Method for automatically determining importance of medication counseling information by using artificial intelligence and system for executing same

Abstract

A system for automatically determining importance of medication counseling information by using artificial intelligence, according to one embodiment of the present invention, may comprise: a medication guide providing server that provides a medication guide; an importance estimation model generating device that, upon receiving the medication guide from the medication guide providing server, executes analysis of the medication guide to predict importance grades for respective pieces of medication counseling information, generates a model, and then trains the model by correcting the importance grades using a user's feedback; and a medication guide providing server that, when the user purchases a medicine or receives a medicine by means of a prescription, receives a medication guide for the medicine from the medication guide providing server, provides the medication guide for the medicine to the importance estimation model generating device, and, upon receiving the importance grades for respective phrases in the medication guide from the importance estimation model generating device, generates and provides the medication guide according to the importance grades.

Description

Method for automatically determining the importance of medication consultation information using artificial intelligence and a system for executing the same

[0001] The present invention relates to a method for automatically determining the importance of medication consultation information using artificial intelligence and a system for executing the same, and more specifically, to a method for automatically determining the importance of medication consultation information using artificial intelligence and a system for executing the same, which enables automatic determination of the importance level of medication consultation information by creating a model that can automatically determine the importance level of each medication consultation information using artificial intelligence.

[0002] In general, the dispensing and sale of medicines to patients must be accompanied by guidance on how to administer the medicines provided to the patient.

[0003] The above medication guide provides information such as the name of the drug, dosage / administration, efficacy / effect, storage method, side effects, and interactions. Recently, there has been a demand for additional information such as measures to take in case of side effects, side effects of the drug, precautions when taking the drug, and how to take the drug.

[0004] However, it is realistically difficult for pharmacists to provide all of this information to each and every patient who visits the pharmacy.

[0005] In order to maintain the professionalism of pharmacists and the high-quality pharmacy services that the public demands, pharmacists' medication guidance is of utmost importance.

[0006] If effective medication guidance is not provided at pharmacies, patients will not take their medications properly, which increases the probability of treatment failure. This will lead to repeated visits to medical institutions due to treatment failure, which is expected to lead to increased waste of medical costs socially and economically.

[0007] Here, medication guidance for the relevant medication is provided based on this patient-specific prescription database. Therefore, patient medication guidance includes information on the medication's efficacy / effects and dosage instructions. Pharmacists directly inform patients about this information based on this information.

[0008] However, medication maps stored offline do not provide patients with sufficient or necessary information, which is emerging as a barrier to the activation of medication maps.

[0009] In other words, pharmacists must faithfully provide medication guidance to patients to ensure safe use of medications, and patients must carefully listen to the pharmacist's guidance. However, medications vary in efficacy and effectiveness, and even within the same medication, they have diverse indications. In some cases, they are even prescribed for purposes other than those approved.

[0010] Additionally, side effects of medications can vary depending on age or other factors, and it is difficult to identify in advance whether a person is vulnerable to the side effects of a particular medication.

[0011] For example, even if a patient with liver disease is prescribed a drug with concerns about hepatotoxicity, it may be difficult to fully inform the doctor or pharmacist of the risk if the patient does not inform the doctor or pharmacist in advance, and even if the patient does inform the doctor or pharmacist in advance, the doctor or pharmacist may not be able to provide appropriate guidance on response.

[0012] Furthermore, medication information sheets are written based on specialized pharmaceutical knowledge and often contain content that requires a high level of literacy. Therefore, the content and expression of medication information sheets should vary depending on the patient, and therefore, there is an urgent need for a medication information system that goes beyond the limitations of offline systems.

[0013] The present invention relates to a method for automatically determining the importance of medication consultation information using artificial intelligence, which creates a model that can automatically determine the importance level of each medication consultation information using artificial intelligence, and a system for executing the same.

[0014] The objectives of the present invention are not limited to those mentioned above. Other objectives and advantages of the present invention not mentioned above can be understood through the following description and will be more clearly understood through the embodiments of the present invention. Furthermore, it will be readily apparent that the objectives and advantages of the present invention can be realized by the means and combinations thereof set forth in the claims.

[0015] To achieve this purpose, a system for automatically determining the importance of medication counseling information using artificial intelligence may include a medication counseling server that provides medication counseling, a medication counseling server that receives medication counseling from the medication counseling server, analyzes the medication counseling to predict an importance level for each medication counseling information, creates a model, and then trains the model by correcting the importance level using user feedback, and a medication counseling server that receives medication counseling for a medication from the medication counseling server when a user purchases a medication or receives a medication through a prescription, provides the medication counseling for the medication to the importance level calculation model creation device, and, upon receiving an importance level for each phrase of the medication counseling from the importance level calculation model creation device, creates and provides medication counseling according to the importance level.

[0016] In addition, a method for automatically determining the importance of medication counseling information using artificial intelligence to achieve this purpose may include a step in which, when a medication counseling information generating device receives a medication counseling information from a medication counseling information providing server, the device analyzes the medication counseling information to predict an importance level for each medication counseling information and generates and provides a model; a step in which the medication counseling information providing server receives medication counseling information about a drug from the medication counseling information providing server when a user purchases a drug or receives a drug through a prescription; a step in which the medication counseling information providing server provides medication counseling information about a drug to the device for generating an importance level model; a step in which, when the device for generating an importance level for each phrase of the medication counseling information receives an importance level for each phrase of the medication counseling information from the device for generating an importance level model, the device generates and provides medication counseling information according to the importance level; and a step in which the device for generating an importance level model trains the model by correcting the importance level using user feedback.

[0017] According to the present invention as described above, there is an advantage in that a model can be created that can automatically determine the importance level of each medication consultation information through artificial intelligence, thereby automatically determining the importance level of medication consultation information.

[0018] Figure 1 is a network configuration diagram for explaining a system for automatically determining the importance of medication consultation information using artificial intelligence according to one embodiment of the present invention.

[0019] FIG. 2 is a block diagram illustrating the internal structure of an importance calculation model generation device according to one embodiment of the present invention.

[0020] Figure 3 is a flowchart illustrating one embodiment of a method for automatically determining the importance of medication consultation information using artificial intelligence according to the present invention.

[0021]

[0022] The above-described objects, features, and advantages will be described in detail below with reference to the accompanying drawings, so that those skilled in the art can easily practice the technical idea of ​​the present invention. In describing the present invention, if it is determined that a detailed description of a known technology related to the present invention may unnecessarily obscure the gist of the present invention, a detailed description thereof will be omitted. Hereinafter, preferred embodiments of the present invention will be described in detail with reference to the accompanying drawings. In the drawings, the same reference numerals are used to indicate the same or similar components.

[0023] As used herein, the term "medication guide" refers to a document containing information about a medication prescribed by a doctor or pharmacist. This document is provided to help patients take the medication safely and effectively. Typically, a medication guide may include the following:

[0024] Figure 1 is a network configuration diagram for explaining a system for automatically determining the importance of medication consultation information using artificial intelligence according to one embodiment of the present invention.

[0025] Referring to Fig. 1, the system for automatically determining the importance of medication consultation information using artificial intelligence includes a medication information provision server (100), an importance calculation AI model generation device (200), and a medication information generation device (300).

[0026] The medication information provision server (100) is a server located at a pharmaceutical manufacturer, medical institution, health authority, etc. that provides medication information.

[0027] A medication guide may include drug information, uses and effects, dosage, precautions and warnings, storage instructions, and consultation with a physician. Drug information may include the brand name, generic name, dosage, and administration method of the prescribed medication.

[0028] Uses and Effects may include information about the intended use of the drug and its effects. Directions for Use include instructions on how to take the drug.

[0029] This may include dosage, time of administration, method of administration, and duration of administration. Precautions and warnings may include information about precautions to take while taking the drug or side effects.

[0030] This may include interactions with certain diseases and precautions when taking with other medications. Storage instructions provide guidance on how to properly store the medication. This may include storage temperature, humidity, and keeping it away from direct sunlight.

[0031] The importance assessment AI model generation device (200) collects various types of medication instructions through communication with the medication instruction provision server (100).

[0032] This importance assessment AI model generation device (200) can obtain medication information from pharmaceutical manufacturers, medical institutions, health authorities, etc. It can also collect medication information from online databases, medical literature, academic papers, etc.

[0033] Thereafter, the importance estimation AI model generation device (200) analyzes the medication information and generates an importance estimation model (210) to determine the importance of medication consultation information. At this time, the types of medication consultation information may include dosage method, side effects, storage method, dosage cycle, dosage timing, dosage period, drug interactions, etc.

[0034] The above dosage timing is the most important information stated in the instructions and is essential to maximize the effectiveness of the drug and minimize side effects.

[0035] It's important to understand how to take your medication. For example, you need to know whether to take it before or after a meal.

[0036] It's important to understand the dosage schedule, which includes how many times a day you should take the medication and how long you should take it for. This is because the number of doses and the interval between doses can affect effectiveness and safety.

[0037] Storage methods are crucial for maintaining the stability of pharmaceuticals. This means ensuring the proper temperature, humidity, and storage location to ensure effective results.

[0038] To this end, the importance estimation AI model generation device (200) analyzes the medication instructions using text analysis techniques. During this process, the importance estimation AI model generation device (200) utilizes natural language processing (NLP) technology to process text, understand sentence structures, and extract keywords.

[0039] The importance assessment AI model generation device (200) analyzes the syntax of the medication information and assigns tags to determine the importance based on whether the type of medication consultation information affects the disease improvement effect, side effects, and safety.

[0040] In one embodiment, the importance estimation AI model generation device (200) can predict the importance level as high when the type of medication consultation information is the time of medication, because it directly affects the treatment effect and side effects of a disease.

[0041] For example, in the case of “take medicine after a meal,” the importance estimation AI model generation device (200) determines the type of medication consultation information based on the timing of taking the medicine, and since the timing of taking the medicine directly affects the effectiveness and side effects, the importance level can be predicted as high.

[0042] In another embodiment, the AI ​​model generation device (200) for calculating importance may determine the importance level as high or medium because, if the type of medication consultation information is a method of taking, it may affect the treatment effect and side effects of a disease, but has an effect on adaptation rather than the timing of taking.

[0043] For example, in the case of “take with a set amount of water,” the AI ​​model generation device (200) determines the type of medication consultation information as the dosage method, and in the case of the dosage method, it may affect the treatment effect and side effects of the disease, but since it has an effect on adaptation rather than the timing of taking, it may determine the importance level as high or medium.

[0044] In another embodiment, the importance assessment AI model generation device (200) may determine the importance level as medium level when the type of medication consultation information is a dosing cycle, as it directly affects the treatment effect and safety of a disease.

[0045] For example, in the case of “Take it twice a day,” the AI ​​model generation device (200) determines the type of medication consultation information as the dosage cycle, and in the case of the dosage cycle, it can predict the importance level as an intermediate level because it affects the treatment effect and safety of the disease.

[0046] In addition, the importance assessment AI model generation device (200) analyzes the syntax of the medication information to determine the type of medication consultation information, and then assigns a tag to determine the importance according to the type of syntax.

[0047] In one embodiment, the importance estimation AI model creation device (200) can predict the importance level of the corresponding medication consultation information as high by analyzing the syntax of the medication information and if the type of the syntax is a command sentence, a negative sentence, or an enumeration sentence.

[0048] For example, in the case of the command “Take it,” the importance estimation AI model creation device (200) can determine the type of medication consultation information based on the time of taking the medication and predict the importance level as high.

[0049] For another example, in the case of a negative sentence such as “Do not take before meals,” the importance estimation AI model generation device (200) can determine the type of medication consultation information based on the time of taking the medication and predict the importance level as high.

[0050] As another example, in the case of the enumerated sentence “Take in the morning and evening,” the importance estimation AI model creation device (200) can determine the type of medication consultation information based on the time of taking and predict the importance level as high.

[0051] In another embodiment, the importance estimation AI model creation device (200) can analyze the syntax of the medication information and, if the type of syntax is a conditional sentence or an explanatory sentence, predict the importance level of the corresponding medication consultation information as a medium level.

[0052] For example, in the case of the conditional statement, "If symptoms occur, consult a doctor.", the type of medication consultation information can be determined as additional information and the importance level can be predicted to be medium.

[0053] Additionally, the importance estimation AI model generation device (200) parses the syntax of the medication information and calculates the frequency of word appearance (TF) in each sentence or paragraph of the medication information if the type of medication consultation information is about precautions and interactions. This indicates the importance of the word. For example, if the word "medicine" appears twice in the sentence "Do not take the medicine with food after taking the medicine," the TF for "medicine" is 2.

[0054] Then, the importance estimation AI model creation device (200) calculates the inverse document frequency (IDF) by taking the inverse of the number of sentences or paragraphs in which each word appears in the entire medication instructions. At this time, the inverse document frequency (IDF) is a value obtained by taking the inverse of the frequency, and the inverse document frequency (IDF) of a commonly appearing word is low, and the inverse document frequency (IDF) of a rarely appearing word is high.

[0055] That is, the importance estimation AI model generation device (200) can calculate the inverse document frequency (IDF) using [Mathematical Formula 1] below.

[0056]

[0057] [Mathematical Formula 1]

[0058]

[0059]

[0060] N: Number of entire sentences or paragraphs

[0061] DF(w): The number of sentences or paragraphs in which a specific word w appears

[0062]

[0063] The reason for adding 1 to the denominator in the above [Mathematical Formula 1] is to prevent the result of the natural logarithm function from becoming infinite because the denominator becomes 0 if a certain word does not appear in the entire document.

[0064] Therefore, the inverse document frequency (IDF) value indicates how rare a specific word is in the entire document set, and the higher the value, the higher the importance of the word.

[0065] The importance calculation AI model creation device (200) can calculate the importance score of a word using word frequency (TF) and inverse document frequency (IDF(W)) as in [Mathematical Formula 2], and predict the importance level according to the importance score.

[0066]

[0067] [Equation 2]

[0068] Generative AI model (w) = TF (w) * IDF (w)

[0069] Generative AI model (w): importance scores of words,

[0070] TF(w): frequency of a specific word w,

[0071] IDF(w): rarity of a particular word w,

[0072]

[0073] The generative AI model above considers the frequency of a word's appearance within a document, while also considering its rarity within the document set to indicate its importance. A higher TF indicates a word's importance within the document, while a higher IDF indicates a word's uniqueness and importance within the document set.

[0074] For example, if the TF for the word "medicine" is 10 and the IDF is 1.5, the generative AI model for that word would be 15. Therefore, this value evaluates the importance of the word by reflecting its frequent occurrence within a document but its rarity across the entire document set.

[0075] In addition, the importance estimation AI model generation device (200) can predict the importance grade based on the score according to the generated AI model (Term Frequency-Inverse Document Frequency) and the expert evaluation score in each case of the type of medication consultation information such as precautions, interactions, and others by analyzing the syntax of the medication information.

[0076] In one embodiment, the importance estimation AI model creation device (200) can predict the importance level by analyzing the syntax of the medication information and using the likelihood of side effects and the severity of side effects when the type of medication consultation information is a side effect.

[0077] In the above embodiment, when the type of medication consultation information is a side effect, the importance calculation AI model creation device (200) extracts words related to side effects from the medication guidance, calculates the frequency of the words related to side effects, and can predict an importance level based on the frequency of the words related to side effects.

[0078] For example, the AI ​​model creation device (200) for calculating importance can assign a corresponding side effect occurrence probability grade of 10% or more, 1% or more but less than 10%, 0.1% or more but less than 1%, and 0.01% or more but less than 0.1%, respectively, according to the principles of notation of pharmaceutical approval information, when a word related to a side effect is expressed as 'very common' in the medication guide, when a word related to a side effect is expressed as 'common' in the medication guide, when a word related to a side effect is expressed as 'uncommon' or 'occasionally' in the medication guide, when a word related to a side effect is expressed as 'rarely' in the medication guide, and when a word related to a side effect is expressed as 'very rarely' in the medication guide, respectively, in accordance with the principles of notation of pharmaceutical approval information, and when a specific occurrence frequency is presented together with a word related to a side effect, the numerical value can be directly extracted and assigned to the side effect occurrence probability.

[0079] As another example, the importance estimation AI model creation device (200) can identify sentences or paragraphs containing expressions recommending discontinuation, such as “stop administration,” “must stop,” and “do not continue taking,” from text data extracted from the medication instructions, and predict the importance level of the corresponding medication consultation information as high.

[0080] As described above, the importance estimation AI model creation device (200) predicts the importance level for medication consultation information, and then receives expert feedback and user feedback on the prediction result of the importance estimation model (210) to improve the importance estimation model (210).

[0081] The importance estimation AI model creation device (200) predicts the importance level for medication consultation information, and then receives expert feedback and user feedback on the prediction results of the importance estimation model (210) to improve the importance estimation model (210).

[0082] The importance estimation AI model generation device (200) receives feedback on the medication instructions provided by the medication instructions generation device (300). To this end, the importance estimation model learning unit (203) provides a feedback form or interface for user feedback input. The feedback is primarily provided in text format and may include user opinions or improvement requests.

[0083] First, the importance assessment AI model creation device (200) assigns an emotional score to each predetermined emotional word among the words used in the feedback, if any. For example, a positive score is assigned to each word if it conveys a positive emotion, and a negative score is assigned if it conveys a negative emotion.

[0084] As described above, the importance calculation AI model creation device (200) can calculate the emotional score of the entire sentence by adding up the emotional scores of each word used in the feedback, and determine the type of feedback based on the emotional score.

[0085] For example, the importance estimation AI model creation device (200) can compare the emotional score and a predetermined threshold score, and determine the type of feedback as positive feedback if the emotional score is higher than the predetermined threshold score, and determine the type of feedback as negative feedback if the emotional score is lower than the predetermined threshold score.

[0086] The importance assessment AI model generation device (200) integrates the original medication information data and feedback data to create a dataset, and readjusts the importance rating of medication consultation information based on the feedback. For example, the importance assessment model learning unit (203) maintains the rating if the user's feedback is positive, and modifies the rating if it is negative.

[0087] The medication guide generation device (300) receives medication guides for a medicine from a medication guide provision server (100) when a user purchases a medicine or receives a medicine through a prescription, and provides the medication guides for the medicine to an AI model generation device (200) for calculating importance of the medication guides.

[0088] When the medication information generation device (300) receives the importance level for each phrase of the medication information from the importance calculation AI model generation device (200), it can generate the medication information according to the importance level.

[0089] In one embodiment, the medication guidance generation device (300) can change the order of medication consultation information and provide it in order of importance level from high to low.

[0090] FIG. 2 is a block diagram illustrating the internal structure of an importance calculation model generation device according to one embodiment of the present invention.

[0091] Referring to FIG. 2, the importance estimation AI model generation device (200) includes a medication information collection unit (201), an importance estimation model generation unit (202), and an importance estimation model learning unit (203).

[0092] Figure 3 is a flowchart illustrating one embodiment of a method for automatically determining the importance of medication consultation information using artificial intelligence according to the present invention.

[0093] Referring to FIG. 3, when a device for generating an importance estimation model receives a medication guide from a medication guide provision server, it analyzes the medication guide, predicts an importance level for each medication consultation information, and generates and provides a model (step S310).

[0094] The medication information provision server receives medication information about a drug from the medication information provision server when a user purchases a drug or receives a drug through a prescription (step S320).

[0095] The above medication information provision server provides medication information for a drug to the importance calculation model creation device (step S330).

[0096] When the above medication information provision server receives the importance level for each phrase of the medication information from the importance calculation model generation device, it generates and provides the medication information according to the importance level (step S340).

[0097] The above-mentioned importance estimation model creation device trains the model by correcting the importance level using user feedback (step S350).

[0098] While described with reference to limited embodiments and drawings, the present invention is not limited to the above-described embodiments, and various modifications and variations are possible based on this disclosure by those skilled in the art. Accordingly, the scope of the present invention should be understood solely by the scope of the claims set forth below, and all equivalent or equivalent modifications thereof are deemed to fall within the scope of the present invention.

Claims

1. Medication information provision server that provides medication information; An importance calculation model generation device that, when receiving a medication guide from the above medication guide providing server, analyzes the medication guide, predicts the importance level of each medication consultation information, generates a model, and then trains the model by correcting the importance level using user feedback; and A system for automatically determining the importance of medication counseling information using artificial intelligence, characterized in that the system includes a medication guidance provision server that receives medication guidance for a medication from a medication guidance provision server when a user purchases a medication or receives a medication through a prescription, provides the medication guidance for the medication to an importance estimation model generation device, and generates and provides medication guidance according to the importance estimation grade when receiving an importance grade for each phrase of the medication guidance from the importance estimation model generation device.

2. A step of generating an importance estimation model in which, when a medication information provision device receives a medication information from a medication information provision server, the device analyzes the medication information to predict the importance level for each medication consultation information and generates and provides a model; A step in which a medication information provision server receives medication information for a drug from a medication information provision server when a user purchases a drug or receives a drug through a prescription; A step in which the above medication information provision server provides medication information for a medicine to a model generating device for calculating importance; When the above medication information provision server receives the importance grade of each phrase of the medication information from the importance calculation model generation device, a step of generating and providing the medication information according to the importance grade; and The above-mentioned importance estimation model generation device is characterized in that it includes a step of training the model by correcting the importance grade using the user's feedback. A method for automatically determining the importance of medication consultation information using artificial intelligence.

Images