Active detection system for adverse drug reactions using periodic conversion database and artificial intelligence
The active adverse drug reaction detection system using a period conversion database and AI technology addresses the inefficiencies in current methods by constructing a predictive model from electronic medical records, enabling timely and accurate adverse drug reaction detection and improving drug safety management.
Patent Information
- Application Number
- PCT/KR2024/002623
- Authority / Receiving Office
- WO · WO
- Patent Type
- Applications
- Current Assignee / Owner
- Priority Date
- 2023-05-30
- Filing Date
- 2024-02-29
- Publication Date
- 2024-12-05
AI Technical Summary
Current methods for detecting adverse drug reactions are inefficient and lack timely detection, leading to potential harm to patients, as they rely on voluntary reporting and are resource-intensive, making it difficult to systematically manage and prevent adverse events.
An active adverse drug reaction detection system using a period conversion database and artificial intelligence technology, which constructs a regularly updated database from electronic medical records and utilizes machine learning models like Random Forest, Logistic Regression, and neural networks to predict adverse reactions, providing timely and accurate alerts.
The system enables efficient and timely detection of adverse drug reactions, reducing the risk of harm by providing predictive alerts and improving drug safety management, reducing the need for extensive resource allocation and improving post-marketing surveillance efficiency.
Smart Images

Figure KR2024002623_05122024_PF_FP_ABST
Abstract
Description
Active Adverse Drug Reaction Detection System Using Periodic Conversion Database and Artificial Intelligence
[0001] The present invention relates to the field of artificial intelligence, and more specifically, to a method and device for active detection of adverse drug reactions using artificial intelligence technology.
[0002] Adverse drug reactions (ADRs) are defined as harmful and unintended reactions to medicines, which can pose a risk to a patient's health and range in severity from mild to very serious. Recently, with the increase in global drug sales, the number of ADRs has also increased. This has led to calls for countermeasures, and drug safety assessments and domestic and international regulatory agencies are researching the causes of ADRs and proactive countermeasures. The World Health Organization (WHO) maintains VigiBase (https: / / who-umc.org / vigibase / ), a global drug safety information database that collects information on ADRs occurring worldwide. The U.S. Food and Drug Administration (FDA), a regulatory agency responsible for ensuring the safety and efficacy of drugs, operates a database called the FDA Adverse Event Reporting System (FAERS) to collect information on ADRs reported to the FDA. SIDER (Side Effect Resource, http: / / sideeffects.embl.de / about / ), managed by researchers at the University of Copenhagen in Denmark, is a database containing information on currently marketed drugs and their adverse reactions. In Korea, the Ministry of Food and Drug Safety operates KAERS (https: / / kaers.drugsafe.or.kr / ). This database of various adverse drug reactions provides standardized information on adverse drug reactions.
[0003] Recently, research utilizing artificial intelligence has been conducted in the field of adverse drug reactions, and the mainly utilized technologies include machine learning, deep learning, and natural language processing. According to a study that searched papers published from January 1, 2015, to July 9, 2021, and reviewed those containing key terms related to artificial intelligence and drug safety in the titles or abstracts, the main areas of application among the searched papers included identification of adverse effects and drug reactions (57.6%), safety report processing (21.2%), extraction of drug-drug interactions (7.6%), identification of populations at high risk of drug toxicity or personalized treatment guidelines (7.6%), prediction of adverse effects (3.0%), simulation of clinical trials (1.5%), and incorporation of prediction uncertainty in diagnostic classifiers. It is expected that the use of artificial intelligence in the field of adverse drug reactions and drug reaction prediction will further increase.
[0004] To proactively detect adverse drug reactions, information on medication use and subsequent reactions (diagnostic test data, disease diagnosis data, nursing records, etc.) are essential. A database converted to the Common Data Model standard using electronic medical records can collect data from multiple medical institutions. Furthermore, regularly updated databases allow for more efficient utilization of data from these institutions.
[0005] In this regard, Republic of Korea Patent Publication No. 2008-0042256 has been published.
[0006] The present disclosure is conceived in response to the aforementioned background technology, and aims to provide a method and device for active detection of adverse drug reactions using artificial intelligence technology.
[0007] The technical problems of the present disclosure are not limited to the technical problems mentioned above, and other technical problems not mentioned will be clearly understood by those skilled in the art from the description below.
[0008] According to one embodiment of the present disclosure, a method for generating a prediction result for an adverse drug reaction using artificial intelligence technology performed by a computing device is disclosed. The method may include: a step of constructing a periodic database scheduled to be updated in predetermined time period units using raw data for detecting adverse drug reactions; a step of determining a plurality of input variables for learning an artificial intelligence-based prediction model for detecting the adverse drug reaction based on data stored in the periodic database; and a step of training the prediction model using a learning dataset including the determined input variables and adverse reaction information for a specific drug and a specific disease, such that the prediction model outputs adverse reaction information for the specific drug and the specific disease in response to receiving the input variables.
[0009] In one embodiment, the step of constructing the periodic conversion database may include: a step of obtaining raw data for detecting the adverse drug reaction from an electronic medical record (EMR), wherein the raw data includes at least one of drug administration-related data, disease diagnosis data, diagnostic test data, or patient nursing record data; a step of converting terms of the raw data to correspond to a common data standard specification of the prediction model; and a step of constructing the periodic conversion database that is updated in predetermined time period units using at least one of a database linkage technique, a file linkage technique, or a change history classification linkage technique.
[0010] In one embodiment, the raw data may additionally be obtained from at least one of FAERS, KAERS, WHO-ART, SIDER or EU-ADR.
[0011] In one embodiment, the input variables may include at least one of a demographic variable, a drug-related variable, a diagnostic test-related variable, a nursing record-related variable, or a drug side effect diagnosis-related variable.
[0012] In one embodiment, the demographic variable includes at least one of gender, age, or region, the drug-taking related variable includes at least one of drug type, drug-taking period, drug-taking dosage, or drug-taking history, the diagnostic test-related variable includes at least one of a test through photography, a blood test, or a urine test, the nursing record-related variable is a variable obtained from a nursing record and obtained through at least one technique among optical character recognition (OCR), natural language processing (NLP), or text mining, and the drug-side effect diagnosis-related variable may include information related to the diagnosis of a disease caused by a drug-side effect.
[0013] In one embodiment, the input variables may be determined using at least one of a first technique for determining variable importance by using perturbed input data in which at least some of the variables have changed and output data output from the prediction model in response to the perturbed input data, a second technique for selecting input variables corresponding to principal components by transforming data in a high-dimensional space into a low-dimensional space, or a third technique for selecting input variables in a direction that reduces impurity on a decision tree.
[0014] In one embodiment, the input variables may be preprocessed based on: a step of handling outliers, a step of performing missing value imputation, a step of performing categorization or recategorization, and a step of performing normalization or standardization.
[0015] In one embodiment, the step of training the prediction model for detecting the adverse drug reaction may include: a step of determining the prediction model for detecting the adverse drug reaction from among a plurality of candidate models; and a step of training the prediction model so that the prediction model outputs whether the adverse drug reaction is for a specific drug or a specific disease, or the probability of occurrence of the adverse drug reaction for a specific drug or a specific disease, using the preprocessed input variable as input data.
[0016] In one embodiment, the prediction model may be a model selected from candidate models including: Random Forest, Logistic Regression, Support Vector Machine, Decision Tree, Naive Bayes, K-Nearest Neighbor, XGBoost, LightGBM, CatBoost, Convolutional Neural Network (CNN), Recurrent Neural Network (RNN), Long Short-Term Memory (LSTM), Transformer, and Generative Pretrained Transformer (GPT).
[0017] In one embodiment, the prediction model may include: obtaining output results from each of the candidate models using a predetermined test dataset; performing a performance evaluation on the output results from each of the candidate models; and determining one of the candidate models as the prediction model based on the result of the performance evaluation.
[0018] In one embodiment, the adverse reaction information for the specific drug and the specific disease may include: first information indicating whether there is an adverse drug reaction for the specific drug, second information indicating the probability of occurrence of an adverse drug reaction for the specific drug, third information indicating whether there is an adverse drug reaction for the specific disease, and fourth information indicating the probability of occurrence of an adverse drug reaction for the specific disease.
[0019] In one embodiment, the method may include a step of transmitting an output result of the prediction model for the adverse drug reaction to a user terminal.
[0020] In one embodiment, the output result transmitted to the user terminal may include at least one output result of: a first result indicating a dual classification of an adverse drug reaction, classified as detected or not detected, a second result indicating a probability notation for an adverse drug reaction, or a third result indicating a three-level category notation linking a drug, an adverse reaction, and a risk.
[0021] In one embodiment, the output result from the prediction model can be used to retrain the prediction model based on the accuracy evaluated for the output result.
[0022] In one embodiment, feedback information corresponding to the output result transmitted from a user terminal that has received the output result from the prediction model can be used to retrain the prediction model.
[0023] In one embodiment, a computer program stored in a computer-readable storage medium is disclosed. The computer program, when executed by a computing device, causes the computing device to perform operations for generating a prediction result for an adverse drug reaction using artificial intelligence technology, the operations including: constructing a periodic database scheduled to be updated in predetermined time intervals using raw data for detecting adverse drug reactions; determining, based on data stored in the periodic database, a plurality of input variables for learning an artificial intelligence-based prediction model for detecting the adverse drug reaction; and training the prediction model using a learning dataset including the determined input variables and adverse reaction information for a specific drug and a specific disease, such that the prediction model outputs adverse reaction information for the specific drug and the specific disease in response to receiving the input variables.
[0024] A computing device according to one embodiment is disclosed. The computing device may include at least one processor and a memory. The at least one processor may perform the following operations: constructing a periodic database scheduled to be updated in predetermined time period units using raw data for detecting adverse drug reactions; determining, based on data stored in the periodic database, a plurality of input variables for learning an artificial intelligence-based predictive model for detecting adverse drug reactions; and training the predictive model using the determined input variables and a learning dataset including adverse reaction information for a specific drug and a specific disease, such that the predictive model outputs adverse reaction information for the specific drug and the specific disease in response to receiving the input variables.
[0025] A method for generating prediction results for adverse drug reactions using artificial intelligence technology performed by a computing device according to one embodiment is disclosed. The method may include: obtaining input data for adverse drug reaction detection based on data stored in a periodic database scheduled to be updated at predetermined time intervals using raw data for adverse drug reaction detection; and obtaining adverse reaction information for a specific drug and a specific disease from the input data using an artificial intelligence-based prediction model.
[0026] The method and device according to one embodiment of the present disclosure can be used universally without distinction between inpatient / outpatient or specific drugs, and can contribute to the management of adverse drug reactions and the creation of a virtuous cycle in the field of drug safety.
[0027] FIG. 1 schematically illustrates a block diagram of a computing device according to one embodiment of the present disclosure.
[0028] FIG. 2 illustrates an exemplary structure of an artificial intelligence-based model according to one embodiment of the present disclosure.
[0029] FIG. 3 illustrates an example of a method for generating output results of a model including predictions for adverse drug reactions according to one embodiment of the present disclosure.
[0030] FIG. 4 illustrates a flowchart of constructing a periodic conversion database including the principles of periodic data loading and change history table loading of an electronic medical record and a periodic conversion database according to one embodiment of the present disclosure.
[0031] FIG. 5 illustrates the concept of linking files for constructing a periodic conversion database according to one embodiment of the present disclosure.
[0032] FIG. 6 illustrates an overall process leading to database construction, development of artificial intelligence for detecting adverse drug reactions, and notification of abnormality detection results according to one embodiment of the present disclosure.
[0033] FIG. 7 is a schematic diagram of a computing environment according to one embodiment of the present disclosure.
[0034] Various embodiments are described with reference to the drawings. In this specification, various descriptions are provided to provide an understanding of the present disclosure. Before describing specific details for implementing the present disclosure, it should be noted that components not directly related to the technical spirit of the present disclosure have been omitted to the extent that they do not obscure the technical spirit of the present invention. Furthermore, terms or words used in this specification and claims should be interpreted with meanings and concepts consistent with the technical spirit of the present invention, based on the principle that the inventor can define the concept of appropriate terms to best describe his or her invention.
[0035] As used herein, the terms "component," "module," "system," and "part" refer to computer-related entities, hardware, firmware, software, a combination of software and hardware, or an execution of software, and may be used interchangeably. For example, a component may be, but is not limited to, a procedure running on a processor, a processor, an object, a thread of execution, a program, and / or a computer. For example, both an application running on a computing device and the computing device may be a component. One or more components may reside within a processor and / or a thread of execution. A component may be localized within a single computer. A component may be distributed between two or more computers. Furthermore, these components may execute from various computer-readable media having various data structures stored therein. Components may communicate via local and / or remote processes, for example, by signals comprising one or more data packets (e.g., data from one component interacting with another component in a local system, a distributed system, and / or data transmitted via signals to another system over a network such as the Internet).
[0036] Furthermore, the term "or" is intended to mean an inclusive "or" rather than an exclusive "or." That is, unless otherwise specified or clear from context, "X employs A or B" is intended to mean either of the natural inclusive permutations. That is, if X employs A; X employs B; or X employs both A and B, "X employs A or B" can apply to any of these cases. Furthermore, the term "and / or" as used herein should be understood to refer to and include all possible combinations of one or more of the associated items listed.
[0037] Additionally, the terms "comprises" and / or "comprising" should be understood to imply the presence of the features and / or components in question. However, it should be understood that the terms "comprises" and / or "comprising" do not exclude the presence or addition of one or more other features, components, and / or groups thereof. Furthermore, unless otherwise specified or clear from the context to refer to the singular form, the singular in the specification and claims should generally be construed to mean "one or more."
[0038] And, the term “at least one of A or B” or “at least one of A and B” should be interpreted to mean “if it includes only A”, “if it includes only B”, or “if it is combined in the composition of A and B”.
[0039] Those skilled in the art should further appreciate that the various illustrative logical components, blocks, modules, circuits, means, logics, and algorithms described in connection with the embodiments disclosed herein may be implemented as electronic hardware, computer software, or combinations of both. To clearly illustrate the interchangeability of hardware and software, various illustrative components, blocks, means, logics, modules, circuits, and steps have been described above generally in terms of their functionality. Whether such functionality is implemented as hardware or software depends upon the particular application and design constraints imposed on the overall system. Skilled artisans may implement the described functionality in varying ways for each particular application. However, such implementation decisions should not be interpreted as causing a departure from the scope of the present disclosure.
[0040] The description of the disclosed embodiments is provided to enable a person skilled in the art to make or use the present invention. Various modifications to these embodiments will be apparent to those skilled in the art. The general principles defined herein may be applied to other embodiments without departing from the scope of the present disclosure. Therefore, the present invention is not limited to the embodiments disclosed herein. The present invention is to be construed in the widest scope consistent with the principles and novel features disclosed herein.
[0041] In the present disclosure, terms expressed as N, such as first, second, or third, are used to distinguish at least one entity. For example, entities expressed as first and second may be the same or different from each other.
[0042] The term “adverse drug reaction (ADR)” used in this disclosure refers to a harmful and unintended reaction that occurs when a medicine, etc. is administered or used, and in cases where a causal relationship with the medicine, etc. cannot be ruled out, and may include not only adverse reactions occurring at normal doses, but also adverse drug reactions and withdrawal symptoms that occur when a drug is intentionally or accidentally overdosed or abused, and cases where the expected pharmacological effect does not appear.
[0043] “Adverse drug reaction monitoring” refers to activities to establish rational drug use and prevent damage related to drugs in advance by quickly and systematically collecting or evaluating various adverse events that occur when using drugs, taking countermeasures, and delivering safety information and response results to medical professionals, consumers, etc.
[0044] Examples of ADRs include side effects and adverse drug events (ADEs). An adverse drug reaction (ADD) is the opposite of the principal action, which occurs when a drug is used for a specific purpose. It refers to any unintended effect that occurs when a drug is administered at normal dosages. As a term encompassing all actions other than those intended for therapeutic purposes, it is a concept that can be actively utilized in a variety of fields that encompass the study of all actions, regardless of their harmfulness.
[0045] An adverse drug event (ADE) is an undesirable and unintended sign, symptom, or disease that occurs during the administration or use of a drug, regardless of a causal relationship with the drug.
[0046] Adverse drug reactions (ADRs) are harmful and unintended reactions that occur during the administration or use of medicines, etc. according to normal usage, and in cases where a causal relationship with the medicine cannot be ruled out. Since ADRs are predictable and preventable, they are a concept for which management by pharmacists and medical staff is important.
[0047] Drug metabolism occurs mostly in the liver, where metabolism reactions such as oxidation / reduction, hydrolysis, hydration, conjugation, condensation, or isomerization of drugs occur in the endoplasmic reticulum of hepatocytes by a metabolic enzyme system called cytochrome P450 (CYP450). The four main parameters of pharmacokinetics generally include absorption, distribution, metabolism, and excretion. Dose, frequency, route of administration, tissue distribution, and protein binding of drugs to receptors can all affect drug metabolism. In addition, pathological factors including diseases of the digestive system, cardiopulmonary system, respiratory system, and renal excretory system can also affect drug metabolism.
[0048] The embodiments in the present disclosure may include a step of deriving predictable ADRs in order of causality (e.g., certain, probable, possible, unlikely, conditional) based on the type of drug administered for an indication, demographic background, race / individual genotype / phenotype, history, etc., classification based on the degree of reaction (e.g., mild, moderate, severe), or frequency of occurrence.
[0049] FIG. 1 schematically illustrates a block diagram of a computing device (100) according to one embodiment of the present disclosure.
[0050] A computing device (100) according to one embodiment of the present disclosure may include a processor (110) and a memory (130).
[0051] The configuration of the computing device (100) illustrated in FIG. 1 is merely a simplified example. In one embodiment of the present disclosure, the computing device (100) may include other configurations for performing the computing environment of the computing device (100), and only some of the disclosed configurations may constitute the computing device (100).
[0052] The computing device (100) in the present disclosure may refer to any type of node that constitutes a system for implementing embodiments of the present disclosure. The computing device (100) may refer to any type of user terminal or any type of server. The components of the computing device (100) described above are exemplary, and some may be excluded or additional components may be included. For example, if the computing device (100) described above includes a user terminal, an output unit (not shown) and an input unit (not shown) may be included within the scope of the computing device (100).
[0053] The computing device (100) in the present disclosure can perform technical features according to embodiments of the present disclosure to be described later. For example, the computing device (100) can use an artificial intelligence-based predictive model that utilizes input data corresponding to a patient's medical history, drug allergy history, administered drugs, and administered dosages, etc., to generate a predictive result including adverse drug reactions (ADRs) that may occur in a patient corresponding to the input data (e.g., a specific drug triggers excessive immune system activation in a specific individual, inducing a cellular immune response cascade).
[0054] As an additional example, the computing device (100) may obtain information on the patient's demographic information, medication history, medical history, and blood test data obtained from a database (e.g., 65 years old, Asian male, use of serotonin reuptake inhibitors (SSRIs), divalproate, and statin drugs for the past 26 months, experience of abdominal pain and constipation as adverse drug reactions), and generate an ADR prediction result corresponding to the drug or a specific disease based on the obtained information, thereby providing corresponding data including a cost benefit analysis of drug use.
[0055] As an additional example, the computing device (100) can more efficiently collect raw data from multiple medical institutions by constructing a periodic conversion database converted to the specifications of the common data model using electronic medical records. The computing device (100) can be configured to perform more up-to-date detection of adverse drug reactions by utilizing the regularly updated periodic conversion database.
[0056] In one embodiment of the present disclosure, the computing device (100) can obtain the results of performing base sequence analysis (e.g., Next Generation Sequencing) from a server or an external entity, etc. In another embodiment, the computing device (100) can also perform base sequence analysis on proteins or genes (e.g., DNA or RNA) obtained from a biological sample derived from a subject. As used herein, the term base sequence analysis can be performed by any type of technique capable of analyzing the sequence of bases, and can include, but is not limited to, whole genome sequencing, whole exome sequencing, or whole transcriptome sequencing.
[0057] As used herein, the term "subject" may mean an object or individual for obtaining a biological sample including saliva, hair root, proteins, genes and / or combinations thereof.
[0058] As used herein, the term "sample" refers to a sample obtained from a subject or individual whose genetic genome and / or genetic phenotype is to be determined, and may be used without limitation, for example, cells or tissues obtained by biopsy, etc., blood, whole blood, serum, plasma, saliva, cerebrospinal fluid, various secretions, urine and / or feces, etc. Preferably, the sample may be selected from the group consisting of blood, plasma, serum, saliva, nasal fluid, sputum, ascites, vaginal secretions and / or urine, and more preferably, blood, plasma or serum. The sample may be preprocessed before being used for detection or diagnosis. For example, the preprocessing method may include homogenization, filtration, distillation, extraction, concentration, inactivation of interfering components, and / or addition of reagents. In the present disclosure, the biological sample may be, but is not limited to, tissue, cell, whole blood, and / or blood.
[0059] In one embodiment, the processor (110) may be configured with at least one core and may include a processor for data analysis and / or processing, such as a central processing unit (CPU), a general purpose graphics processing unit (GPGPU), or a tensor processing unit (TPU) of the computing device (100).
[0060] The processor (110) can read a computer program stored in the memory (130) to generate a prediction result including the occurrence and / or probability of an adverse drug event (ADE), according to one embodiment of the present disclosure.
[0061] According to one embodiment of the present disclosure, the processor (110) may perform operations for learning a neural network. The processor (110) may perform calculations for learning a neural network, such as processing input data for learning in deep learning (DL), extracting features from the input data, calculating errors, and updating weights of the neural network using backpropagation. At least one of the CPU, GPGPU, and TPU of the processor (110) may process learning of a network function. For example, the CPU and GPGPU may together process learning of a network function and classification of data using the network function. In addition, in one embodiment of the present disclosure, processors of a plurality of computing devices may be used together to process learning of a network function and classification of data using the network function. In addition, a computer program executed in a computing device according to one embodiment of the present disclosure may be a CPU, GPGPU, or TPU executable program.
[0062] Additionally, the processor (110) can typically process the overall operation of the computing device (100). For example, the processor (110) can process data, information, or signals input or output through components included in the computing device (100) or run application programs stored in a storage unit, thereby providing or processing appropriate information or functions to the user.
[0063] According to one embodiment of the present disclosure, the memory (130) can store any form of information generated or determined by the processor (110) and any form of information received by the computing device (100). According to one embodiment of the present disclosure, the memory (130) can be a storage medium that stores computer software that causes the processor (110) to perform operations according to embodiments of the present disclosure. Accordingly, the memory (130) can mean computer-readable media for storing software codes required to perform embodiments of the present disclosure, data that is the target of execution of the codes, and execution results of the codes.
[0064] According to one embodiment of the present disclosure, the memory (130) may mean any type of storage medium. For example, the memory (130) may include at least one type of storage medium among a flash memory type, a hard disk type, a multimedia card micro type, a card type memory (e.g., SD or XD memory, etc.), a random access memory (RAM), a static random access memory (SRAM), a read-only memory (ROM), an electrically erasable programmable read-only memory (EEPROM), a programmable read-only memory (PROM), a magnetic memory, a magnetic disk, and an optical disk. The computing device (100) may also operate in relation to web storage that performs the storage function of the memory (130) on the Internet. The description of the memory described above is merely an example, and the memory (130) used in the present disclosure is not limited to the examples described above.
[0065] The communication unit (not shown) in the present disclosure may be configured regardless of the communication mode, such as wired or wireless, and may be configured with various communication networks, such as a personal area network (PAN) and a wide area network (WAN). In addition, the communication unit may operate based on the known World Wide Web (WWW), and may also utilize a wireless transmission technology used for short-distance communication, such as infrared (IrDA: Infrared Data Association) or Bluetooth.
[0066] The computing device (100) in the present disclosure may include any type of user terminal and / or any type of server. Accordingly, embodiments of the present disclosure may be performed by a server and / or a user terminal.
[0067] A user terminal may include any type of terminal capable of interacting with a server or other computing device. User terminals may include, for example, mobile phones, smart phones, laptop computers, personal digital assistants (PDAs), slate PCs, tablet PCs, and ultrabooks. A server may include any type of computing system or computing device, such as, for example, a microprocessor, a mainframe computer, a digital processor, a handheld device, or a device controller.
[0068] In a further embodiment, the aforementioned server may also refer to an entity that stores and manages medication administration information and responses to administration, electronic medical records (EMR), diagnostic test data, disease diagnosis data, nursing records, etc. The server can more efficiently collect medical data from various medical institutions through a database converted to the specifications of a common data model, and may include a storage unit (not shown) for storing related information so that the latest adverse drug reactions can be detected by utilizing a regularly updated database. For example, the storage unit may be included within the server or under the management of the server. In another example, the storage unit may be implemented in a form that exists outside the server and can communicate with the server. In this case, the storage unit may be managed and controlled by an external server different from the server.
[0069] FIG. 2 illustrates an exemplary structure of an artificial intelligence-based model according to one embodiment of the present disclosure.
[0070] Throughout this specification, the terms predictive model, artificial intelligence-based predictive model, artificial intelligence model, artificial intelligence-based model, computational model, neural network, network function, and neural network may be used interchangeably.
[0071] A neural network can be composed of a set of interconnected computational units, generally referred to as nodes. These nodes can also be referred to as neurons. A neural network consists of at least one node. The nodes (or neurons) that make up a neural network can be interconnected by one or more links.
[0072] Within a neural network, one or more nodes connected via links can form a relationship between input nodes and output nodes. The concept of input nodes and output nodes is relative, meaning that any node that is in an output node relationship with one node can also be in an input node relationship with another node, and vice versa. As described above, the relationship between input nodes and output nodes can be created based on links. One input node can be connected to one or more output nodes via links, and vice versa.
[0073] In a relationship between input nodes and output nodes connected through a single link, the data of the output node can have its value determined based on the data input to the input node. Here, the link interconnecting the input nodes and output nodes can have a weight. The weight can be variable and can be varied by the user or an algorithm so that the neural network can perform a desired function. For example, when one or more input nodes are interconnected to one output node through each link, the output node can determine the output node value based on the values input to the input nodes connected to the output node and the weight set on the link corresponding to each input node.
[0074] As described above, a neural network is a network in which one or more nodes are interconnected through one or more links, forming input and output node relationships within the network. The characteristics of a neural network can be determined based on the number of nodes and links within the network, the relationships between the nodes and links, and the weights assigned to each link. For example, if two neural networks have the same number of nodes and links but different weight values for the links, the two neural networks can be perceived as different from each other.
[0075] A neural network can be composed of a set of one or more nodes. A subset of the nodes comprising the neural network can form a layer. Some of the nodes comprising the neural network can form a layer based on their distances from the initial input node. For example, a set of nodes that are n distances from the initial input node can form n layers. The distance from the initial input node can be defined by the minimum number of links required to reach the node from the initial input node. However, this definition of a layer is arbitrary for illustrative purposes, and the degree of a layer within a neural network can be defined in a different way than described above. For example, a layer of nodes can be defined by its distance from the final output node.
[0076] In one embodiment of the present disclosure, a set of neurons or nodes may be defined as a layer.
[0077] An initial input node may refer to one or more nodes within a neural network into which data is directly input without going through links with other nodes. Alternatively, within a neural network, it may refer to nodes that do not have other input nodes connected by links in the relationship between nodes based on links. Similarly, a final output node may refer to one or more nodes within a neural network that do not have output nodes in their relationship with other nodes. Furthermore, a hidden node may refer to nodes that constitute a neural network other than the initial input node and the final output node.
[0078] A neural network according to one embodiment of the present disclosure may be a neural network in which the number of nodes in an input layer may be the same as the number of nodes in an output layer, and the number of nodes decreases and then increases as it progresses from the input layer to a hidden layer. In addition, a neural network according to another embodiment of the present disclosure may be a neural network in which the number of nodes in an input layer may be less than the number of nodes in an output layer, and the number of nodes decreases as it progresses from the input layer to the hidden layer. In addition, a neural network according to another embodiment of the present disclosure may be a neural network in which the number of nodes in an input layer may be greater than the number of nodes in an output layer, and the number of nodes increases as it progresses from the input layer to the hidden layer. A neural network according to another embodiment of the present disclosure may be a neural network in the form of a combination of the above-described neural networks.
[0079] A deep neural network (DNN) can refer to a neural network that includes multiple hidden layers in addition to input and output layers. Using a DNN, one can identify latent structures in data. This includes gene sequence structures, amino acid sequence structures, protein sequence structures, causality between drugs and ADRs and / or the probability of association between symptoms and ADRs, and the latent structures of photos, text, videos, and audio (e.g., what objects are in the photo, what the content and emotion of the text are, what the content and emotion of the audio are, etc.). A DNN can include a convolutional neural network (CNN), a recurrent neural network (RNN), an autoencoder, a generative adversarial network (GAN), a restricted Boltzmann machine (RBM), a deep belief network (DBN), a Q-network, a U-network, and a Siamese network. The description of the deep neural network described above is only an example, and the present disclosure is not limited thereto.
[0080] For example, the artificial intelligence-based prediction model of the present disclosure may include a Random Forest, a Logistic Regression, a Support Vector Machine, a Decision Tree, Naive Bayes, a K-Nearest Neighbor, XGBoost, LightGBM, CatBoost, a Recurrent Neural Network (RNN), a Long Short Term Memory (LSTM) network, a Bidirectional Long Short Term Memory (BiLSTM) network, a Transformer, a Generative Pre-trained Transformer (GPT), a Bidirectional Encoder Representations from Transformers (BERT), a SpanBERT, a Gated Recurrent Unit (GRU), or a Bidirectional Gated Recurrent Unit (BiGRU).
[0081] The artificial intelligence-based prediction model of the present disclosure can be expressed by a network structure of any structure described above, including an input layer, a hidden layer, and an output layer.
[0082] The neural network that can be used in the artificial intelligence-based model of the present disclosure can be trained by at least one of supervised learning, unsupervised learning, semi-supervised learning, reinforcement learning, federated learning for distributed deep learning, and incremental learning. Training of the neural network can be a process of applying knowledge to the neural network to perform a specific operation. For example, a predictive model according to an embodiment of the present disclosure can be trained by a semi-supervised learning method that applies a mask to the patient's genetic genomic information, medication, and symptoms or at least some of the symptoms indicating potential adverse drug reactions, and then matches the masked symptoms (e.g., adverse drug reactions). At this time, at least some of the administered drugs and potential adverse drug reactions may be used as training data for the prediction model as labeled data, and the remaining some may be used as training data for the prediction model as unlabeled data.
[0083] Neural networks can be trained to minimize output errors. This process involves repeatedly inputting training data into the neural network, calculating the neural network output and target error for the training data, and backpropagating the neural network error from the output layer to the input layer to update the weights of each node in the neural network to reduce the error. In supervised learning, training data with the correct answer for each training data is used (i.e., labeled training data). In unsupervised learning, the correct answer may not be labeled for each training data. For example, in supervised learning for data classification, the training data may be data with each category labeled. Labeled training data is input into the neural network, and the error can be calculated by comparing the output (category) of the neural network with the labels of the training data.
[0084] As another example, in unsupervised learning for data classification, errors can be calculated by comparing input training data with the output of a neural network. The calculated errors are backpropagated through the neural network (i.e., from the output layer to the input layer), and the connection weights of each node in each layer of the neural network can be updated based on the learning rate. The amount of change in the updated connection weights of each node can be determined by the learning rate. The neural network's calculation of the input data and backpropagation of the error can constitute a learning cycle (epoch). The learning rate can be applied differently depending on the number of iterations of the neural network's learning cycle. For example, a high learning rate can be used early in the neural network's training process to quickly achieve a certain level of performance, thereby improving efficiency. A lower learning rate can be used later in the training process to improve accuracy.
[0085] In neural network training, training data can typically be a subset of real-world data (i.e., the data to be processed using the trained neural network). Therefore, there can be a learning cycle where errors on the training data decrease but errors on the real-world data increase. Overfitting is a phenomenon where excessive training on the training data leads to increased errors on the real-world data. For example, a neural network trained on yellow cats may fail to recognize cats when shown non-yellow colors, a type of overfitting. Overfitting can increase errors in machine learning algorithms. Various optimization methods can be used to prevent overfitting. These methods include increasing the training data, regularization, dropout, which disables some nodes in the network during the learning process, and utilizing batch normalization layers.
[0086] A computer-readable medium storing a data structure according to one embodiment of the present disclosure is disclosed. The aforementioned data structure can be stored in a storage unit in the present disclosure, executed by a processor, and transmitted and received by a communication unit.
[0087] A data structure can refer to the organization, management, and storage of data to enable efficient access and modification. A data structure can refer to the organization of data to solve a specific problem (e.g., data retrieval, data storage, or data modification in the shortest possible time). A data structure can also be defined as the physical or logical relationships between data elements designed to support specific data processing functions. Logical relationships between data elements can include connections between user-defined data elements. Physical relationships between data elements can include actual relationships between data elements physically stored on a computer-readable storage medium (e.g., persistent storage). Specifically, a data structure can include a collection of data, relationships between data, and functions or commands applicable to the data. An effectively designed data structure allows a computing device to perform operations while minimizing the use of its resources. Specifically, a computing device can improve the efficiency of operations, reading, inserting, deleting, comparing, exchanging, and searching through an effectively designed data structure.
[0088] Data structures can be categorized as linear or nonlinear, depending on their form. A linear data structure can be a structure in which only one data item is linked to the next. Linear data structures can include lists, stacks, queues, and deques. A list can refer to a series of data sets with an internal order. Lists can also include linked lists. A linked list is a data structure in which data is linked in a single line, each item having a pointer. In a linked list, a pointer can contain information about the next or previous item. Linked lists can be expressed as singly linked lists, doubly linked lists, or circular linked lists, depending on their form. A stack can be a data listing structure with limited data access. A stack can be a linear data structure in which data operations (e.g., insertion or deletion) can only be performed at one end of the data structure. Data stored in a stack can be a Last-in-First-out (LIFO) data structure. A queue is a data structure with limited access to data. Unlike a stack, it can be a first-in, first-out (FIFO) data structure, with later data being retrieved later. A deck can be a data structure that can process data at both ends.
[0089] A nonlinear data structure can be a structure in which multiple pieces of data are connected behind a single piece of data. Nonlinear data structures can include graph data structures. A graph data structure can be defined by vertices and edges, and an edge can include a line connecting two different vertices. A graph data structure can include a tree data structure. A tree data structure can be a data structure in which there is only one path connecting two different vertices among multiple vertices included in the tree. In other words, it can be a data structure that does not form a loop in a graph data structure.
[0090] Throughout this specification, the terms predictive model, artificial intelligence-based model, computational model, neural network, network function, and neural network may be used interchangeably. Hereinafter, the term neural network will be used interchangeably. The data structure may include a neural network. Furthermore, the data structure including the neural network may be stored on a computer-readable medium. The data structure including the neural network may also include preprocessed data for processing by the neural network, data input to the neural network, weights of the neural network, hyperparameters of the neural network, data obtained from the neural network, activation functions associated with each node or layer of the neural network, a loss function for learning the neural network, and the like. The data structure including the neural network may include any of the components described above. That is, the data structure including the neural network may be configured to include all or any combination of preprocessed data for processing by the neural network, data input to the neural network, neural network weights, neural network hyperparameters, data obtained from the neural network, activation functions associated with each node or layer of the neural network, and loss functions for learning the neural network. In addition to the aforementioned configurations, the data structure including the neural network may include any other information that determines the characteristics of the neural network. In addition, the data structure may include all forms of data used or generated in the computational process of the neural network, and is not limited to the aforementioned. The computer-readable medium may include a computer-readable recording medium and / or a computer-readable transmission medium. The neural network may be configured as a set of interconnected computational units, which may generally be referred to as nodes. These nodes may also be referred to as neurons. The neural network is configured to include at least one node.
[0091] The data structure may include data input to a neural network. The data structure including the data input to the neural network may be stored on a computer-readable medium. The data input to the neural network may include training data input during the neural network training process and / or input data input to the neural network after training has been completed. The data input to the neural network may include data that has undergone preprocessing and / or data that is the target of preprocessing. Preprocessing may include a data processing process for inputting data to the neural network. Accordingly, the data structure may include data that is the target of preprocessing and data generated by the preprocessing. The above-described data structure is merely an example, and the present disclosure is not limited thereto.
[0092] The data structure may include weights of a neural network (in this specification, weights and parameters may be used interchangeably). And the data structure including the weights of the neural network may be stored in a computer-readable medium. The neural network may include a plurality of weights. The weights may be variable and may be varied by a user or an algorithm so that the neural network can perform a desired function. For example, when one or more input nodes are interconnected to one output node by respective links, the output node may determine a data value output from the output node based on the values input to the input nodes connected to the output node and the weights set for the links corresponding to each input node. The above-described data structure is merely an example, and the present disclosure is not limited thereto.
[0093] By way of example and not limitation, the weights may include weights that vary during the neural network training process and / or weights that have completed neural network training. The weights that vary during the neural network training process may include weights at the start of the training cycle and / or weights that vary during the training cycle. The weights that have completed neural network training may include weights that have completed the training cycle. Accordingly, a data structure including the weights of a neural network may include a data structure including weights that vary during the neural network training process and / or weights that have completed neural network training. Therefore, the above-described weights and / or combinations of each weight are included in the data structure including the weights of a neural network. The above-described data structures are merely examples and the present disclosure is not limited thereto.
[0094] A data structure including neural network weights can be stored in a computer-readable storage medium (e.g., memory, hard disk) after going through a serialization process. Serialization can be a process of converting a data structure into a form that can be stored on the same or different computing devices and later reconstructed and used. A computing device can serialize the data structure to transmit and receive data over a network. The serialized data structure including neural network weights can be reconstructed on the same computing device or another computing device through deserialization. The data structure including neural network weights is not limited to serialization. Furthermore, the data structure including neural network weights can include a data structure that increases computational efficiency while minimizing the use of computing device resources (e.g., a B-Tree, an R-Tree, a Trie, an m-way search tree, an AVL tree, a Red-Black Tree in nonlinear data structures). The foregoing is merely an example, and the present disclosure is not limited thereto.
[0095] The data structure may include hyperparameters of a neural network. Furthermore, the data structure including the hyperparameters of the neural network may be stored on a computer-readable medium. The hyperparameters may be variables that can be changed by the user. The hyperparameters may include, for example, a learning rate, a cost function, the number of learning cycle iterations, weight initialization (e.g., setting a range of weight values to be initialized), and the number of hidden units (e.g., the number of hidden layers, the number of nodes in the hidden layer). The above-described data structure is merely an example, and the present disclosure is not limited thereto.
[0096] A transformer may be considered as a network function for a prediction model according to one embodiment of the present disclosure. For example, the prediction model may be operated based on a transformer. Such a prediction model may be operated using, for example, a recurrent neural network with an attention algorithm or a transformer with an attention algorithm.
[0097] In one embodiment, a transformer may be composed of an encoder that encodes embedded data and a decoder that decodes the encoded data. The transformer may have a structure that receives a series of data, performs encoding and decoding steps, and outputs a series of data of different types. In one embodiment, the series of data may be processed into a form operable by the transformer. The process of processing the series of data into a form operable by the transformer may include an embedding process. Expressions such as data tokens, embedding vectors, and embedding tokens may refer to data embedded in a form operable by the transformer.
[0098] To encode and decode a series of data, a transformer can process the encoders and decoders within the transformer using an attention algorithm. An attention algorithm can refer to an algorithm that, for a given query, calculates the similarity for one or more keys, reflects this similarity in the values corresponding to each key, and then weights and adds the values to calculate an attention value.
[0099] Depending on how the query, key, and value are set, various types of attention algorithms can be categorized. For example, if attention is obtained by setting the query, key, and value all to the same value, this could be a self-attention algorithm. If attention is obtained by reducing the dimensionality of the embedding vector to process a series of input data in parallel and then generating individual attention heads for each segmented embedding vector, this could be a multi-head attention algorithm.
[0100] In one embodiment, the transformer may be composed of modules that perform multiple multi-head self-attention algorithms or multi-head encoder-decoder algorithms. In one embodiment, the transformer may also include additional components other than attention algorithms, such as embedding layers, normalization layers, and softmax layers. Methods for constructing a transformer using attention algorithms may include methods disclosed in Vaswani et al., Attention Is All You Need, 2017 NIPS, which is incorporated herein by reference.
[0101] A transformer can be applied to various data domains, such as embedded natural language, embedded sequence information, segmented image data, and audio waveforms, to transform a series of input data into a series of output data. To transform data with various data domains into a series of data that can be input to a transformer, the transformer can embed the data. The transformer can process additional data expressing the relative positional relationship or phase relationship between a series of input data. Alternatively, vectors expressing the relative positional relationship or phase relationship between the input data can be additionally reflected in the series of input data to embed the series of input data. In one example, the relative positional relationship between a series of input data may include, but is not limited to, word order within a natural language sentence, the relative positional relationship between each segmented image, and the time order of segmented audio waveforms. The process of adding information expressing the relative positional relationship or phase relationship between a series of input data may be referred to as positional encoding.
[0102] An example of how to embed data into a transformer is disclosed in Dosovitskiy, et al., AN IMAGE IS WORTH 16X16 WORDS: TRANSFORMERS FOR IMAGE RECOGNITION AT SCALE, which is incorporated herein by reference.
[0103] The computing device and computing apparatus in the present disclosure may be used interchangeably.
[0104] FIG. 3 illustrates an example of a method for generating output results of a model including predictions for adverse drug reactions according to one embodiment of the present disclosure.
[0105] Embodiments of the present disclosure may encompass embodiments for training a prediction model and embodiments for actually obtaining inference (e.g., predictions including possible adverse drug reactions and / or their probability of occurrence) results through the trained prediction model.
[0106] In this disclosure, the operation of a prediction model may be described as either a process of training a prediction model or a process of inferring adverse drug reactions using the prediction model (e.g., a process of generating prediction results). This description is provided for convenience of explanation. Even if it is described as either a process of performing learning or a process of performing inference, it should be interpreted as encompassing either the learning or inference process.
[0107] In one embodiment, the steps illustrated in FIG. 3 may be performed by a computing device (100). In a further embodiment, the steps illustrated in FIG. 3 may be implemented by multiple entities, such as some of the steps illustrated in FIG. 3 being performed on a user terminal and others being performed on a server.
[0108] In one embodiment according to the present disclosure, a computing device (100) can build a periodic database scheduled to be updated in predetermined time period units using raw data for detecting adverse drug reactions (S310).
[0109] In one embodiment, raw data may refer to data acquired prior to being databased. For example, raw data may be used to refer to data acquired from various channels, such as electronic medical records.
[0110] In one embodiment, the raw data may be obtained from at least one of FAERS, KAERS, WHO-ART, SIDER or EU-ADR.
[0111] In one embodiment, raw data may be obtained from an electronic medical record (EMR). The raw data may include at least one of medication-related data, disease diagnosis data, diagnostic test data, or nursing record data. For example, diagnostic test data may refer to data obtained through any type of test or diagnosis. For example, diagnostic test data may include genetic test records. Specific examples of diagnostic test data, medication-related data, disease diagnosis data, and nursing record data will be described below.
[0112] In one embodiment, the computing device (100) may receive a list of possible adverse drug reactions (ADRs) from an external rule-based or artificial intelligence-based predictive model that is different from the predictive model according to the present disclosure.
[0113] In one embodiment, the provided predictable adverse drug reactions (ADRs) may be sorted based on specific criteria (e.g., causality and / or degree of reaction such as Mild, Moderate, Severe, etc.) based on the raw data, and the computing device (100) may further process the provided ADRs into output data including, in addition, the 'probability of occurrence' of the ADRs.
[0114] In one embodiment, the computing device (100) can transform terms contained in raw data to correspond to common data standard specifications of a database or predictive model. For example, dystonia, athetosis, choreoathetosis, and chorea can be defined as similar or identical ADRs under the same category, and the transformation can be provided with a clear distinction from causes other than drug-related, as distinguished from such abnormal movements suspected to be drug side effects. For example, the computing device (100) can classify cerebral palsy, focal intracranial diseases, Huntington's disease, hepatic encephalopathy, hepatic cirrhosis, etc., which are different from the above-mentioned athetosis, as diseases or ADRs distinct from the above-mentioned dystonia, athetosis, choreoathetosis, and chorea. As an additional example, if the cause of the above-mentioned abnormal movement is not specifically described in the nursing record data, the computing device (100) may not use the data as learning data for predicting adverse drug reactions, or may classify the term as an ADR with low causality with the drug.
[0115] In one embodiment of the present disclosure, a periodic database may refer to a data repository configured to store and update information such as patient demographic information, medications and indications administered, adverse drug reactions (ADRs), and medical history at optimal time intervals determined by a predetermined or predictive model. The periodic database may refer to a data repository constructed by collecting all data regarding newly entered or additionally entered ADRs and patient information. For example, a user of the periodic database may search for a type of drug that causes a specific ADR—for example, seizures—and, if the user searches for the drug bupropion that the patient is taking, may also receive statistical summaries of other possible ADRs or clinical cases that showed seizure symptoms after taking the drug, along with the dosages and patient information. For example, an artificial intelligence-based model according to one embodiment of the present disclosure may include a model for determining a predetermined time period for which the periodic database is to be converted or updated. An artificial intelligence-based model that takes as input the amount of data stored in a database, the type of data, and / or the content of the data can be operated to output a periodic conversion point in a periodic conversion database.
[0116] In one embodiment, the computing device (100) can build a periodic database using at least one of a database linkage technique, a file linkage technique, or a change history distinction linkage technique.
[0117] In one embodiment, the database linking technique may mean an automated method of data linking that enables accurate or efficient linking of the age of onset of a disease, the number of months of duration, the age of symptom improvement, and the age of complete recovery, based on information about the patient's age, the time of disease onset, and the duration of the disease at the time of data entry or data acquisition.
[0118] In one embodiment, the file linking technique may mean a method of linking electronic medical record (EMR) documents with different extensions in a file format that is most suitable for running a prediction model and simultaneously running an external model different from the prediction model without losing the information they contain.
[0119] In one embodiment, the change history classification linkage technique may mean a method of organically distinguishing and linking change history related to ADR based on information such as newly occurred symptoms, disappeared symptoms, symptom severity, added or excluded medications, smoking status, drinking status, newly discovered medical history, etc. for the same patient.
[0120] In one embodiment according to the present disclosure, the update cycle of a periodic conversion database updated using at least one of a database linkage technique, a file linkage technique, or a change history-based linkage technique may be based on a predetermined time period. For example, the update cycle of the periodic conversion database may be changed from once a month to once a week depending on the user's needs, or the update cycle may be shortened or extended depending on the sufficiency or insufficiency of resources.
[0121] In one embodiment, the computing device (100) may obtain one or more of demographic data, medication-related data, disease diagnosis data, diagnostic test data, and nursing record data from a public database (e.g., FAERS, KAERS, WHO-ART, SIDER, EU-ADR, etc.). The public database is not limited to the disclosed database.
[0122] In additional embodiments, the diagnostic test-related data may include, for example, genetic sequences obtained from a patient-derived specimen and experimentally sequenced.
[0123] In one embodiment, the computing device (100) may utilize any form of big data analysis technique to construct the periodic database. Big data analysis techniques may include extracting or generating predictive information by identifying hidden patterns and relationships in large amounts of data. For example, big data analysis techniques may include text mining, which extracts specific information present in unstructured documents; opinion mining, which extracts opinions on a specific topic by crawling websites and other information related to a specific topic; and / or web mining, which extracts desired information from weblogs and / or search terms.
[0124] In one embodiment of the present disclosure, a computing device (100) can determine a plurality of input variables for learning an artificial intelligence-based model for detecting adverse drug reactions based on data stored in a periodic conversion database (S320).
[0125] In one embodiment, an input variable may refer to a variable or feature used in the learning or inference of an artificial intelligence model. For example, an input variable may include at least one of demographic variables, medication-related variables, diagnostic test-related variables, nursing record-related variables, or drug side effect diagnosis-related variables.
[0126] In one embodiment, a demographic variable is a variable that quantitatively or qualitatively represents a phenomenon related to a population. A demographic variable may include, for example, at least one of age or region.
[0127] In one embodiment, medication-related variables may represent quantitative or qualitative indicators that can be defined in relation to a user's medication intake. For example, medication-related variables may include at least one of the following: medication type, medication duration, medication dosage, or medication history.
[0128] In one embodiment, diagnostic test-related variables may represent quantitative or qualitative indicators derived from the results of various types of diagnostic or testing. For example, diagnostic test-related variables may include at least one of imaging tests, blood tests, or urine tests.
[0129] In one embodiment, the nursing record-related variables may refer to quantitative or qualitative variables obtained from nursing records generated during the treatment and / or nursing process. For example, the nursing record-related variables may be generated based on information contained in treatment or nursing record charts within a hospital. For example, the nursing record-related variables may be obtained through at least one of optical character recognition (OCR), natural language processing (NLP), or text mining techniques. In a further embodiment, the drug side effect diagnosis-related variables may include information related to the diagnosis of a disease caused by a drug side effect.
[0130] In one embodiment, imaging-based examination data that may be included in diagnostic examination-related variables may be obtained from ultrasound imaging, X-ray (simple radiography), CT (computed tomography), MRI (magnetic resonance imaging), fMRI (functional magnetic resonance image), PET (positron emission tomography), etc.
[0131] In one embodiment, information related to disease diagnosis due to adverse drug reactions, which may be included in the variables related to drug side effect diagnosis, may refer to data collected when unintended or unexpected symptoms resulting from the administered drug are diagnosed as a disease. In a further embodiment, the predictive model according to the present disclosure may suggest coadministration to prevent unintended symptoms that may result from the administered drug.
[0132] A more detailed description of the method of selecting input variables is provided later in Fig. 6.
[0133] In one embodiment, parameter information related to the preprocessing method may be obtained from data output from an external, separate prediction model different from the prediction model in the present disclosure.
[0134] In a further example, preprocessing for learning or inference of the above-described prediction model of the present disclosure (e.g., a preprocessing process of blood test data variables) may be performed using parameters generated by an external prediction model that operates on patient demographic data and disease information as input data. For example, when a 36-year-old white female is a triple-negative (HER-2, ER, PR negative) breast cancer (TNBC), the prediction accuracy for the occurrence probability of a specific ADR (e.g., mucosal inflammation, erythrodysesthesia, and increased blood ALT (Alanine aminotransferase) level) may be improved by using parameter information generated by the external prediction model.
[0135] In additional embodiments, the ADR prediction model may be driven by real-time updates from databases (e.g., FAERS, KAERS, WHO-ART, SIDER, EU-ADR) on medications administered to a specific individual, specific indications, and adverse drug reactions that have occurred, as well as reports of new research papers. In additional embodiments, the ADR prediction model may be driven by real-time updates from clinical response and progress data entered into international clinical trial databases (e.g., ClinicalTrials.gov), and may also be driven by real-time updates from research literature on genetic and drug responses in human or non-human entities.
[0136] In one embodiment, the data (e.g., combined processed data such as demographic information, genetic information, administered medications, and known / unknown side effects) that serve as the basis for driving an artificial intelligence or rule-based external prediction model that provides input parameter information to the ADR prediction model are exemplary only and are not limited to the present disclosure.
[0137] In one embodiment of the present disclosure, the computing device (100) may perform preprocessing by applying a mask to at least one of data corresponding to an adverse drug reaction (ADR), data related to a medication period, and a portion of demographic information (e.g., gender, age of onset, race) for learning and / or inferring a predictive model.
[0138] In one embodiment, the computing device (100) can determine the category of a region to be masked in the learning dataset. For example, the region to be masked may include data of one or more categories. The computing device (100) can cause the prediction model to perform masking-based learning based on the determined masking target category. In a further embodiment, the above-described masking-based preprocessing and learning of the prediction model can contribute to more accurately outputting predicted values of multiple adverse drug reactions (ADRs) in descending order of occurrence probability, for example, when two patients with similar medical histories differ only in gender, or when the gender is the same but in different age groups.
[0139] In one embodiment, the computing device (100) may determine an input variable for a prediction model among a plurality of variables based on a first technique for determining variable importance by using perturbed input data in which at least some of the variables have been changed and output data output from the prediction model in response to the perturbed input data. For example, the computing device (100) may generate perturbed input data in which at least some of the variables included in the input data have been changed. The computing device (100) may input a plurality of input data including input data in which the variables have not been changed and perturbed input data in which the variables have been changed into an artificial intelligence-based model, and may determine which variable among the plurality of variables will be selected by comparing the outputs of the models. For example, the computing device (100) may compare outputs (first output and second output) of each of the models of the first input data in which the first variable has not been changed and the first perturbed input data in which the first variable has been changed, to determine the importance of the changed variable. Based on this importance, input variables to be used in an AI-based predictive model can be determined. In one embodiment, the computing device (100) may determine input variables by inputting multiple input data, including input data with unchanged variables and perturbed input data with changed variables, to each of multiple AI-based models.
[0140] In one embodiment, the computing device (100) can determine input variables for a prediction model among a plurality of variables based on a second technique that selects input variables corresponding to principal components by converting data in a high-dimensional space into a low-dimensional space.
[0141] In one embodiment, the computing device (100) can determine input variables for a prediction model among a plurality of variables based on a third technique for selecting input variables in a direction that reduces impurity on a decision tree.
[0142] In one embodiment, the computing device (100) can determine input variables based on Boruta SHAP, Principal Component Analysis (PCA), Feature importance, Permutation importance, and / or Gini importance. Determination of such input variables can be included in the preprocessing process.
[0143] In one embodiment, the preprocessing step may be performed based on at least one of a technique for handling outliers, a technique for performing missing value imputation, a technique for performing categorization or recategorization, and a technique for performing normalization or standardization. Through these steps, input variables may be determined and / or preprocessed. The preprocessed input variables may be input into a predictive model to perform learning and / or inference.
[0144] In one embodiment of the present disclosure, the computing device (100) can train a prediction model to output adverse reaction information for a specific drug and a specific disease in response to the input variables being input by using a learning dataset including determined input variables and adverse reaction information for a specific drug and a specific disease (S330).
[0145] In one embodiment, the method for training the prediction model for detecting the adverse drug reaction may include a step of determining the prediction model for detecting the adverse drug reaction from among a plurality of candidate models, and training the prediction model so that the prediction model outputs whether or not there is an adverse drug reaction for a specific drug or a specific disease, or the probability of occurrence of an adverse drug reaction for a specific drug or a specific disease, using the preprocessed input variable as input data.
[0146] In one embodiment, the plurality of candidate models may be one or more models selected from among candidate models consisting of Random Forest, Logistic Regression, Support Vector Machine, Decision Tree, Naive Bayes, K-Nearest Neighbor, XGBoost, LightGBM, CatBoost, Convolutional Neural Network (CNN), Recurrent Neural Network (RNN), Long Short-Term Memory (LSTM), Bidirectional Long Short-Term Memory (BiLSTM), Transformer, and Generative Pre-trained Transformer (GPT).
[0147] In one embodiment, the prediction model may obtain output results from each of the candidate models using a predetermined test dataset, and, based on a performance evaluation of the output results from each of the candidate models, determine one of the candidate models as the prediction model.
[0148] In additional embodiments, the performance evaluation of the output results of the candidate models may be based on AUC (Area Under the Curve), accuracy, F1-score, precision, recall, sensitivity and / or specificity.
[0149] In one embodiment, the input variables used for training and / or inference of the prediction model may be preprocessed by performing outlier handling, missing value imputation, categorization / recategorization, and / or normalization / standardization.
[0150] In an additional example, large volumes of data including medications taken, possible adverse drug reactions, and patient demographic information can be obtained and updated in real time from public data DBs (e.g., FAERS, KAERS, WHO-ART, SIDER, EU-ADR).
[0151] In an additional embodiment, the prediction model according to the present disclosure can output meaningful prediction results for ADRs with a low probability of occurrence, considering the underlying health issue of the target patient and the fact that the patient is an infant or a child, even for administered drugs and indications with a relatively large amount of accumulated data, despite the vast amount of data.
[0152] In one embodiment, the prediction model can improve learning and / or inference efficiency and / or accuracy by referencing the prediction results of adverse drug reactions from the prediction model for each epoch performed multiple times. Specifically, the above-described efficiency and / or accuracy improvements can be achieved by receiving feedback on correct data regarding adverse drug reactions confirmed to have a high clinical causality through supervised, semi-supervised, and / or unsupervised learning.
[0153] As a specific example, in the case where the administration of aminosalicylic acid (5-ASA) and CAM inhibitor antibody drug (Vedolizumab) as treatment drugs for ulcerative colitis diagnosed within 6 months in a 47-year-old male is considered as a treatment option, if the prediction model according to the present disclosure outputs 34.4% for tiredness, 30.2% for joint pain, 28.9% for headache, and 19.3% for bronchitis as ADR prediction results, a medical professional may decide to administer 5-ASA and Vedolizumab to the patient to treat ulcerative colitis, and may also be suggested a combination drug to prevent them based on the expected occurrence probability and fatality of the expected ADRs.
[0154] In an additional embodiment, the ADR prediction results may be reflected in the update of the loss function or hyperparameters of the prediction model based on data on adverse drug reactions that occurred or did not occur as a result of monitoring the progress of treatment. For example, if the 47-year-old male patient with ulcerative colitis actually developed moderate bronchitis as a result of the prediction and was clearly causally related, feeding this data back to the prediction model may be a necessary process.
[0155] Specifically, in one embodiment, the computing device (100) can predict, based on information included in the periodic conversion database, possible moderate ADRs caused by drug Z administered to a patient X with cardiovascular disease and indication Y, including i. inflammatory bowel disease (IBD) with a probability of occurrence of 3.8%, ii. diarrhea with a probability of occurrence of 13.2%, and iii. dyspepsia with a probability of occurrence of 10.9% as mild ADRs.
[0156] Further explanation of the learning and operation of the prediction model will be described later in FIGS. 4 to 7.
[0157] FIG. 4 illustrates a construction flowchart of a periodic conversion database (420) including the periodic data loading and change history table loading principles of an electronic medical record (410) and a periodic conversion database (420) according to one embodiment of the present disclosure.
[0158] In one embodiment, FIG. 4 illustrates a feature in which a computing device (100) periodically loads data and change history of an order communication system (OCS) and / or an electronic medical record (EMR) (411) into a periodic database (420) from the OCS and / or EMR. In one embodiment, FIG. 4 illustrates that the computing device (100) can store data using an Operational Data Store (ODS) (421), an Interface (ITF) (422), and a Common Data Model (CDM) (423).
[0159] In one embodiment, ODS (421) may be a central database that serves as an operational data repository, providing snapshots of the latest data from various transactional systems for operational reporting. In one embodiment, ODS (421) may collect or receive information from newly added or modified OCS and / or EMR (411) at predetermined time intervals. In one embodiment, ODS (421) may be used as a database designed to integrate data from various sources.
[0160] The ODS (421) that receives the OCS and / or EMR (411) information that has been newly added or changed can transmit the information to be stored as a CDM (423) schema through the ITF (422) schema. The common data model (CDM) (423) may refer to a data model with the same structure and specifications that can be applied to medical data with different structures held by each medical institution. The common data model (CDM) (423) may allow the same analysis code to be individually executed by each data holding institution, thereby enabling distributed collaborative research that integrates distributed data.
[0161] In one embodiment according to the present disclosure, the periodic database (420) may store information such as patient demographic information, medications and indications administered to patients, adverse drug reactions (ADRs), and medical history at predetermined or optimal time intervals determined by the predictive model. The periodic database (420) may refer to a data repository that collects and constructs data on newly entered or additionally entered ADRs and patient information.
[0162] FIG. 5 illustrates the concept of linking files for constructing a periodic conversion database according to one embodiment of the present disclosure.
[0163] In one embodiment, the computing device (100) can build a periodic database (420) using at least one of a database linkage (510) technique, a file linkage (520) technique, and / or a change history separation linkage (530) technique.
[0164] In one embodiment, in relation to the database linkage (510), the computing device (100) can periodically load data from interface tables (511) in which data is entered on an electronic medical record (EMR) of a medical institution. The loaded data can be transferred to operational data store (ODS) database interface tables (513) through an operational data store (ODS) database loader (512). The computing device (100) can link the loaded data to a common data model (CDM) table on a CDM DM by applying a CDM transformation (514) to the ODS DB interface tables (513). Linkage between a database related to an EMR and a common data model database can be achieved in the manner described above.
[0165] The database linkage (510) technique can be utilized in the same or similar manner in the file linkage (520) technique and / or the change history classification linkage (530) technique described later.
[0166] In one embodiment, the file linking (520) technique may refer to an automated data management method that can more accurately and / or more efficiently link the age of onset of a disease, the number of months of duration, the age of symptom improvement, and the age of complete recovery, based on information about the patient's age at the time of data entry, the time of disease onset, and the duration of the disease. Specifically, with respect to the file linking (520), the computing device (100) may upload the data to the ODS DB through periodic loading of data on an EMR, a CSV (comma-separated values) file, or a Parquet file (521) of a medical institution. The loaded data may be uploaded to the interface tables (523) of the ODS DB through the loader (522) of the ODS DB. In one embodiment, files related to the EMR loaded into the ODS DB interface tables (523) may be stored in the format of CDM tables (525) of the CDM database through CDM conversion (524).
[0167] In one embodiment, the file linking (520) technique may mean a method of linking, for example, electronic medical record (EMR) documents (521) with different extensions, in a file format that is most suitable for driving the prediction model and simultaneously driving an external model different from the prediction model without loss of information contained therein.
[0168] In one embodiment, the change history classification linkage (530) technique may refer to a method of organically distinguishing and linking change history related to ADR based on, for example, newly occurred symptoms, disappeared symptoms, symptom severity, added or excluded medications, smoking status, drinking status, newly discovered medical history, etc. for the same patient. In one embodiment, in relation to the change history classification linkage (530), the computing device (100) may periodically load data from interface tables (531) and change history tables (531) in which data is entered on an electronic medical record (EMR) of a medical institution. The computing device (100) may upload to a temporary logical area DB of a data warehouse through an ODS DB loader (532). The computing device (100) may load data into interface tables and change history tables (533) of the ODS DB through the ODS DB loader (532). The loaded data can be stored in the CDM database (535) after going through CDM transformation (534). In one embodiment, the change-based linking (530) technique uses change history tables that store changed data to target and link changed data, thereby efficiently utilizing computing resources in data linking.
[0169] In one embodiment according to the present disclosure, the update cycle of the periodic conversion database (420) updated using at least one of the database linkage (510) technique, the file linkage (520) technique, and the change history classification linkage (530) technique may correspond to a predetermined time period. For example, the update cycle of the periodic conversion database (420) may be changed from once a month to once a week and / or daily according to the user's needs. For example, the update cycle of the periodic conversion database (420) may be determined based on the quantitative amount of computing resources of the computing device (100). For example, the computing device (100) may shorten or extend the update cycle depending on whether computing resources are sufficient or insufficient.
[0170] In one embodiment, the computing device (100) may obtain one or more of demographic data, medication-related data, disease diagnosis data, diagnostic test data, and nursing record data from a public database (e.g., FAERS, KAERS, WHO-ART, SIDER, EU-ADR, etc.). The public database is not limited to the databases disclosed herein.
[0171] FIG. 6 illustrates an overall process leading to database construction, development of artificial intelligence for detecting adverse drug reactions, and notification of abnormality detection results according to one embodiment of the present disclosure.
[0172] “Adverse drug reaction (ADR)” refers to a harmful and unintended reaction that occurs when administering or using a medicine, etc., and refers to cases where a causal relationship with the medicine, etc. cannot be ruled out. It may include not only adverse reactions that occur at normal doses, but also adverse drug reactions and withdrawal symptoms that occur when a drug is intentionally or accidentally overdosed or abused, and cases where the expected pharmacological effect does not appear.
[0173] “Adverse drug reaction monitoring” refers to activities to establish rational drug use and prevent drug-related damage in advance by quickly and systematically collecting and evaluating various adverse events that occur when using drugs, taking countermeasures, and delivering safety information and response results to medical professionals, consumers, etc.
[0174] Related concepts to ADRs include side effects and adverse drug events (ADEs). An adverse drug reaction (ADD) is the opposite of the principal action, which occurs when a drug is used for a specific purpose. It refers to any unintended effect that occurs when a drug is administered at normal dosages. As a term encompassing all actions other than those intended for therapeutic purposes, it is a concept that can be actively utilized in a variety of fields that encompass the study of all actions, regardless of their harmfulness.
[0175] An adverse drug event (ADE) refers to an undesirable and unintended sign, symptom, or disease that occurs during the administration or use of a drug, regardless of a causal relationship with the drug.
[0176] Adverse drug reactions (ADRs) are harmful and unintended reactions that occur during the administration and use of medicines according to normal usage, and in cases where a causal relationship with the medicine cannot be ruled out. Since ADRs are predictable and preventable, they are a concept for which management by pharmacists and medical staff is important.
[0177] Drug metabolism occurs mostly in the liver, and metabolism occurs through oxidation, reduction, or hydrolysis reactions of drugs by a metabolic enzyme system called cytochrome P450 (CYP450) in the endoplasmic reticulum of hepatocytes.
[0178] The examples in the present disclosure may include a step of deriving predictable ADRs in the order of causality (certain / probable / possible / unlikely / conditional) or classification according to the degree of reaction (mild / moderate / severe) and frequency of occurrence, based on the type of drug administered and the genotype / phenotype of each race / subject, history, etc.
[0179] In one embodiment, a computing device (100) can implement an active adverse drug reaction detection system through the processes of building a cycle conversion database (610), developing artificial intelligence for detecting adverse drug reactions (620), and notifying the results of the adverse drug reaction (630).
[0180] In one embodiment, the process for building a periodic database (610) may include terminology unification and / or terminology mapping (616) for raw data obtained from electronic medical records (612) and adverse drug reaction specifications (614). Terminology mapping (616) may refer to a process of unifying data formats (e.g., formats) for raw data having various formats. Terminology mapping (616) may refer to a process of unifying data representations for raw data having various expressions but the same or similar meaning.
[0181] For example, the term mapping (616) process may be based on an AI-based language model that determines whether multiple input data are synonyms or have a synonymous relationship with each other. In this example, the AI-based language model may extract features for two or more input data, compare the features in a vector space, and determine whether the two or more input data are synonyms or have a synonymous relationship with each other based on the results of the comparison. In this way, terms that are synonyms or have a synonymous relationship can be mapped and managed.
[0182] In one embodiment, the periodic database (618) may be built, for example, according to the OMOP-CDM standard. In one embodiment, the periodic database (618) may be preset to be updated at predetermined time intervals (e.g., daily, weekly, monthly, and / or yearly).
[0183] In one embodiment, the development of artificial intelligence for drug adverse reaction detection (620) may include a process of modeling and / or validating a predictive model.
[0184] In one embodiment, the computing device (100) can perform the processes of data preprocessing (622), artificial intelligence model learning (624), artificial intelligence model verification (626), and artificial intelligence model selection (628).
[0185] In one embodiment, data preprocessing (622) may include determining input variables to be used in training an artificial intelligence model among a number of unified variables in a periodic database. As illustrated in FIG. 6 , data preprocessing (622) may include feature selection, outlier determination, missing value imputation, categorization or recategorization, and normalization or standardization.
[0186] In one embodiment, the process associated with AI model training (624) may include selecting a specific candidate algorithm from among multiple candidate algorithms, selecting a specific learning method from among multiple learning methods, and tuning or optimizing hyperparameters. For example, the learning methods may include federated learning and / or incremental learning.
[0187] In one embodiment, the process associated with AI model validation (626) may include a methodology for validation using an N:1 test dataset and / or a methodology for validation through evaluation from an external server. In one embodiment, AI model validation (626) may also be performed using metrics related to AI model performance, such as AUC, accuracy, and F-1 score.
[0188] In one embodiment, the computing device (100) may select a specific artificial intelligence model from among multiple artificial intelligence models (e.g., algorithms) through artificial intelligence model verification (626).
[0189] In one embodiment, the predictive model of the computing device (100) can process data input from an input layer to a hidden layer. The results of residual connections between multiple hidden layers can be transmitted to an output layer.
[0190] According to one embodiment of the present disclosure, a predictive model may include an encoder and a decoder. For example, a first input data set including input variables or a combination of input variables, such as demographic data, may be input to the encoder, and the encoder and decoder may be pre-trained such that the output of the decoder, which receives the output of the encoder, corresponds to the first input data set. Through such pre-training, the encoder may be configured to accurately or efficiently generate features or vectors for the input variables.
[0191] According to a further embodiment of the present disclosure, the predictive model may include an encoder and a decoder. For example, a first input data set including input variables or a combination of input variables, such as demographic data, is input to the encoder, and a second input data set including data on administered drugs and adverse drug reactions, along with the output of the encoder, is input to the decoder, and the predictive model may be pre-trained such that the decoder outputs prediction results including information on administered drugs and possible clinicopathological reactions or occurrence probabilities associated with the ADRs.
[0192] In one embodiment, the predictive model can categorize each piece of information in the form of an embedding, and perform tasks necessary to generate output results based on the assigned category. For example, category embedding of family history information, medical history information, and genetic information can be performed in a separate embedding layer within the predictive model to reflect information on the probability of adverse drug reactions.
[0193] The effect of embodiments of the present disclosure can be achieved by efficiently and accurately predicting adverse reactions to specific drugs, thereby helping patients and medical professionals, for example, to design proactive countermeasures to prevent the occurrence of ADRs or minimize their impact on life phenomena in patients with diverse information. Furthermore, the computing device (100) may have a structure similar to a specific deep learning network that can be arbitrarily assumed. However, the prediction model in the present disclosure is not limited to an artificial intelligence model and may also be a rule-based operating model.
[0194] In one embodiment, in one embodiment for the process for the abnormality detection result notification (630), the computing device (100) can transmit the output result of the prediction model for the adverse drug reaction to a user terminal (transmitting / receiving device). For example, the output result of the prediction model transmitted to the user terminal can include at least one output result of a first result indicating a binary classification of the adverse drug reaction as detected or not detected, a second result indicating a probability notation for the adverse drug reaction, or a third result indicating a three-level category notation connecting the drug, the adverse reaction, and the risk. For example, the three-level category notation connecting “drug - ADR - risk” output by the above prediction model can output a three-level category notation as a severe adverse event of a single overdose of 10 mg or more of Wellbutrin (bupropion) - seizure - 71.0%.
[0195] In one embodiment, the output result output from the prediction model can be used to retrain the prediction model based on the accuracy evaluated for the output result. As a specific example, if a 29-year-old male patient with ankylosing spondylitis who received 40 mg of adalimumab once a week is provided with an output result indicating that there is a 35.2% probability that a benign, malignant, or unidentified neoplasm, including a polyp, will develop as an ADR in the patient, and if the patient did not actually develop the neoplasm, the prediction model according to the present disclosure can be provided with the above feedback data, that is, the 29-year-old male patient with ankylosing spondylitis who received 40 mg of adalimumab once a week did not develop the unidentified neoplasm as retraining data.
[0196] In one embodiment, feedback information corresponding to the output result transmitted from a user terminal that has received the output result from the prediction model may be used to retrain the prediction model. As a specific example, in the case where follow-up data has been collected for a patient for whom accurate information on demographic information and medical history was absent, the follow-up data may be provided to a user terminal that has received the ADR prediction output result from the prediction model, and the corresponding feedback information may be utilized to obtain more accurate prediction results from the prediction model or to retrain the prediction model.
[0197] Management of adverse drug reactions, if dependent on reporting channels such as the Adverse Drug Reaction Reporting System, requires voluntary reporting by patients and medical institutions. Therefore, if the reporter is unaware of the adverse drug reaction, it may not be reported, and if no timely action is taken against the adverse drug reaction, it may cause fatal additional damage to the patient. This has limitations.
[0198] An active adverse drug reaction detection system according to one embodiment of the present disclosure is a system capable of timely detection of adverse drug reactions. This system not only prevents potential harm due to failure to recognize adverse drug reactions but also enables simple and systematic management of adverse drug reactions. The system is particularly effective in detecting specific drugs for which adverse reactions have been reported in hospitalized patients, but can be utilized universally regardless of whether the patient is inpatient or outpatient, or whether the drug is specific. Therefore, the system can be useful for institutions lacking personnel to manage adverse drug reactions, as well as for individual patients taking medications.
[0199] Pharmaceutical companies conduct post-marketing surveillance (PMS) to monitor the long-term safety of drugs. However, this process requires significant time, manpower, and money, and major limitations include difficulties recruiting study subjects and collecting data. Therefore, active cooperation from medical professionals and patients is essential. A technique according to one embodiment of the present disclosure can reduce the time, manpower, and cost of PMS, and further contribute to creating a virtuous cycle in the field of drug safety.
[0200] FIG. 7 is a schematic diagram of a computing environment according to one embodiment of the present disclosure.
[0201] A component, module, or unit in the present disclosure includes a routine, procedure, program, component, data structure, or the like that performs a specific task or implements a specific abstract data type. Furthermore, those skilled in the art will readily appreciate that the methods presented in the present disclosure can be implemented in other computer system configurations, including single-processor or multiprocessor computing devices, minicomputers, mainframe computers, as well as personal computers, handheld computing devices, microprocessor-based or programmable consumer electronics, and the like, each of which may be operatively connected to one or more associated devices.
[0202] The embodiments described in this disclosure can also be implemented in distributed computing environments, where certain tasks are performed by remote processing devices that are connected through a communications network. In a distributed computing environment, program modules may be located in both local and remote memory storage devices.
[0203] Computing devices typically include a variety of computer-readable media. Computer-readable media can be any media accessible by a computer, including volatile and nonvolatile media, transitory and non-transitory media, removable and non-removable media. By way of example, and not limitation, computer-readable media can include computer-readable storage media and computer-readable transmission media.
[0204] Computer-readable storage media includes volatile and nonvolatile media, transitory and non-transitory media, removable and non-removable media implemented in any method or technology for storing information such as computer-readable instructions, data structures, program modules or other data. Computer-readable storage media includes, but is not limited to, RAM, ROM, EEPROM, flash memory or other memory technology, CD-ROM, digital video disks (DVD) or other optical disk storage, magnetic cassettes, magnetic tape, magnetic disk storage or other magnetic storage devices, or any other medium that can be accessed by a computer and used to store the desired information.
[0205] Computer-readable transmission media typically includes any information delivery media that embodies computer-readable instructions, data structures, program modules, or other data in a modulated data signal, such as a carrier wave or other transport mechanism. The term modulated data signal means a signal that has one or more of its characteristics set or changed so as to encode information in the signal. By way of example, and not limitation, computer-readable transmission media includes wired media, such as a wired network or direct-wired connection, and wireless media, such as acoustic, RF, infrared, or other wireless media. Combinations of any of the above are also intended to be included within the scope of computer-readable transmission media.
[0206] An exemplary environment (2000) implementing various aspects of the present invention is illustrated, including a computer (2002), which includes a processing unit (2004), a system memory (2006), and a system bus (2008). The computer (2000) herein may be used interchangeably with a computing device. The system bus (2008) connects system components, including but not limited to the system memory (2006), to the processing unit (2004). The processing unit (2004) may be any of a variety of commercially available processors. Dual processors and other multiprocessor architectures may also be utilized as the processing unit (2004).
[0207] The system bus (2008) may be any of several types of bus structures that may be additionally interconnected to a memory bus, a peripheral bus, and a local bus using any of a variety of commercial bus architectures. The system memory (2006) includes read-only memory (ROM) (2010) and random access memory (RAM) (2012). A basic input / output system (BIOS) is stored in non-volatile memory (2010), such as ROM, EPROM, or EEPROM, and contains basic routines that help transfer information between components within the computer (2002), such as during start-up. The RAM (2012) may also include high-speed RAM, such as static RAM, for caching data.
[0208] The computer (2002) also includes an internal hard disk drive (HDD) (2014) (e.g., EIDE, SATA), a magnetic floppy disk drive (FDD) (2016) (e.g., for reading from or writing to a removable diskette (2018)), a solid state drive (SSD), and an optical disk drive (2020) (e.g., for reading from or writing to a CD-ROM disk (2022) or other high-capacity optical media such as a DVD). The hard disk drive (2014), the magnetic disk drive (2016), and the optical disk drive (2020) may be connected to the system bus (2008) by a hard disk drive interface (2024), a magnetic disk drive interface (2026), and an optical drive interface (2028), respectively. The interface (2024) for implementing an external drive includes, for example, at least one or both of Universal Serial Bus (USB) and IEEE 1394 interface technologies.
[0209] These drives and their associated computer-readable media provide nonvolatile storage of data, data structures, computer-executable instructions, and the like. In the case of the computer (2002), the drives and media correspond to storing any data in a suitable digital format. While the description of computer-readable storage media above refers to HDDs, removable magnetic disks, and removable optical media such as CDs or DVDs, those skilled in the art will appreciate that other types of computer-readable storage media, such as zip drives, magnetic cassettes, flash memory cards, cartridges, and the like, may also be used in the exemplary operating environment, and that any such media may contain computer-executable instructions for performing the methods of the present invention.
[0210] A number of program modules, including an operating system (2030), one or more application programs (2032), other program modules (2034), and program data (2036), may be stored in the drive and RAM (2012). All or portions of the operating system, applications, modules, and / or data may also be cached in RAM (2012). It will be appreciated that the present invention may be implemented in various commercially available operating systems or combinations of operating systems.
[0211] A user may enter commands and information into the computer (2002) via one or more wired / wireless input devices, such as a keyboard (2038) and a pointing device such as a mouse (2040). Other input devices (not shown) may include a microphone, an IR remote control, a joystick, a game pad, a stylus pen, a touch screen, and the like. These and other input devices are often connected to the processing unit (2004) via an input device interface (2042) that is connected to the system bus (2008), but may be connected by other interfaces such as a parallel port, an IEEE 1394 serial port, a game port, a USB port, an IR interface, and the like.
[0212] A monitor (2044) or other type of display device is also connected to the system bus (2008) via an interface, such as a video adapter (2046). In addition to the monitor (2044), the computer typically includes other peripheral output devices (not shown), such as speakers, a printer, and so on.
[0213] The computer (2002) may operate in a networked environment using logical connections to one or more remote computers, such as remote computer(s) (2048), via wired and / or wireless communications. The remote computer(s) (2048) may be a workstation, a server computer, a router, a personal computer, a portable computer, a microprocessor-based entertainment device, a peer device, or other conventional network node, and generally include many or all of the components described for the computer (2002), although for simplicity, only the memory storage device (2050) is shown. The logical connections shown include wired / wireless connections to a local area network (LAN) (2052) and / or a larger network, such as a wide area network (WAN) (2054). Such LAN and WAN networking environments are common in offices and companies and facilitate enterprise-wide computer networks, such as intranets, all of which may be connected to a worldwide computer network, such as the Internet.
[0214] When used in a LAN networking environment, the computer (2002) is connected to a local network (2052) via a wired and / or wireless communication network interface or adapter (2056). The adapter (2056) may facilitate wired or wireless communications to the LAN (2052), which may also include a wireless access point installed therein for communicating with the wireless adapter (2056). When used in a WAN networking environment, the computer (2002) may include a modem (2058), be connected to a communications server on the WAN (2054), or have other means for establishing communications over the WAN (2054), such as via the Internet. The modem (2058), which may be internal or external and wired or wireless, is connected to the system bus (2008) via a serial port interface (2042). In a networked environment, program modules described for the computer (2002) or portions thereof may be stored in a remote memory / storage device (2050). It will be appreciated that the network connections depicted are exemplary and that other means of establishing a communications link between the computers may be used.
[0215] The computer (1602) operates to communicate with any wireless device or object that is arranged and operates via wireless communication, such as a printer, a scanner, a desktop and / or portable computer, a portable data assistant (PDA), a communication satellite, any equipment or location associated with a radio-detectable tag, and a telephone. This includes at least Wi-Fi and Bluetooth wireless technologies. Accordingly, the communication may be a predefined structure, as in a conventional network, or simply an ad hoc communication between at least two devices.
[0216] It should be understood that the specific order or hierarchy of steps in the presented processes is merely an example of exemplary approaches. It should be understood that the specific order or hierarchy of steps in the processes may be rearranged within the scope of the present disclosure based on design priorities. The method claims of the present disclosure provide elements of various steps in a sample order, but are not intended to be limited to the specific order or hierarchy presented.
[0217] As described above, the relevant contents have been described in the best form for carrying out the invention.
[0218] It can be used in devices and systems for managing patients who have adverse reactions to drugs.
Claims
1. A method for generating prediction results for adverse drug reactions using artificial intelligence technology performed by a computing device. A step of constructing a periodic database scheduled to be updated at predetermined time intervals using raw data for detecting adverse drug reactions; A step of determining a plurality of input variables for learning an artificial intelligence-based prediction model for detecting adverse drug reactions based on data stored in the above periodic conversion database; and A step of training the prediction model so that the prediction model outputs adverse reaction information for the specific drug and the specific disease in response to receiving the input variables, using the determined input variables and a learning dataset including adverse reaction information for the specific drug and the specific disease; Including, The steps for building the above periodic conversion database are: A step of converting the terms of the raw data to correspond to the common data standard specifications of the above prediction model; and A step of constructing a periodic conversion database that is updated in predetermined time period units using at least one of a database linkage technique, a file linkage technique, or a change history classification linkage technique; including, and The above input variables are, Determined by using a technique for determining variable importance by using perturbation input data in which at least some of the variables have changed and output data output from the prediction model in response to the perturbation input data, method.
2. In paragraph 1, The steps for building the above periodic conversion database are: A step of obtaining raw data for detecting the adverse drug reaction from an electronic medical record (EMR), wherein the raw data includes at least one of drug administration-related data, disease diagnosis data, diagnostic test data, or patient nursing record data; including more, method.
3. In paragraph 1, The above raw data is additionally obtained from at least one of FAERS, KAERS, WHO-ART, SIDER or EU-ADR. method.
4. In paragraph 1, The above input variables are, Including at least one of demographic variables, medication-related variables, diagnostic test-related variables, nursing record-related variables, or adverse drug reaction-related variables. method.
5. In paragraph 4, The above demographic variables include at least one of gender, age, or region, The above drug-taking related variables include at least one of drug type, drug-taking period, drug-taking dosage, or drug-taking history. The above diagnostic test related variables include at least one of a test through photography, a blood test, or a urine test, The above nursing record-related variables are variables obtained from nursing records and are obtained through at least one technique among optical character recognition (OCR), natural language processing (NLP), or text mining, and The above drug side effect diagnosis related variables include information related to the diagnosis of diseases caused by drug side effects. method.
6. In paragraph 1, The above input variables are: Step for handling outliers; A step of performing missing value imputation; A step of performing categorization or recategorization; and A step of performing normalization or standardization; Preprocessed based on, method.
7. In paragraph 6, The step of training the above prediction model for detecting the above adverse drug reaction is: A step of determining a prediction model for detecting the adverse drug reaction among a plurality of candidate models; A step of training the prediction model so that the prediction model outputs whether there is an adverse drug reaction to a specific drug or a specific disease, or the probability of occurrence of an adverse drug reaction to a specific drug or a specific disease, using the preprocessed input variables as input data; including, method.
8. In paragraph 1, The above prediction model: The model selected from candidate models consisting of Random Forest, Logistic Regression, Support Vector Machine, Decision Tree, Naive Bayes, K-Nearest Neighbor, XGBoost, LightGBM, CatBoost, Convolutional Neural Network (CNN), Recurrent Neural Network (RNN), Long Short-Term Memory (LSTM), Transformer, and Generative Pretrained Transformer (GPT). method.
9. In paragraph 8, A step of obtaining output results from each of the candidate models using a predetermined test dataset; A step of performing a performance evaluation on the output results from each of the above candidate models; and A step of determining one of the candidate models as the prediction model based on the results of the performance evaluation; Based on which the above prediction model is determined, method.
10. In paragraph 1, Information on adverse reactions to the above specific drugs and specific diseases: First information indicating whether there is an adverse drug reaction to a specific drug; Second information indicating the probability of occurrence of adverse drug reactions to a specific drug; Third-party information indicating whether there is an adverse drug reaction for a specific disease; and Fourth information indicating the probability of occurrence of adverse drug reactions for a specific disease; including, method.
11. In paragraph 1, A step of transmitting the output result of the prediction model for the above adverse drug reaction to a user terminal; including more, method.
12. In paragraph 11, The output result transmitted to the user terminal is: A first outcome representing a dual classification of adverse drug reactions, classified as detected or undetected; A second outcome indicating the probability of adverse drug reactions; or A third outcome, representing a three-tiered categorical notation linking drugs, adverse reactions, and risks; Contains at least one output result, method.
13. In paragraph 1, The output results from the above prediction model are: Used to retrain the prediction model based on the accuracy evaluated for the above output results. method.
14. In paragraph 1, Feedback information corresponding to the output result transmitted from the user terminal that received the output result from the above prediction model is used to retrain the prediction model. method.
15. A computer program stored in a computer-readable storage medium, wherein the computer program, when executed by a computing device, causes the computing device to perform operations for generating prediction results for adverse drug reactions using artificial intelligence technology, the operations comprising: An operation of constructing a periodic database scheduled to be updated at predetermined time intervals using raw data for detecting adverse drug reactions; An operation of determining a plurality of input variables for learning an artificial intelligence-based prediction model for detecting adverse drug reactions based on data stored in the above periodic conversion database; and An operation of training the prediction model so that the prediction model outputs adverse reaction information for the specific drug and the specific disease in response to receiving the input variables, using the determined input variables and a learning dataset including adverse reaction information for the specific drug and the specific disease; Including, The actions for building the above periodic conversion database are: An operation of converting the terms of the raw data to correspond to the common data standard specifications of the above prediction model; and An operation of constructing a periodic conversion database that is updated in predetermined time period units using at least one of a database linkage technique, a file linkage technique, or a change history classification linkage technique; including, and The above input variables are, Determined by using a technique for determining variable importance by using perturbation input data in which at least some of the variables have changed and output data output from the prediction model in response to the perturbation input data, A computer program stored on a computer-readable storage medium.
16. As a computing device, at least one processor; and memory; Includes, At least one processor of the above: An operation of constructing a periodic database scheduled to be updated at predetermined time intervals using raw data for detecting adverse drug reactions; An operation of determining a plurality of input variables for learning an artificial intelligence-based prediction model for detecting adverse drug reactions based on data stored in the above periodic conversion database; and An operation of training the prediction model so that the prediction model outputs adverse reaction information for the specific drug and the specific disease in response to receiving the input variables, using the determined input variables and a learning dataset including adverse reaction information for the specific drug and the specific disease; , and The actions for building the above periodic conversion database are: An operation of converting the terms of the raw data to correspond to the common data standard specifications of the above prediction model; and An operation of constructing a periodic conversion database that is updated in predetermined time period units using at least one of a database linkage technique, a file linkage technique, or a change history classification linkage technique; including, and The above input variables are, Determined by using a technique for determining variable importance by using perturbation input data in which at least some of the variables have changed and output data output from the prediction model in response to the perturbation input data, Computing device.
17. A method for generating prediction results for adverse drug reactions using artificial intelligence technology performed by a computing device, A step of obtaining input data for detecting adverse drug reactions based on data stored in a periodic database scheduled to be updated at predetermined time intervals using raw data for detecting adverse drug reactions; and A step of obtaining adverse reaction information for a specific drug and a specific disease from the input data using an artificial intelligence-based prediction model; Includes, The above periodic conversion database is: An operation of converting the terms of the raw data to correspond to the common data standard specifications of the above prediction model; and An operation of constructing a periodic conversion database that is updated in predetermined time period units using at least one of a database linkage technique, a file linkage technique, or a change history classification linkage technique; is created based on, and Multiple input variables for learning the above prediction model are: Determined by using a technique for determining variable importance by using perturbation input data in which at least some of the variables have changed and output data output from the prediction model in response to the perturbation input data, method.
Citation Information
Patent Citations
Adverse dug events surveillance system
KR1020080042256A
Thermal resistant connector, battery module, battery pack and vehicle including the same
KR1020230043052A
Wireless sensor system for indoor, and method for using the same in greenhouse
KR1020240006203A
Ballet shoes for practice
KR1020240018106A
Harmful chemical solution leakage sensor with IoT
KR102126902B1
Cited By
Multi-dimensional transaction behavior big data fusion method and system
CN121903763A
