificial intelligence-based medication recommendation method according to an embodiment of the present invention;
FIG. 2 is a flowchart of an artificial intelligence based medication recommendation method in accordance with an embodiment of the present invention;
FIG. 3 is another flowchart of an artificial intelligence based medication recommendation method in accordance with an embodiment of the present invention;
FIG. 4 is another flowchart of an artificial intelligence based medication recommendation method in accordance with an embodiment of the present invention;
FIG. 5 is another flowchart of an artificial intelligence based medication recommendation method in accordance with an embodiment of the present invention;
FIG. 6 is another flowchart of an artificial intelligence based medication recommendation method in accordance with an embodiment of the present invention;
FIG. 7 is another flowchart of an artificial intelligence based medication recommendation method in accordance with an embodiment of the present invention;
FIG. 8 is another flowchart of an artificial intelligence based medication recommendation method in accordance with an embodiment of the present invention;
FIG. 9 is a schematic diagram of an artificial intelligence based medication recommendation device in accordance with an embodiment of the present invention;
FIG. 10 is a schematic diagram of a computer device according to an embodiment of the invention.
Detailed Description
The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are some, not all, embodiments of the present invention. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
The artificial intelligence based medication recommendation method provided by the embodiment of the invention can be applied to an application environment shown in fig. 1. Specifically, the medication recommendation method based on artificial intelligence is applied to a CDSS system, the CDSS system comprises a client and a server shown in fig. 1, and the client and the server are communicated through a network and are used for intelligently analyzing current disease data and user portrait data according to a medication recommendation request triggered by the client, so that the obtained target recommended medication is ensured, and the accuracy, efficiency and pertinence of the recommended medication are ensured. The client is also called a user side, and refers to a program corresponding to the server and providing local services for the client. The client may be installed on, but is not limited to, various personal computers, laptops, smartphones, tablets, and portable wearable devices. The server may be implemented as a stand-alone server or as a server cluster consisting of a plurality of servers.
In an embodiment, as shown in fig. 2, an artificial intelligence-based medication recommendation method is provided, which is described by taking the server in fig. 1 as an example, and includes the following steps:
s201: and acquiring a medication recommendation request, wherein the medication recommendation request comprises a disease type, current disease data and user portrait data.
Wherein the medication recommendation request is a request for triggering execution of a medication recommendation service. As an example, a server of the CDSS system may receive a medication recommendation request triggered by a user through a client, and obtain a type of a medical condition, current medical condition data, and user portrait data in the medication recommendation request.
Wherein, the disease type is the type corresponding to the disease of the patient. As an example, the types of conditions include, but are not limited to, diabetes, hypertension, and chronic kidney disease. As can be understood, the determination of the disease type is helpful for determining a more accurate medication recommendation model for medication recommendation, and improves the accuracy of the medication recommendation.
Wherein, the current disease data is data corresponding to the current physical state of the patient. The current disease data refers to data corresponding to a detection index corresponding to a disease type, which is detected by a patient in a medical place in real time. For example, for the type of diabetes, measured data corresponding to indices such as blood glucose, glycated hemoglobin, glycated serum protein, and antibodies need to be collected in real time.
The user profile data is profile data for reflecting patient-related information, and the user profile data may include data affecting the disease diagnosis effect, such as age, sex and other data, and may also include data affecting medication recommendation, for example, when the patient income is low, medication with low cost may be more preferred.
S202: and acquiring at least two target medication recommendation models with better target analysis results according to the target analysis results of the original medication recommendation models corresponding to the disease types.
The original medication recommendation model is a medication recommendation model which is stored in the CDSS system in advance and corresponds to the disease type, and can output medication recommendation according to disease data. For each disease type, the CDSS system accesses at least one original medication recommendation model corresponding to the disease type, so that the CDSS system can intelligently analyze current disease data corresponding to the disease type based on the previously accessed original medication recommendation model, and the pertinence and the effectiveness of the recommended medication obtained by analysis are improved. It can be understood that different original medication recommendation models corresponding to the same disease type may be data obtained by performing model training on clinical medical data (including clinical disease data and corresponding clinical medication data) by using different neural network models, so that each original medication recommendation model is based on clinical medical guidelines to avoid clinical errors of recommended medication offenders, and also needs to implement personalized medication recommendation by using medical big data to learn medication schemes of similar patients to ensure effectiveness of medication recommendation.
Because the neural network models adopted by different original medication recommendation models are different and the training processes are different, different medication recommendation results may exist when the original medication recommendation models obtained through training perform intelligent medication recommendation on the same current disease data. Therefore, the CDSS system needs to perform comprehensive evaluation on each original medication recommendation model in advance, that is, the original medication recommendation models are comprehensively evaluated based on clinical medical data of similar patients to obtain a target analysis result, and the target analysis result is stored in the CDSS system, so that the CDSS system can quickly query the target analysis result of the original medication recommendation model according to the disease type in the medication recommendation process, perform descending order on all the original medication recommendation models according to the target analysis result, and select at least two original medication recommendation models with the target analysis result before to determine the original medication recommendation model as the target medication recommendation model adopted by the medication recommendation. The number of the target medication recommendation models can be set autonomously according to actual needs, and the former at least two original medication recommendation models in the target analysis results in descending order are determined as the target medication recommendation models, so that the effectiveness of the medication recommendation results can be guaranteed when intelligent medication recommendation is performed based on the target medication recommendation models.
S203: and analyzing and processing the current disease data by adopting at least two target medication recommendation models to obtain at least two original recommended medications.
The original recommended medication is an output result of the target medication recommendation model after analyzing and processing the current disease data. As an example, after obtaining at least two target medication recommendation models, the CDSS system analyzes and processes the current condition data by using the at least two target medication recommendation models in parallel to obtain the original recommended medication output by the at least two target medication recommendation models.
S204: and performing characteristic analysis on the original recommended medication to obtain characteristic analysis values corresponding to at least two characteristics to be analyzed.
The feature to be analyzed is a feature which is configured in advance and needs to be analyzed, and the feature to be analyzed can be understood as a feature for evaluating the quality of the medication recommendation. In this example, the characteristics to be analyzed include, but are not limited to, the cost of medication, the duration of the treatment effect, and side effects.
As an example, after obtaining at least two original recommended medications, the CDSS system performs feature analysis on the original recommended medications by using a preset feature analysis strategy to obtain feature analysis values corresponding to at least two features to be analyzed. In this example, the performing, by the CDSS system, the feature analysis on the original recommended medication by using the preset feature analysis policy may include: firstly, based on an original recommended medication inquiry system database or an associated database, feature original data corresponding to features to be analyzed in original recommended medication is obtained, for example, actual cost corresponding to the features to be analyzed, i.e., medication cost required for one treatment course, is obtained. And then, carrying out normalization processing on the feature original data corresponding to all the features to be analyzed so as to convert the feature original data corresponding to all the features to be analyzed into the quantity of dimensionless quantities, thereby ensuring the feasibility of calculating the subsequent recommended evaluation value.
S205: and determining a recommendation tendency type based on the user portrait data, and acquiring the feature weights corresponding to at least two features to be analyzed according to the recommendation tendency type.
The recommendation tendency type is a medication recommendation type which is analyzed according to the user portrait data of the patient and used for reflecting the tendency of the patient. As an example, the medication recommendation types include, but are not limited to, a low cost recommendation type with a higher cost weight, a high efficacy recommendation type with a higher efficacy time weight, and a low risk recommendation type with a higher side effect weight. In this example, at least two feature weights corresponding to the features to be analyzed are pre-configured for each medication recommendation type, and the corresponding feature weights are stored in the system database.
As an example, the CDSS system intelligently analyzes user portrait data by using a preset recommendation tendency analysis strategy to determine a recommendation tendency type corresponding to the user portrait data; and then according to the recommendation tendency type, acquiring the recommendation tendency type to query a system database, and acquiring the preset feature weight corresponding to each feature to be analyzed.
S206: and acquiring recommended evaluation values of at least two original recommended medications based on the feature analysis values and the feature weights corresponding to the at least two features to be analyzed, and determining the original recommended medication with the maximum recommended evaluation value as the target recommended medication.
Wherein the target recommended medication is a recommended medication scheme that best matches the user profile data, and is selected from at least two original recommended medications.
In this embodiment, after the CDSS system is based on the feature analysis values and the feature weights corresponding to the at least two features to be analyzed, a weighting algorithm is used to perform weighting operation on the feature analysis values and the feature weights corresponding to the at least two features to be analyzed, so as to obtain at least two recommended evaluation values of the original recommended medication; and comparing the at least two recommended evaluation values, determining the maximum value of the recommended evaluation values, and determining the original recommended medication corresponding to the maximum value of the recommended evaluation values as the target recommended medication. The recommended evaluation value is a numerical value obtained after weighting operation is carried out on the feature analysis values and the feature weights corresponding to at least two features to be analyzed, the fitting degree of the user portrait data of the patient can be reflected, the higher the recommended evaluation value is, the more the original recommended medication fits the actual requirements of the patient, and the pertinence and the effectiveness of medication recommendation are improved.
As an example, a CDSS system may employ P- ∑ SiWiCarrying out weighting operation on the characteristic analysis values and the characteristic weights corresponding to the at least two characteristics to be analyzed to obtain recommended evaluation values of the at least two original recommended medicaments, wherein the recommended evaluation values are simple and convenient to calculate in the obtaining process; where P is the recommended evaluation value, SiFor the feature analysis value corresponding to the ith feature to be analyzed, WiFor the ith feature pair to be analyzedThe corresponding feature weight.
As another example, after the target analysis result of each original medication recommendation model is obtained in step S202, the target analysis result may be further processed by using a model weight conversion strategy to obtain a model weight corresponding to each target analysis result, and then the CDSS system may use P-Q ∑ SiWiPerforming weighting operation on the characteristic analysis values and the characteristic weights corresponding to at least two characteristics to be analyzed to obtain at least two recommended evaluation values of the original recommended medication, wherein Q is the model weight corresponding to the original medication recommended model, P is the recommended evaluation value, and S isiFor the feature analysis value corresponding to the ith feature to be analyzed, WiAnd the feature weight corresponding to the ith feature to be analyzed. The method has the advantages that target analysis results of the original medication recommendation model are converted into model weights, and then the model weights, the feature analysis values and the feature weights are used for determining the recommendation evaluation value, so that the recommendation evaluation value can effectively reflect effectiveness of medication recommendation and fit degree of user portrait data, and effectiveness and pertinence of intelligent medication recommendation are improved.
In the medication recommendation method based on artificial intelligence provided by this embodiment, at least two target medication recommendation models with better target analysis results are obtained according to the target analysis results of the original medication recommendation models corresponding to the disease types, and then the target medication recommendation models are used to analyze and process current disease data to obtain original recommended medication, so that effectiveness and pertinence of the original recommended medication can be ensured; and then, carrying out feature analysis on the original recommended medication, determining a feature analysis value corresponding to the feature to be analyzed, determining a feature weight corresponding to the feature to be analyzed according to the recommendation tendency type determined by analyzing the user portrait data, and carrying out weighting operation based on the feature analysis value corresponding to the feature to be analyzed and the corresponding feature weight to determine a recommendation evaluation value, so that the recommendation evaluation value can effectively reflect the effectiveness of medication recommendation and the attaching degree with the user portrait data, and the effectiveness and pertinence of intelligent medication recommendation are improved.
In one embodiment, as shown in fig. 3, before obtaining the medication recommendation request, the artificial intelligence-based medication recommendation method further includes the following steps:
s301: retrospective data corresponding to the type of condition is obtained, the retrospective data including historical condition data and actual medication data.
The retrospective data refers to data used for evaluating the quality of an original medication recommendation model so as to determine a target analysis result. As an example, the retrospective data may be obtained from a system database of a hospital information system or a CDSS system, or may be some published medical data set, which is specifically structured data.
In this example, the retrospective data is a visit record, i.e. data determined by the patient on a clinic basis, including in particular historical condition data and actual medication data. Historical condition data refers to data relating to the condition recorded at the time of patient visit, including symptoms and test examination result information. The actual medication data refers to medication data given by a doctor during a visit, and specifically includes data such as a disease diagnosis result, prescription medication, treatment course time and the like. It can be understood that, in the process of comprehensively evaluating the original medication recommendation model, the effect of treating the disease symptoms corresponding to the historical disease data needs to be analyzed in combination with the actual medication data, and therefore, the retrospective data also includes short-term effect data and long-term effect data. The short-term effect data refers to data indicating that the examination index has reached the standard within a short period of time after the current visit, and for example, data indicating whether the glycated hemoglobin has reached the standard within about 3 months after the treatment may be determined as the short-term effect data. The long-term effect data refers to data on whether complications or adverse events occur for a long time after the visit, and the proportion of the data can be determined according to the results of follow-up outpatient or in-patient diagnosis. It will be appreciated that the temporal boundaries of the short term effects data and the long term effects data may be determined based on historical experience.
As an example, the CDSS system may query a system database of the hospital information system or the CDSS system or crawl a hospital data set according to the disease type, and obtain all historical visit records corresponding to the disease type; checking all historical treatment records, and judging whether the historical treatment records comprise necessary data such as historical disease data, actual medication data, short-term effect data, long-term effect data and the like; if the historical encounter record contains all necessary data, the historical encounter record is determined as retrospective data that can be used to evaluate the original medication recommendation model to ensure the feasibility of subsequent evaluation of the original medication recommendation model.
S302: analyzing and processing historical disease data by adopting an original medicine recommendation model corresponding to the disease type to obtain historical recommended medicine data, matching the historical recommended medicine data with actual medicine data to obtain a matching result, and dividing retrospective data into a first data set conforming to model recommendation and a second data set not conforming to the model recommendation based on the matching result.
Because the original medication recommendation model is a model capable of outputting medication recommendations according to the disease data, the CDSS system can analyze and process the historical disease data by using the original medication recommendation model to obtain the historical recommended medication data. Namely, the historical recommended medication data refers to medication data output by analyzing and processing historical disease data by adopting an original medication recommendation model.
After acquiring the historical recommended medication data, the CDSS system can perform matching processing on the historical recommended medication data and the actual medication data to determine whether the historical recommended medication data is matched with the actual medication data, and divide retrospective data into a first data set conforming to model recommendation and a second data set not conforming to the model recommendation according to a matching result, namely if the historical recommended medication data is matched with the actual medication data, classify the retrospective data into the first data set conforming to the model recommendation; if the historical recommended medication data does not match the actual medication data, the retrospective data is categorized into a second data set that does not conform to the model recommendation.
S303: and respectively counting the short-term effectiveness index standard-reaching rate and the long-term complication incidence rate of the retrospective data in the first data set and the second data set, and carrying out significance verification to obtain an effect analysis result.
Wherein the short-term effectiveness index achievement rate is used for evaluating whether the model recommends whether the short-term effectiveness index is up according to retrospective data from the visit level or the patient level. Long-term complication incidence is used to assess from an interview level or patient level, based on retrospective data, whether a model recommendation is in compliance with the impact on whether long-term complications occur. The visit level refers to a one-time visit record of a patient and is mainly suitable for evaluating whether the short-term effectiveness index reaches the standard or not, for example, whether the short-term effectiveness index reaches the standard or not in a short time after each visit. The patient grade refers to all visit records for the same patient, and is mainly used for evaluating whether long-term complications occur. As an example, the recommended compliance rate of the actual medication data and the historical recommended medication data of the same patient may be determined according to the multiple-visit records of the same patient, and the relationship between the recommended compliance rate of the patient level and the long-term complication occurrence rate may be further analyzed according to the recommended compliance rate. For example, if the same patient has 10 visits, that is, 10 visit records are formed, wherein 6 times of actual medication data are consistent with historical recommended medication data, that is, the recommended compliance rate is 60%, and then significance verification is performed according to the compliance model recommendation rate, so that an effect analysis result reflecting the relationship between the compliance model recommendation rate and the long-term complication occurrence rate is obtained. The significance test (significance test) is to make an assumption about the parameters of the population (random variables) or the distribution form of the population in advance, and then use the sample information to determine whether the assumption (alternative assumption) is reasonable, i.e. determine whether the true situation of the population is significantly different from the original assumption.
As an example, the CDSS system respectively counts the short-term effectiveness index achievement rate and the long-term complication occurrence rate corresponding to the first data set and the second data set according to retrospective data in the first data set and the second data set, so as to perform significance verification based on the counted short-term effectiveness index achievement rate and the counted long-term complication occurrence rate, thereby obtaining an effect analysis result, and the effect analysis result can effectively reflect a relationship between the short-term effectiveness index achievement rate and the long-term complication occurrence rate and the model compliance rate, so as to perform subsequent comprehensive evaluation.
S304: and if the effect analysis result meets the preset analysis condition, performing difference analysis on retrospective data in the first data set and the second data set to obtain the target short-term mixed factors and the target long-term mixed factors.
The preset analysis condition is a preset condition which needs to be subjected to subsequent difference analysis. As an example, when the effect analysis result is P-value obtained by significance check, the preset analysis result may be set to be less than a preset value, such as less than 0.05.
Since the difference between the short-term effectiveness index achievement rate and the long-term complication rate of the retrospective data in the first data set and the second data set may be caused by treatment factors or non-treatment factors, the treatment factors are related to medication, and the non-treatment factors are not related to medication, in order to avoid the influence of the non-treatment factors on the effectiveness of the recommended medication of the model, the retrospective data in the first data set and the second data set needs to be analyzed differently to obtain the target short-term complications and the target long-term complications. Confounding factors may be understood as non-therapeutic factors; the target short-term confounding factors refer to the first few non-treatment factors which have the greatest or greater influence on the short-term effectiveness index standard-reaching rate; the target long-term confounding factors refer to the first few non-therapeutic factors that have the greatest or greater effect on the incidence of long-term complications.
For example, the retrospective data includes N confounding factors such as F1, F2, F3 … … Fn, and when performing difference analysis on the retrospective data in the first data set and the second data set, and when analyzing any confounding factor, respectively counting the short-term effectiveness index standard-reaching rate and the long-term complication incidence rate corresponding thereto to obtain an effect analysis result corresponding to the confounding factor; and according to the result of the effect analysis, determining the first few confounding factors which have the greatest influence or have a greater influence on the short-term effectiveness index standard-reaching rate as target short-term confounding factors, and determining the first few confounding factors which have the greatest influence or have a greater influence on the long-term complication incidence as target short-term confounding factors.
S305: and performing tendency analysis on retrospective data in the first data set and the second data set based on the target short-term mixed factors and the target short-term mixed factors, and acquiring and storing a target analysis result of the original medication recommendation model.
In this example, based on the target short-term confounding factors and the target long-term confounding factors determined by the analysis, the tendency analysis is performed on the retrospective data in the first data set and the second data set, so as to weaken the influence of the confounding factors on the short-term effectiveness index achievement rate and the long-term complication occurrence rate, and make the obtained target analysis result reasonable. In this example, after obtaining the target analysis result of the original medication recommendation model, the CDSS system needs to store the target analysis result in the system database, so that after obtaining the medication recommendation request, the CDSS system can call the target analysis result of the original medication recommendation model, and quickly determine at least two target medication recommendation models with better target analysis results, thereby ensuring the effectiveness of selecting a target medication recommendation model that is clinically verified and learns the medication schemes of similar patients.
In an embodiment, as shown in fig. 4, the matching processing of the historical recommended medication data and the actual medication data in step S302 to obtain a matching result specifically includes the following steps:
s401: at least one recommended medication category is obtained based on the historical recommended medication data, and at least one actual medication category is obtained based on the actual medication data.
S402: and if all the recommended medication categories are matched with all the actual medication categories, the matching result is the model recommendation.
S403: and if at least one recommended medication category is not matched with the actual medication category, the matching result is not in accordance with the model recommendation.
In this example, the recommended medication category may be determined according to historical recommended medication data, the actual medication category may be determined according to actual medication data, and all recommended medication categories and all actual medication categories may be compared for matching; if all recommended medication categories are matched with all actual medication categories, namely the combination of the recommended medication categories is matched with the combination of the actual medication categories, the matching result is the model recommendation; if at least one recommended medication category is matched with the actual medication category, namely the combination of the recommended medication categories is not matched with the combination of the actual medication categories, the matching result is not in accordance with the model recommendation.
For example, drugs for diabetes can be classified into drug classes such as biguanides and sulfonylureas, and biguanide drugs are considered to be used when the patients actually take drugs containing metformin. It will be appreciated that the actual medication data may be defined as a combination of medication classes, such as biguanides + sulfonylureas, i.e., a single medication class of medication or multiple medication classes of medication may be used simultaneously. The historical recommended medication data and the actual medication data of the doctor are matched, namely are completely consistent in combination of medication categories, and if the medication categories are different in number or name, the medication categories are considered not matched. It can be understood that, because different drugs have different drug effects and the dosages of the drugs are not consistent under the same medication category, whether the dosages of the drugs are matched or not is not required to be considered, and whether historical recommended medication data and actual medication data are matched or not can be determined based on whether the medication categories are matched or not, which is beneficial to simplifying the calculation amount of the matching processing process and improving the processing efficiency.
In an embodiment, the retrospective data further comprises short-term effects data and long-term effects data. As shown in fig. 5, the step S303 of counting the short-term effectiveness index achievement rate and the long-term complication occurrence rate respectively for the retrospective data in the first data set and the second data set, and performing significance check to obtain an effect analysis result specifically includes the following steps:
s501: a first short-term effectiveness index achievement rate and a first long-term complication incidence rate are obtained based on retrospective data in the first dataset.
S502: a second short-term effectiveness index achievement rate and a second long-term complication incidence rate are obtained based on retrospective data in the second data set.
S503: and performing significance verification on the first short-term effectiveness index standard-reaching rate, the first long-term complication occurrence rate, the second short-term effectiveness index standard-reaching rate and the second long-term complication occurrence rate to obtain an effect analysis result.
The short-term effectiveness index standard-reaching rate refers to the ratio of the number of samples reaching the short-term effectiveness index standard to the total number of all samples. The long-term complication rate refers to the ratio of the number of samples with long-term complications to the total number of all samples.
As an example, assuming that the amount of retrospective data corresponding to the type of the condition acquired by the CDSS system is a, the CDSS system may divide the first data set and the second data set according to the matching result of the historical recommended medication data and the actual medication data, assuming that the amount of retrospective data in the first data set is a1 and the amount of retrospective data in the second data set is a 2.
Correspondingly, the CDSS system judges whether a standard range corresponding to the short-term effectiveness index is reached or not according to the short-term effect data in the retrospective data, if the short-term effect data reaches the standard range corresponding to the short-term effectiveness index, the short-term effectiveness index is determined to reach the standard, the number of samples meeting the short-term effectiveness index is updated, even if the number of samples meeting the short-term effectiveness index reaches the standard is added by 1, until all the retrospective data are analyzed completely, the number of samples meeting the short-term effectiveness index is obtained to be B1, and the number of samples meeting the short-term effectiveness index in the first data set is B2.
Correspondingly, the CDSS system judges whether a standard of the long-term complications is met or not according to the long-term effect data in the retrospective data, if the long-term effect data meets the standard of the long-term complications, the long-term complications are determined to occur, the number of samples of the long-term complications is updated, even if the number of samples of the long-term complications is increased by 1, until all the retrospective data are analyzed, the number of samples of the long-term complications is obtained to be C, wherein the number of samples of the long-term complications occurring in the first data set is C1, and the number of samples of the long-term complications occurring in the second data set is C2.
In this example, for retrospective data in the first data set and the second data set, the short-term effectiveness index achievement rate and the long-term complication occurrence rate are respectively counted, and significance verification is performed, which specifically includes: (1) the CDSS system acquires a first short-term effectiveness index standard-reaching rate and a first long-term complication incidence rate based on retrospective data in the first data set, and if the first short-term effectiveness index standard-reaching rate is Q1, Q1 is B1/A1; assuming that the first long-term complication incidence is P1, P1 ═ C1/a 1. (2) The CDSS system acquires a second short-term effectiveness index standard-reaching rate and a second long-term complication incidence rate based on retrospective data in a second data set, and if the second short-term effectiveness index standard-reaching rate is Q2, Q2 is B2/A2; assuming that the second long-term complication incidence is P2, P2 ═ C2/a 2. (3) The CDSS system adopts a chi-square inspection method to carry out significance verification on the first short-term effectiveness index achievement rate Q1, the first long-term complication occurrence rate P1, the second short-term effectiveness index achievement rate Q2 and the second long-term complication occurrence rate P2, and an effect analysis result is obtained.
As shown in table one below, for all retrospective data for which the short-term effectiveness indicator meets the standard, the flag is "yes" if the model recommendation is met; if the model recommendation is not met, marking as 'no', counting the number of samples B1 and B2 in the first data set and the second data set, respectively calculating a first short-term effectiveness index achievement rate Q1 and a second short-term effectiveness index achievement rate Q2 based on B1 and B2, and then calculating P-value1 according to the difference value of Q1 and Q2 and through significance check. Accordingly, retrospective data for long-term complications occurrences are labeled "yes" if they are consistent with model recommendations; if the model recommendation is not met, marking as 'no', counting the number of samples C1 and C2 in the first data set and the second data set, respectively calculating a first long-term complication incidence P1 and a second long-term complication incidence P2 based on C1 and C2, and then calculating P-value2 through significance check according to the difference value of P1 and P2. In the example, whether the model recommendation is met or not is used as an exposure variable, whether the short-term effectiveness index is met or not and whether the long-term complication is generated or not are used as outcome variables, P-value between the exposure variable and the outcome variable is calculated through a chi-square checking method, if the P-value is less than 0.05, the significance between the exposure variable and the outcome variable is shown, and an effect analysis result is obtained, so that the purpose of analyzing whether the model recommendation is related to the relationship between the short-term effectiveness index meeting rate and the long-term complication generation rate is achieved.
Figure BDA0002515873310000121
In an embodiment, as shown in fig. 6, step S304, namely, performing a difference analysis on the retrospective data in the first data set and the second data set to obtain the target short-term confounding factor and the target long-term confounding factor, specifically includes the following steps:
s601: confounders to be analyzed are determined.
The confounding factor to be analyzed is the confounding factor which needs to be analyzed and processed at this time. For example, if the retrospective data includes N confounders such as F1, F2, F3 … … Fn, one confounder may be randomly selected at a time as the confounder to be analyzed, e.g., for glycation achievement rate of diabetes, the initial glycated hemoglobin may be determined as its confounder to be analyzed.
S602: and screening out a first analysis subset and a second analysis subset corresponding to the confounders to be analyzed from retrospective data in the first data set and the second data set based on the confounders to be analyzed.
In this example, the CDSS system may divide the confounding factor to be analyzed into a plurality of first analysis subsets corresponding to the classification criteria according to a preset classification criterion, divide the second data set into a plurality of second analysis subsets corresponding to the classification criteria, and divide the retrospective data into the corresponding first analysis subsets and second analysis subsets according to the condition that a specific value corresponding to the confounding factor in the retrospective data satisfies the corresponding classification criterion.
S603: and respectively counting the short-term effectiveness index standard-reaching rate and the long-term complication incidence rate of retrospective data in the first analysis subset and the second analysis subset, carrying out significance verification, and determining the obtained effect analysis result as a mixed analysis result corresponding to the mixed factors to be analyzed.
It can be understood that the actual processes of steps S603 and S303 are substantially the same, and the detailed implementation scheme may refer to steps S501 to S503, which differ in that the data sets are different, and for avoiding repetition, the detailed description is not repeated here.
S604: and acquiring the target short-term confounding factors and the target long-term confounding factors according to confounding analysis results corresponding to confounding factors to be analyzed.
In this example, after statistics and significance check of short-term effectiveness index achievement rate and long-term complication occurrence rate are performed on retrospective data in the first analysis subset and the second analysis subset, the corresponding effect analysis result is determined as a mixed analysis result corresponding to the mixed factor to be analyzed, then according to the mixed analysis result, the mixed factors affecting the short-term effectiveness index achievement rate most or the first bits having a large influence are determined as target short-term mixed factors, and the mixed factors affecting the long-term complication occurrence rate most or the first bits having a large influence are determined as target short-term mixed factors. For example, initial glycated hemoglobin is the largest confounding factor for glycation achievement, and therefore, is identified as the target short-term confounding factor; the initial risk of complications is the largest confounding factor for the incidence of complications and, therefore, is identified as the target long-term confounding factor.
In an embodiment, as shown in fig. 7, in step S305, performing a tendency analysis on retrospective data in the first data set and the second data set based on the target short-term confounders and the target short-term confounders, and obtaining and storing a target analysis result of the original medication recommendation model, specifically, the method includes the following steps:
s701: and layering the first data set and the second data set based on the target short-term confounding factors and the target short-term confounding factors to obtain a layered data set.
S702: and performing tendency analysis on retrospective data in the hierarchical data sets to obtain a tendency result corresponding to each hierarchical data set.
S703: and acquiring and storing a target analysis result corresponding to the original medication recommendation model based on the tendency result corresponding to the hierarchical data set.
The hierarchical data set is formed by dividing the first data set and the second data set according to the target short-term confounding factor and the target long-term confounding factor. As an example, when the initial glycated hemoglobin is a target short-term confounding factor, the initial glycated hemoglobin may be divided into three hierarchical data sets of an initial standard, an initial standard and a severe standard, the short-term effectiveness index standard-reaching rates that meet the model recommendation and do not meet the model recommendation in each hierarchical data set are compared, that is, the hierarchical data sets are used as a processing group (a deal group) and a control group (a control group) for comparison, and the tendency score matching manner is used for controlling other confounding factors to obtain the tendency result in each hierarchical data set. As another example, when the initial concurrency risk is a target long-term confounding factor, the initial concurrency risk may be divided into three hierarchical data sets of low risk, medium risk and high risk, the long-term complication occurrence rates of two cases that meet the model recommendation and do not meet the model recommendation in each hierarchical data set are compared, that is, the hierarchical data sets are used as a processing group (deal group) and a control group (control group) for comparison, and the tendency score matching manner is adopted to control other confounding factors, so as to obtain the tendency result in each hierarchical data set.
In this example, after obtaining the tendency result in each hierarchical data set, the CDSS system compares the tendency result with a pre-set qualification evaluation threshold for determining whether the result is qualified, so as to determine whether the corresponding hierarchical data set is qualified; then, the proportion of the qualified quantity and the total quantity corresponding to the hierarchical data set is counted to obtain a target analysis result, so that the curative effect effectiveness of the original medication recommendation model is objectively analyzed based on objectively existing short-term effect data and long-term effect data in retrospective data, and the objectivity of the target analysis result is ensured.
In one embodiment, as shown in fig. 8, step S205, determining a recommendation tendency type based on the user profile data, specifically includes the following steps:
s801: a determination is made as to whether the user representation data includes an existing trend type.
S802: if the user portrait data contains an existing trend type, the existing trend type is determined to be a recommended trend type.
S803: if the user portrait data does not contain the existing tendency type, determining similar crowds based on the user portrait data, and determining a common tendency type corresponding to the similar crowds as a recommended tendency type; or acquiring related webpage data based on the user portrait data, counting the feature tags of the related webpage data, and acquiring the recommendation tendency type.
The existing trend type is a trend type explicitly indicated in the user image data. For example, the existing tendency types are configured on an entry interface of the CDSS system for selection by a user, so that the user can configure the corresponding tendency types through the entry interface.
As an example, the CDSS system analyzes user image data by adopting a character matching algorithm or other algorithms, judges whether the user image data comprises an existing tendency type, directly determines the existing tendency type as a recommended tendency type if the user image data comprises the existing tendency type, and determines the characteristic weight corresponding to each feature to be analyzed in original recommended medication based on the recommended tendency type, so that the target recommended medication determined based on the finally determined recommended evaluation value is more matched with the user image data, and the pertinence and the effectiveness of the target recommended medication are improved.
As an example, the CDSS system analyzes user portrait data by adopting a character matching algorithm or other algorithms, judges whether the user portrait data comprises an existing tendency type, determines similar people based on the user portrait data if the user portrait data does not comprise the existing tendency type, and determines a common tendency type corresponding to the similar people as a recommended tendency type. The similar population is the population closest to or most similar to the user portrait data of the patient, and specifically, the similar population is the population closest to or most similar to the user portrait data corresponding to the disease type. The similar population comprises a plurality of similar users which are closest to or most similar to the user portrait data of the patient, each similar user corresponds to a medication tendency type, and the medication tendency type refers to a tendency type which is determined by the similar users. The common tendency type corresponding to the similar population refers to the medication tendency type with the largest proportion determined by counting the medication tendency types of all similar users in all similar populations.
For example, user representation data includes, but is not limited to, gender, age, address, occupation, consumption habits, and athletic data. The CDSS system determines similar crowds based on the user portrait data, and the process of determining the common tendency type corresponding to the similar crowds as the recommendation tendency type comprises the following steps: (1) and performing distance calculation on the user portrait data of the patient and the user portrait data of any existing user stored in the system database by adopting a distance algorithm to obtain a similar distance corresponding to the two user portrait data. (2) And determining the existing users with the similar distance reaching the preset distance threshold value as similar users, and forming a similar crowd based on all the similar users. (3) And then carrying out statistical analysis on the medication tendency types of all similar users in the similar population, determining the medication tendency type with the largest proportion as a common tendency type, and determining the common tendency type as the recommended tendency type of the patient. The distance algorithm includes, but is not limited to, a euclidean distance algorithm. It is to be understood that since the user profile data of the similar population is closest or most similar to the user profile data corresponding to the patient, so that the trend types thereof are also relatively similar, the common trend type of the similar population may be determined as the recommended trend type of the patient.
As an example, the CDSS system analyzes user portrait data by adopting a character matching algorithm or other algorithms, judges whether the user portrait data comprises an existing tendency type, and if the user portrait data does not comprise the existing tendency type, acquires associated webpage data based on the user portrait data, and counts feature tags of the associated webpage data to acquire a recommended tendency type. The related webpage data refers to content data in a related webpage related to disease treatment corresponding to the disease type. The feature tag of the associated web page data refers to a tag which is configured in advance for the content data in each associated web page and is related to the feature to be analyzed, for example, a tag with low cost, short curative time, less side effects and the like.
For example, the process of obtaining associated web page data based on the user portrait data, counting feature tags of the associated web page data, and obtaining the recommendation tendency type includes: (1) and acquiring all historical webpage data which are accessed by the patient and correspond to the identity card number and/or the mobile phone number based on the identity card number and/or the mobile phone number in the user portrait data. (2) And determining historical webpage data related to the disease type in the user portrait data in all historical webpage data as related webpage data. (3) And counting the number of the feature tags in all the associated webpage data, and determining the recommendation tendency type according to the number. It can be understood that the feature labels in each associated webpage data can be obtained by adopting a Jieba word segmentation tool and a TF-IDF algorithm, that is, the Jieba word segmentation tool can be firstly adopted to scan, segment and label the part of speech of the text information in the associated webpage data to obtain word segmentation results; and extracting keywords from the word segmentation result by adopting a TF-IDF algorithm to obtain a characteristic label corresponding to the associated webpage data.
It should be understood that, the sequence numbers of the steps in the foregoing embodiments do not imply an execution sequence, and the execution sequence of each process should be determined by its function and inherent logic, and should not constitute any limitation to the implementation process of the embodiments of the present invention.
In an embodiment, an artificial intelligence-based medication recommendation device is provided, and the artificial intelligence-based medication recommendation device corresponds to the artificial intelligence-based medication recommendation method in the embodiment one to one. As shown in fig. 9, the artificial intelligence-based medication recommendation apparatus includes a recommendation request obtaining module 901, a recommendation model determining module 902, an original recommended medication obtaining module 903, a feature analysis value obtaining module 904, a feature weight obtaining module 905, and a target recommended medication obtaining module 906. The functional modules are explained in detail as follows:
a recommendation request obtaining module 901, configured to obtain a medication recommendation request, where the medication recommendation request includes a type of a medical condition, current medical condition data, and user portrait data.
The recommendation model determining module 902 is configured to obtain at least two target medication recommendation models with better target analysis results according to the target analysis result of the original medication recommendation model corresponding to the disease type.
An original recommended medication acquiring module 903, configured to analyze and process the current condition data by using at least two target medication recommendation models, and acquire at least two original recommended medications.
The characteristic analysis value obtaining module 904 is configured to perform characteristic analysis on the originally recommended medication to obtain characteristic analysis values corresponding to at least two characteristics to be analyzed.
The feature weight obtaining module 905 is configured to determine a recommended tendency type based on the user portrait data, and obtain feature weights corresponding to at least two features to be analyzed according to the recommended tendency type.
And a target recommended medication acquiring module 906, configured to acquire recommended evaluation values of at least two original recommended medications based on the feature analysis values and the feature weights corresponding to the at least two features to be analyzed, and determine the original recommended medication with the largest recommended evaluation value as the target recommended medication.
Preferably, the artificial intelligence-based medication recommendation device further comprises a retrospective data acquisition module, a data partitioning module, an effect analysis result acquisition module, a confounding factor determination module and a target analysis result acquisition module.
A retrospective data acquisition module to acquire retrospective data corresponding to a type of condition, the retrospective data including historical condition data and actual medication data.
The data dividing module is used for analyzing and processing historical disease data by adopting an original medication recommendation model corresponding to the disease type, acquiring historical recommended medication data, matching the historical recommended medication data with actual medication data, acquiring a matching result, and dividing retrospective data into a first data set which accords with model recommendation and a second data set which does not accord with the model recommendation based on the matching result.
And the effect analysis result acquisition module is used for respectively counting the short-term effectiveness index standard-reaching rate and the long-term complication incidence rate of the retrospective data in the first data set and the second data set, and performing significance verification to acquire an effect analysis result.
And the confounding factor determining module is used for performing difference analysis on retrospective data in the first data set and the second data set to acquire a target short-term confounding factor and a target long-term confounding factor if the effect analysis result meets a preset analysis condition.
And the target analysis result acquisition module is used for performing tendency analysis on retrospective data in the first data set and the second data set based on the target short-term mixed factors and the target short-term mixed factors, and acquiring and storing a target analysis result of the original medication recommendation model.
Preferably, the data partitioning module includes a medication category obtaining unit, a first matching result obtaining unit, and a second matching result obtaining unit.
And the medication category acquisition unit is used for acquiring at least one recommended medication category based on the historical recommended medication data and acquiring at least one actual medication category based on the actual medication data.
And the first matching result acquisition unit is used for matching all the recommended medication categories with all the actual medication categories, and the matching result is the model recommendation.
And the second matching result acquisition unit is used for judging that the matching result is not in accordance with the model recommendation if at least one recommended medication category is not matched with the actual medication category.
Preferably, the retrospective data also includes short-term effects data and long-term effects data.
The effect analysis result acquisition module comprises a first index acquisition unit, a second index acquisition unit and a significance verification unit.
A first index obtaining unit for obtaining a first short-term effectiveness index achievement rate and a first long-term complication occurrence rate based on retrospective data in the first data set.
And the second index acquisition unit is used for acquiring a second short-term effectiveness index standard-reaching rate and a second long-term complication incidence rate based on the retrospective data in the second data set.
And the significance checking unit is used for performing significance checking on the first short-term effectiveness index standard-reaching rate, the first long-term complication occurrence rate, the second short-term effectiveness index standard-reaching rate and the second long-term complication occurrence rate to obtain an effect analysis result.
Preferably, the confounding factor determining module includes a to-be-analyzed factor determining unit, an analysis subset dividing unit, a confounding analysis result obtaining unit, and a target confounding factor obtaining unit.
And the to-be-analyzed factor determining unit is used for determining the confounding factor to be analyzed.
And the analysis subset dividing unit is used for screening out a first analysis subset and a second analysis subset corresponding to the confounders to be analyzed from the retrospective data in the first data set and the second data set based on the confounders to be analyzed.
And the promiscuous analysis result acquisition unit is used for respectively counting the short-term effectiveness index standard-reaching rate and the long-term complication incidence rate of the retrospective data in the first analysis subset and the second analysis subset, carrying out significance check and determining the acquired effect analysis result as a promiscuous analysis result corresponding to the promiscuous factor to be analyzed.
And the target mixed factor obtaining unit is used for obtaining the target short-term mixed factors and the target long-term mixed factors according to the mixed analysis result corresponding to the mixed factors to be analyzed.
Preferably, the target analysis result acquisition module includes a hierarchical data set acquisition unit, a tendency result acquisition unit, and a target analysis result acquisition unit.
And the hierarchical data set acquisition unit is used for layering the first data set and the second data set based on the target short-term confounding factor and the target short-term confounding factor to acquire a hierarchical data set.
And the tendency result acquisition unit is used for performing tendency analysis on the retrospective data in the hierarchical data sets and acquiring a tendency result corresponding to each hierarchical data set.
And the target analysis result acquisition unit is used for acquiring and storing a target analysis result corresponding to the original medication recommendation model based on the tendency result corresponding to the hierarchical data set.
Preferably, the feature weight obtaining module 905 includes an existing tendency judging unit, a first tendency type determining unit, and a second tendency type determining unit.
An existing tendency judgment unit for judging whether the user portrait data contains an existing tendency type.
And a first tendency type determination unit configured to determine the existing tendency type as the recommended tendency type if the user portrait data includes the existing tendency type.
The second tendency type determining unit is used for determining similar crowds based on the user portrait data if the user portrait data does not contain the existing tendency types, and determining common tendency types corresponding to the similar crowds as recommended tendency types; or acquiring related webpage data based on the user portrait data, counting the feature tags of the related webpage data, and acquiring the recommendation tendency type.
For specific limitations of the artificial intelligence based medication recommendation device, reference may be made to the above limitations of the artificial intelligence based medication recommendation method, which are not described herein again. The modules in the artificial intelligence based medication recommending device can be wholly or partially realized by software, hardware and a combination thereof. The modules can be embedded in a hardware form or independent from a processor in the computer device, and can also be stored in a memory in the computer device in a software form, so that the processor can call and execute operations corresponding to the modules.
In one embodiment, a computer device is provided, which may be a server, and its internal structure diagram may be as shown in fig. 10. The computer device includes a processor, a memory, a network interface, and a database connected by a system bus. Wherein the processor of the computer device is configured to provide computing and control capabilities. The memory of the computer device comprises a nonvolatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program, and a database. The internal memory provides an environment for the operation of an operating system and computer programs in the non-volatile storage medium. The database of the computer device is used for storing data adopted or generated in the process of executing the artificial intelligence-based medication recommendation method. The network interface of the computer device is used for communicating with an external terminal through a network connection. The computer program is executed by a processor to implement an artificial intelligence-based medication recommendation method.
In one embodiment, a computer device is provided, which includes a memory, a processor, and a computer program stored in the memory and executable on the processor, and when the processor executes the computer program, the artificial intelligence based medication recommendation method in the foregoing embodiments is implemented, for example, S201 to S206 shown in fig. 2 or shown in fig. 3 to 8, which is not described herein again to avoid repetition. Alternatively, the processor implements the functions of the modules/units in the embodiment of the medication recommendation device based on artificial intelligence when executing the computer program, for example, the functions of the modules/units shown in fig. 9, and are not described herein again to avoid repetition.
In an embodiment, a computer-readable storage medium is provided, where a computer program is stored on the computer-readable storage medium, and when the computer program is executed by a processor, the method for recommending medication based on artificial intelligence in the foregoing embodiments is implemented, for example, S201-S206 shown in fig. 2, or shown in fig. 3 to 8, which is not described herein again to avoid repetition. Alternatively, the computer program, when executed by the processor, implements the functions of the modules/units in the above-mentioned artificial intelligence-based medication recommendation apparatus, for example, the functions of the modules/units shown in fig. 9, and in order to avoid repetition, the details are not described here again.
It will be understood by those skilled in the art that all or part of the processes of the methods of the embodiments described above can be implemented by hardware instructions of a computer program, which can be stored in a non-volatile computer-readable storage medium, and when executed, can include the processes of the embodiments of the methods described above. Any reference to memory, storage, database, or other medium used in the embodiments provided herein may include non-volatile and/or volatile memory, among others. Non-volatile memory can include read-only memory (ROM), Programmable ROM (PROM), Electrically Programmable ROM (EPROM), Electrically Erasable Programmable ROM (EEPROM), or flash memory. Volatile memory can include Random Access Memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in a variety of forms such as Static RAM (SRAM), Dynamic RAM (DRAM), Synchronous DRAM (SDRAM), Double Data Rate SDRAM (DDRSDRAM), Enhanced SDRAM (ESDRAM), Synchronous Link DRAM (SLDRAM), Rambus Direct RAM (RDRAM), direct bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM).
It will be apparent to those skilled in the art that, for convenience and brevity of description, only the above-mentioned division of the functional units and modules is illustrated, and in practical applications, the above-mentioned function distribution may be performed by different functional units and modules according to needs, that is, the internal structure of the apparatus is divided into different functional units or modules to perform all or part of the above-mentioned functions.
The above-mentioned embodiments are only used for illustrating the technical solutions of the present invention, and not for limiting the same; although the present invention has been described in detail with reference to the foregoing embodiments, it will be understood by those of ordinary skill in the art that: the technical solutions described in the foregoing embodiments may still be modified, or some technical features may be equivalently replaced; such modifications and substitutions do not substantially depart from the spirit and scope of the embodiments of the present invention, and are intended to be included within the scope of the present invention.

Claims (10)

1. A medication recommendation method based on artificial intelligence is characterized by comprising the following steps:
acquiring a medication recommendation request, wherein the medication recommendation request comprises a disease type, current disease data and user portrait data;
acquiring at least two target medication recommendation models with better target analysis results according to the target analysis results of the original medication recommendation models corresponding to the disease types;
analyzing and processing the current disease data by adopting at least two target medication recommendation models to obtain at least two original recommended medications;
performing characteristic analysis on the original recommended medication to obtain characteristic analysis values corresponding to at least two characteristics to be analyzed;
determining a recommendation tendency type based on the user portrait data, and acquiring feature weights corresponding to at least two features to be analyzed according to the recommendation tendency type;
and acquiring recommended evaluation values of at least two original recommended medicaments based on the feature analysis values and the feature weights corresponding to at least two features to be analyzed, and determining the original recommended medicament with the maximum recommended evaluation value as a target recommended medicament.
2. The artificial intelligence based medication recommendation method of claim 1, wherein prior to said obtaining a medication recommendation request, said artificial intelligence based medication recommendation method further comprises:
obtaining retrospective data corresponding to the type of condition, the retrospective data including historical condition data and actual medication data;
analyzing and processing the historical disease data by adopting an original medicine recommendation model corresponding to the disease type to obtain historical recommended medicine data, matching the historical recommended medicine data with the actual medicine data to obtain a matching result, and dividing the retrospective data into a first data set conforming to model recommendation and a second data set not conforming to the model recommendation based on the matching result;
respectively counting short-term effectiveness index standard-reaching rate and long-term complication incidence rate of retrospective data in the first data set and the second data set, and carrying out significance verification to obtain an effect analysis result;
if the effect analysis result meets a preset analysis condition, performing difference analysis on retrospective data in the first data set and the second data set to obtain a target short-term mixed factor and a target long-term mixed factor;
and performing tendency analysis on retrospective data in the first data set and the second data set based on the target short-term confounders and the target short-term confounders, and acquiring and storing a target analysis result of the original medication recommendation model.
3. The artificial intelligence-based medication recommendation method according to claim 2, wherein the matching of the historical recommended medication data and the actual medication data to obtain a matching result comprises:
acquiring at least one recommended medication category based on the historical recommended medication data, and acquiring at least one actual medication category based on the actual medication data;
if all the recommended medication categories are matched with all the actual medication categories, the matching result is in accordance with model recommendation;
and if at least one recommended medication category is not matched with the actual medication category, the matching result is that the model recommendation is not met.
4. The artificial intelligence based medication recommendation method of claim 2, wherein the retrospective data further comprises short-term effect data and long-term effect data;
the step of respectively counting short-term effectiveness index standard-reaching rate and long-term complication incidence rate of the retrospective data in the first data set and the second data set, carrying out significance verification and obtaining an effect analysis result comprises the following steps:
obtaining a first short-term effectiveness index achievement rate and a first long-term complication incidence rate based on retrospective data in the first data set;
obtaining a second short-term effectiveness index achievement rate and a second long-term complication incidence rate based on retrospective data in the second data set;
and performing significance verification on the first short-term effectiveness index achievement rate, the first long-term complication occurrence rate, the second short-term effectiveness index achievement rate and the second long-term complication occurrence rate to obtain an effect analysis result.
5. The artificial intelligence-based medication recommendation method of claim 2, wherein the performing a difference analysis on the retrospective data in the first data set and the second data set to obtain a target short-term confounding factor and a target long-term confounding factor comprises:
determining confounding factors to be analyzed;
screening out a first analysis subset and a second analysis subset corresponding to the confounders to be analyzed from retrospective data in the first data set and the second data set based on the confounders to be analyzed;
respectively counting short-term effectiveness index standard-reaching rate and long-term complication incidence rate of retrospective data in the first analysis subset and the second analysis subset, performing significance verification, and determining an obtained effect analysis result as a mixed analysis result corresponding to the mixed factors to be analyzed;
and acquiring the target short-term confounding factors and the target long-term confounding factors according to the confounding analysis result corresponding to the confounding factors to be analyzed.
6. The artificial intelligence-based medication recommendation method according to claim 2, wherein the performing a trend analysis on retrospective data in the first data set and the second data set based on the target short-term confounders and the target short-term confounders, and obtaining and storing target analysis results of the original medication recommendation model comprises:
layering the first data set and the second data set based on the target short-term confounding factors and the target short-term confounding factors to obtain a layered data set;
performing tendency analysis on retrospective data in the hierarchical data sets to obtain tendency results corresponding to the hierarchical data sets;
and acquiring and storing a target analysis result corresponding to the original medication recommendation model based on the tendency result corresponding to the hierarchical data set.
7. The artificial intelligence based medication recommendation method of claim 1, wherein said determining a recommendation trend type based on said user profile data comprises:
determining whether the user representation data includes an existing trend type;
if the user portrait data contains an existing tendency type, determining the existing tendency type as a recommended tendency type;
if the user portrait data does not contain the existing tendency type, determining a similar crowd based on the user portrait data, and determining a common tendency type corresponding to the similar crowd as a recommended tendency type; or acquiring associated webpage data based on the user portrait data, counting the feature tags of the associated webpage data, and acquiring a recommendation tendency type.
8. An artificial intelligence-based medication recommendation device, comprising:
the recommendation request acquisition module is used for acquiring a medication recommendation request, wherein the medication recommendation request comprises a disease type, current disease data and user portrait data;
the recommendation model determining module is used for acquiring at least two target medication recommendation models with better target analysis results according to the target analysis results of the original medication recommendation models corresponding to the disease types;
the original recommended medication obtaining module is used for analyzing and processing the current disease data by adopting at least two target medication recommending models to obtain at least two original recommended medications;
the characteristic analysis value acquisition module is used for carrying out characteristic analysis on the original recommended medication to acquire characteristic analysis values corresponding to at least two characteristics to be analyzed;
the characteristic weight acquisition module is used for determining a recommendation tendency type based on the user portrait data and acquiring characteristic weights corresponding to at least two characteristics to be analyzed according to the recommendation tendency type;
and the target recommended medication acquiring module is used for acquiring at least two recommended evaluation values of the original recommended medication based on the feature analysis values and the feature weights corresponding to the at least two features to be analyzed, and determining the original recommended medication with the maximum recommended evaluation value as the target recommended medication.
9. A computer device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the artificial intelligence based medication recommendation method of any one of claims 1 to 7 when executing the computer program.
10. A computer-readable storage medium, in which a computer program is stored, which, when being executed by a processor, implements the artificial intelligence based medication recommendation method according to any one of claims 1 to 7.
CN202010475961.3A 2020-05-29 2020-05-29 Medication recommendation method, device, equipment and storage medium based on artificial intelligence Pending CN111667890A (en)

Priority Applications (2)

Application Number Priority Date Filing Date Title
CN202010475961.3A CN111667890A (en) 2020-05-29 2020-05-29 Medication recommendation method, device, equipment and storage medium based on artificial intelligence
PCT/CN2020/099255 WO2021114624A1 (en) 2020-05-29 2020-06-30 Artificial intelligence-based medication recommendation method, apparatus, device, and storage medium

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202010475961.3A CN111667890A (en) 2020-05-29 2020-05-29 Medication recommendation method, device, equipment and storage medium based on artificial intelligence

Publications (1)

Publication Number Publication Date
CN111667890A true CN111667890A (en) 2020-09-15

Family

ID=72385194

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202010475961.3A Pending CN111667890A (en) 2020-05-29 2020-05-29 Medication recommendation method, device, equipment and storage medium based on artificial intelligence

Country Status (2)

Country Link
CN (1) CN111667890A (en)
WO (1) WO2021114624A1 (en)

Cited By (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN112151141A (en) * 2020-09-23 2020-12-29 康键信息技术(深圳)有限公司 Medical data processing method, device, computer equipment and storage medium
CN113241193A (en) * 2021-06-01 2021-08-10 平安科技(深圳)有限公司 Drug recommendation model training method, recommendation method, device, equipment and medium
CN113658659A (en) * 2021-08-20 2021-11-16 平安国际智慧城市科技股份有限公司 Medical information processing method and device, computer equipment and storage medium
CN113707253A (en) * 2021-08-30 2021-11-26 平安国际智慧城市科技股份有限公司 Medical scheme recommendation method, device, equipment and medium
CN116344009A (en) * 2023-05-22 2023-06-27 武汉盛博汇信息技术有限公司 Online diagnosis notification method and device

Families Citing this family (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113626558B (en) * 2021-07-07 2022-10-25 厦门市美亚柏科信息股份有限公司 Intelligent recommendation-based field standardization method and system
CN117198513B (en) * 2023-11-07 2024-03-29 北京烔凡科技有限公司 Health guidance method and system for hypertension patient

Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN106202893A (en) * 2016-06-30 2016-12-07 山东诺安诺泰信息系统有限公司 Method recommended by a kind of medicine
CN106919804A (en) * 2017-03-22 2017-07-04 李学明 Medicine based on clinical data recommends method, recommendation apparatus and server
US20180306797A1 (en) * 2017-04-24 2018-10-25 Washington University Methods of detecting ceramide
CN109801705A (en) * 2018-12-12 2019-05-24 平安科技(深圳)有限公司 Treat recommended method, system, device and storage medium
CN110070420A (en) * 2019-04-28 2019-07-30 广东技术师范大学 Recommended method, device and the computer readable storage medium of drug
CN111191020A (en) * 2019-12-27 2020-05-22 江苏省人民医院(南京医科大学第一附属医院) Prescription recommendation method and system based on machine learning and knowledge graph

Patent Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN106202893A (en) * 2016-06-30 2016-12-07 山东诺安诺泰信息系统有限公司 Method recommended by a kind of medicine
CN106919804A (en) * 2017-03-22 2017-07-04 李学明 Medicine based on clinical data recommends method, recommendation apparatus and server
US20180306797A1 (en) * 2017-04-24 2018-10-25 Washington University Methods of detecting ceramide
CN109801705A (en) * 2018-12-12 2019-05-24 平安科技(深圳)有限公司 Treat recommended method, system, device and storage medium
CN110070420A (en) * 2019-04-28 2019-07-30 广东技术师范大学 Recommended method, device and the computer readable storage medium of drug
CN111191020A (en) * 2019-12-27 2020-05-22 江苏省人民医院(南京医科大学第一附属医院) Prescription recommendation method and system based on machine learning and knowledge graph

Cited By (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN112151141A (en) * 2020-09-23 2020-12-29 康键信息技术(深圳)有限公司 Medical data processing method, device, computer equipment and storage medium
CN113241193A (en) * 2021-06-01 2021-08-10 平安科技(深圳)有限公司 Drug recommendation model training method, recommendation method, device, equipment and medium
CN113658659A (en) * 2021-08-20 2021-11-16 平安国际智慧城市科技股份有限公司 Medical information processing method and device, computer equipment and storage medium
CN113658659B (en) * 2021-08-20 2023-07-21 深圳平安智慧医健科技有限公司 Medical information processing method, medical information processing device, computer equipment and storage medium
CN113707253A (en) * 2021-08-30 2021-11-26 平安国际智慧城市科技股份有限公司 Medical scheme recommendation method, device, equipment and medium
CN113707253B (en) * 2021-08-30 2023-06-27 深圳平安智慧医健科技有限公司 Medical scheme recommendation method, device, equipment and medium
CN116344009A (en) * 2023-05-22 2023-06-27 武汉盛博汇信息技术有限公司 Online diagnosis notification method and device
CN116344009B (en) * 2023-05-22 2023-08-15 武汉盛博汇信息技术有限公司 Online diagnosis notification method and device

Also Published As

Publication number Publication date
WO2021114624A1 (en) 2021-06-17

Similar Documents

Publication Publication Date Title
CN111667890A (en) Medication recommendation method, device, equipment and storage medium based on artificial intelligence
CN108986908B (en) Method and device for processing inquiry data, computer equipment and storage medium
US9443002B1 (en) Dynamic data analysis and selection for determining outcomes associated with domain specific probabilistic data sets
CN112700838A (en) Big data-based medication scheme recommendation method and device and related equipment
CN112017789B (en) Triage data processing method, triage data processing device, triage data processing equipment and triage data processing medium
US20120109683A1 (en) Method and system for outcome based referral using healthcare data of patient and physician populations
WO2020034801A1 (en) Medical feature screening method and apparatus, computer device, and storage medium
US20230035564A1 (en) Diabetes onset and progression prediction using a computerized model
Rashidian et al. Deep learning on electronic health records to improve disease coding accuracy
CN109543718B (en) Method and device for modifying disease type description
Fan et al. Detecting glaucoma in the ocular hypertension study using deep learning
US20090055217A1 (en) Method and system for identifying and communicating a health risk
Manna et al. Diabetes prediction model using cloud analytics
Yousefi et al. Opening the black box: Personalizing type 2 diabetes patients based on their latent phenotype and temporal associated complication rules
Chakraborty et al. A robust framework for accelerated outcome-driven risk factor identification from ehr
Poh et al. Challenges in designing an online healthcare platform for personalised patient analytics
CN116705310A (en) Data set construction method, device, equipment and medium for perioperative risk assessment
Andrews et al. Validating effectiveness of subgroup identification for longitudinal data
CN116313086A (en) Sub-health prediction model construction method, device, equipment and storage medium
Singla et al. Developing clinical decision support system using machine learning methods for type 2 diabetes drug management
CN111274231B (en) Abnormal medical insurance data checking method and device, computer equipment and storage medium
CN113643783A (en) Sub-health population drug recommendation method, system, equipment and storage medium
CN111931091B (en) Page content display method and device, computer equipment and storage medium
CN112599217A (en) Method and device for representing medication behaviors of patient, computer equipment and storage medium
CN113764067B (en) Drug recommendation method, system, device and storage medium

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
REG Reference to a national code

Ref country code: HK

Ref legal event code: DE

Ref document number: 40032372

Country of ref document: HK

SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination