Method and device for predicting occurrence of macce in target patient using artificial intelligence model

Abstract

A method for predicting the occurrence of major cardiac and cerebrovascular events (MACCE) in a target patient using an artificial intelligence model, according to one embodiment of the present invention, may comprise the steps of: acquiring target patient data from an electronic health record (EHR); and predicting occurrence of MACCE in the target patient by inputting the target patient data into at least one artificial intelligence model.

Description

Method and apparatus for predicting the occurrence of MACCE in target patients using an artificial intelligence model

[0001] The present invention relates to a method and apparatus for predicting the occurrence of MACCE in a target patient using an artificial intelligence model. This research was conducted with funding from the Ministry of Science and ICT (government) and support from Seoul National University Bundang Hospital (Project No.: 14-2022-0023; R&D Project: Individual Basic Research (MSIT); Research Project Title: Development of an Algorithm for Predicting Postoperative Cardiovascular Risk in Non-Cardiac Surgery Patients; Project Period: May 1, 2022 – April 30, 2024).

[0002] For reference, the present application claims priority based on Korean patent application filed on January 31, 2025 (Application No. 10-2025-0012625). The entire contents of the said application, which form the basis of this priority, are cited in the present application as reference.

[0003] More than 300 million non-cardiac surgeries are performed annually worldwide, and major cardiac and cerebrovascular events (MACE) occurring in patients undergoing non-cardiac surgery are a significant health issue.

[0004] Due to factors such as the aging population, preoperative and postoperative complications and mortality rates associated with MACCE are increasing, and various studies are underway to predict the occurrence of MACCE.

[0005] In this regard, the revised cardiac risk index (RCRI) has conventionally been used to predict the occurrence of MACCE, but it has limitations in predictive accuracy for personalized risk assessments for individual patients.

[0006] Accordingly, there is a need to develop technology that reduces unnecessary tests and enables rapid surgical planning by providing patient-specific risk assessments for MACCE.

[0007] The problem that the present invention aims to solve is to provide a customized risk assessment by utilizing target patient data that includes comprehensive predictive variables such as the patient's medical records, pre-operative test results, and drug usage history, thereby reducing unnecessary tests, lowering related costs, and enabling rapid surgical planning.

[0008] Furthermore, the problem that the present invention aims to solve is to improve the accuracy and reliability of predicting the occurrence of MACCE in target patients by using an artificial intelligence model trained on patient data classified according to a predetermined time interval for risk assessment.

[0009] However, the problems that the present invention aims to solve are not limited to those mentioned above, and other problems not mentioned will be clearly understood by those skilled in the art to which the present invention pertains from the description below.

[0010] A method for predicting the occurrence of MACCE in a target patient using an artificial intelligence model according to one embodiment of the present invention may include the step of obtaining target patient data from an electronic health record (EHR); and the step of inputting the target patient data into at least one artificial intelligence model to predict the occurrence of MACCE in the target patient.

[0011] Here, the above at least one artificial intelligence model may be trained based on patient data classified according to a predetermined time interval regarding risk assessment.

[0012] In addition, the predetermined time intervals for the above risk assessment may include a first interval corresponding to 3 to 30 days prior to the patient's hospitalization; a second interval corresponding to 3 to 3 months prior to the patient's hospitalization; and a third interval corresponding to 3 to 1 year prior to the patient's hospitalization.

[0013] In addition, the above-mentioned at least one artificial intelligence model may include a first model trained based on patient data corresponding to the first section, a second model trained based on patient data corresponding to the second section, and a third model trained based on patient data corresponding to the third section.

[0014] Here, the step of predicting the occurrence of MACCE in the target patient may include: a step of determining which time interval among the first to third intervals the target patient data is included in; and a step of predicting the occurrence of MACCE in the target patient using a model determined by referring to the time interval among the first to third models in which the target patient data is included.

[0015] Additionally, if it is determined that the target patient data falls within two or more time intervals among the first to third intervals, the step of predicting the occurrence of MACCE in the target patient may include applying different weights to the predicted values ​​derived from the models corresponding to each of the two or more time intervals to predict the occurrence of MACCE in the target patient.

[0016] Meanwhile, the step of predicting the occurrence of MACCE in the target patient may include the step of predicting the occurrence of MACCE in the target patient based on a weighted sum operation of the predicted values ​​derived by inputting the target patient data into each of the first to third models.

[0017] Here, the weight for the predicted value derived through the first model in the weighted sum operation may be set to the highest value.

[0018] An apparatus for predicting the occurrence of MACCE in a target patient using an artificial intelligence model according to another embodiment of the present invention comprises: a memory in which a MACCE occurrence prediction program is stored; and a processor that loads the MACCE occurrence prediction program from the memory and executes the MACCE occurrence prediction program, wherein the processor can obtain target patient data from an electronic health record (EHR) and input the target patient data into at least one artificial intelligence model to predict the occurrence of MACCE in the target patient.

[0019] Here, the above at least one artificial intelligence model may be trained based on patient data classified according to a predetermined time interval regarding risk assessment.

[0020] In addition, the predetermined time intervals for the above risk assessment may include a first interval corresponding to 3 to 30 days prior to the patient's hospitalization; a second interval corresponding to 3 to 3 months prior to the patient's hospitalization; and a third interval corresponding to 3 to 1 year prior to the patient's hospitalization.

[0021] In addition, the above-mentioned at least one artificial intelligence model may include a first model trained based on patient data corresponding to the first section, a second model trained based on patient data corresponding to the second section, and a third model trained based on patient data corresponding to the third section.

[0022] Here, the processor determines which time interval among the first to third intervals the target patient data is included in, and can predict the occurrence of MACCE of the target patient using a model determined by referring to the time interval among the first to third models in which the target patient data is included.

[0023] In addition, if it is determined that the target patient data falls within two or more of the first to third intervals, the processor can predict the occurrence of MACCE in the target patient by applying different weights to the predicted values ​​derived from the models corresponding to each of the two or more time intervals.

[0024] Meanwhile, the processor can predict the occurrence of MACCE in the target patient based on a weighted sum operation of the predicted values ​​derived by inputting the target patient data into each of the first to third models.

[0025] Here, the weight for the predicted value derived through the first model in the weighted sum operation may be set to the highest value.

[0026] A non-transient computer-readable recording medium storing a computer program according to another embodiment of the present invention may include instructions for a processor to perform a method for predicting the occurrence of MACCE, wherein the processor includes the step of obtaining target patient data from an electronic health record (EHR); and the step of inputting the target patient data into at least one artificial intelligence model to predict the occurrence of MACCE in the target patient, wherein the at least one artificial intelligence model is learned based on patient data classified according to a predetermined time interval for risk assessment.

[0027] A computer program stored in a computer-readable recording medium according to another embodiment of the present invention may include instructions for the processor to perform a method for predicting the occurrence of MACCE, wherein the computer program includes the step of obtaining target patient data from an electronic health record (EHR); and the step of inputting the target patient data into at least one artificial intelligence model to predict the occurrence of MACCE in the target patient, wherein the at least one artificial intelligence model is learned based on patient data classified according to a predetermined time interval for risk assessment.

[0028] According to an embodiment of the present invention, through integration with an EHR system, the pre-operative evaluation process can be simplified, and efficient allocation of medical resources, cost management, and improvement of patient treatment outcomes can be achieved.

[0029] In addition, according to an embodiment of the present invention, by using at least one artificial intelligence model trained on patient data classified according to a predetermined time interval for risk assessment to predict the occurrence of MACCE in a target patient, the accuracy and reliability of predicting the occurrence of MACCE can be improved compared to conventional prediction tools such as RCRI.

[0030] FIG. 1 is a block diagram showing a MACCE occurrence prediction device according to an embodiment of the present invention.

[0031] FIG. 2 is a block diagram conceptually illustrating the functions of a MACCE occurrence prediction program according to an embodiment of the present invention.

[0032] FIG. 3 is a flowchart illustrating a method for predicting the occurrence of MACCE in a target patient according to one embodiment of the present invention.

[0033] FIG. 4 is a diagram exemplarily illustrating the main predictive variables of an artificial intelligence model according to one embodiment of the present invention.

[0034] FIG. 5 is a diagram exemplarily illustrating the verification results of an artificial intelligence model according to one embodiment of the present invention.

[0035] The advantages and features of the present invention and the methods for achieving them will become clear by referring to the embodiments described below in detail together with the accompanying drawings. However, the present invention is not limited to the embodiments disclosed below but can be implemented in various different forms. These embodiments are provided merely to ensure that the disclosure of the present invention is complete and to fully inform those skilled in the art of the scope of the invention, and the present invention is defined only by the scope of the claims.

[0036] In describing the embodiments of the present invention, specific descriptions of known functions or configurations will be omitted if it is determined that such detailed descriptions could unnecessarily obscure the essence of the invention. Furthermore, the terms described below are defined in consideration of their functions in the embodiments of the present invention, and these definitions may vary depending on the intentions or practices of the user or operator. Therefore, such definitions should be based on the content throughout this specification.

[0037] FIG. 1 is a block diagram showing a MACCE occurrence prediction device according to an embodiment of the present invention.

[0038] Referring to FIG. 1, the MACCE occurrence prediction device (100) may include a processor (110), an input / output device (120), and a memory (130).

[0039] The processor (110) can control the overall operation of the MACCE occurrence prediction device (100).

[0040] The processor (110) can receive target patient data from an electronic health record (EHR) using an input / output device (120).

[0041] In the present invention, it has been described that the target patient data is input through an input / output device (120), but it is not limited thereto. That is, according to an embodiment, the MACCE occurrence prediction device (100) may include a transceiver (not shown), and the MACCE occurrence prediction device (100) may receive target patient data using the transceiver (not shown), and the target patient data may be generated within the MACCE occurrence prediction device (100).

[0042] The processor (110) receives target patient data from an electronic health record (EHR) and inputs the target patient data into at least one artificial intelligence model to predict the occurrence of MACCE in the target patient.

[0043] The input / output device (120) may include one or more input devices and / or one or more output devices. For example, the input device may include a microphone, keyboard, mouse, touch screen, etc., and the output device may include a display, speaker, etc.

[0044] The memory (130) can store information necessary for the execution of the MACCE occurrence prediction program (200) and the MACCE occurrence prediction program (200).

[0045] In this specification, the MACCE occurrence prediction program (200) may refer to software that includes instructions for predicting the occurrence of MACCE in a target patient using at least one artificial intelligence model.

[0046] The processor (110) can load the MACCE occurrence prediction program (200) and information necessary for the execution of the MACCE occurrence prediction program (200) from memory (130) in order to execute the MACCE occurrence prediction program (200).

[0047] The processor (110) can predict the occurrence of MACCE in a target patient by executing a MACCE occurrence prediction program (200) and receiving target patient data from an electronic health record (EHR).

[0048] In the present invention, major cardiovascular and cerebrovascular events (MACCE) are clinical indicators used to comprehensively evaluate the severe consequences of cardiovascular and cerebrovascular diseases, and may include items such as death, myocardial infarction, stroke, reperfusion therapy or revascularization, and hospitalization for heart failure.

[0049] The function and / or operation of the MACCE occurrence prediction program (200) will be examined in detail through FIG. 2.

[0050] FIG. 2 is a block diagram conceptually illustrating the functions of a MACCE occurrence prediction program according to an embodiment of the present invention.

[0051] Referring to FIG. 2, the MACCE occurrence prediction program (200) may include a patient data acquisition unit (210) and a MACCE occurrence prediction unit (220).

[0052] The patient data acquisition unit (210) and MACCE occurrence prediction unit (220) illustrated in FIG. 2 are conceptually divided to easily explain the functions of the MACCE occurrence prediction program (200), but are not limited thereto. According to embodiments, the functions of the patient data acquisition unit (210) and the MACCE occurrence prediction unit (220) may be merged or separated and may be implemented as a series of instructions included in a single program.

[0053] First, the patient data acquisition unit (210) can acquire target patient data from the electronic health record (EHR).

[0054] Here, the target patients may refer to, but are not limited to, patients scheduled for hospitalization for non-cardiac surgery.

[0055] Furthermore, an Electronic Health Record (EHR) can refer to a system that stores and manages patients' health information and medical records in a digital format. EHRs are characterized by interoperability; through EHRs, not only is it possible to share patient medical information across various medical institutions, but healthcare providers can also update and verify patient information in real time.

[0056] Here, the target patient data according to one embodiment may include data regarding the patient's personal information (age, gender, etc.), past and present medical history (i.e., baseline history), drug prescription history, diagnostic results and test results, surgical history, imaging data, vaccination records, medical records and doctor's notes, etc.

[0057] Meanwhile, according to one embodiment, the patient data acquisition unit (210) can acquire target patient data regarding a specific treatment within a specific hospital from an electronic medical record (EMR).

[0058] Here, the patient data acquisition unit (210) can standardize or convert target patient data acquired from the electronic medical record so that it can be integrated into the electronic health record by using an OMOP CDM (observational medical outcomes partnership common data model). An OMOP CDM model according to one embodiment can convert medical data having different formats and structures into one standardized format, thereby ensuring the accuracy of data interpretation based on the integration of terminology systems.

[0059] In this way, by using an electronic health record or OMOP CDM model to acquire comprehensive target patient data including the patient's medical records, preoperative test results, and drug usage history, not only can broad applicability and scalability be secured, but unique effects can also be achieved, such as providing patient-specific risk assessment.

[0060] Next, the MACCE occurrence prediction unit (220) can predict the occurrence of MACCE in the target patient by inputting the target patient data into at least one artificial intelligence model.

[0061] Here, a Kubernetes-based environment may be established for the efficient training, evaluation, and deployment of at least one artificial intelligence model, and the training (e.g., federated learning, etc.) and management of the at least one artificial intelligence model may be automated in the Kubernetes-based environment.

[0062] In addition, the above-mentioned at least one artificial intelligence model may be trained based on patient data classified according to a predetermined time interval for risk assessment. The patient data classified for training the above-mentioned at least one artificial intelligence model is based on clinical trial results verified based on a time-at-risk assessment technique, which will be described later in FIG. 5.

[0063] Specifically, patient data according to one embodiment may be classified according to a predetermined time interval for risk assessment based on a time-at-risk window, and the predetermined time interval for risk assessment may include a first interval corresponding to 3 to 30 days prior to the patient's admission, a second interval corresponding to 3 days prior to the patient's admission, and a third interval corresponding to 3 days prior to the patient's admission.

[0064] In addition, at least one artificial intelligence model according to one embodiment may include a first model learned based on patient data corresponding to a first interval, a second model learned based on patient data corresponding to a second interval, and a third model learned based on patient data corresponding to a third interval.

[0065] For example, the first to third models may include at least one algorithm among AdaBoost, Decision Trees, Gradient Boosting Machine, Lasso Logistic Regression, and Random Forest.

[0066] However, the algorithms included in the first to third models above are merely examples and may be modified in various ways within the scope of achieving the purpose of the present invention.

[0067] Meanwhile, according to one embodiment of the present invention, the MACCE occurrence prediction unit (220) can determine which time interval among the first to third intervals the target patient data is included in.

[0068] Additionally, the MACCE occurrence prediction unit (220) can predict the occurrence of MACCE in a target patient by using a model determined by referring to a time interval containing target patient data among the first to third models.

[0069] For example, if the target patient data includes data up to 20 days prior to the target patient's hospitalization, the MACCE occurrence prediction unit (220) can predict the MACCE occurrence of the target patient using the first model.

[0070] In another example, if the target patient data includes data up to 2 months prior to the target patient's hospitalization, the MACCE occurrence prediction unit (220) can predict the MACCE occurrence of the target patient using the second model.

[0071] As another example, if the target patient data includes data for at least 3 months prior to the target patient's hospitalization, the MACCE occurrence prediction unit (220) can predict the MACCE occurrence of the target patient using a third model.

[0072] Additionally, according to one embodiment of the present invention, when it is determined that the target patient data is included in two or more time intervals among the first to third intervals, the MACCE occurrence prediction unit (220) can predict the occurrence of MACCE in the target patient by applying different weights to the prediction values ​​derived from the models corresponding to each of the two or more time intervals.

[0073] For example, if the target patient data includes data up to two months prior to the target patient's hospitalization, the target patient data may be determined to be included in the second and third intervals, and the MACCE occurrence prediction unit (220) may predict the occurrence of MACCE in the target patient by inputting the target patient data into the second model and the third model, respectively, and applying different weights to the predicted values ​​derived therefrom. At this time, the weight for the predicted value derived through the second model may be greater than the weight for the predicted value derived through the third model.

[0074] In addition, according to one embodiment of the present invention, the MACCE occurrence prediction unit (220) can predict the occurrence of MACCE in a target patient based on a weighted sum operation of the predicted values ​​derived by inputting target patient data into each of the first to third models.

[0075] At this time, in the weighted sum operation for predicting the occurrence of MACCE in the target patient, the weight for the predicted value derived through the first model may be set to the highest, and the weight for the predicted value derived through the third model may be set to the lowest. This is based on results derived using explainable artificial intelligence (XAI) algorithms (e.g., SHAP, LIME, etc.) to ensure the key predictor variables of the artificial intelligence model and the interpretability of the model, and will be described later in FIG. 4.

[0076] In this way, by predicting the occurrence of MACCE in a target patient using at least one artificial intelligence model trained on patient data classified according to a predetermined time interval for risk assessment, a unique effect can be achieved that can improve the accuracy of predicting the occurrence of MACCE.

[0077] FIG. 3 is a flowchart illustrating a method for predicting the occurrence of MACCE in a target patient according to one embodiment of the present invention.

[0078] Referring to FIG. 3, the patient data acquisition unit (210) can acquire target patient data from an electronic health record (EHR) (S310).

[0079] Next, the MACCE occurrence prediction unit (220) can predict the occurrence of MACCE in the target patient by inputting the target patient data into at least one artificial intelligence model (S320). Here, the at least one artificial intelligence model can be trained based on patient data classified according to a predetermined time interval regarding risk assessment.

[0080] FIG. 4 is a diagram exemplarily illustrating the main predictive variables of an artificial intelligence model according to one embodiment of the present invention.

[0081] Figure 4 illustrates the relative importance of covariates in an artificial intelligence model derived based on explainable artificial intelligence.

[0082] Referring to Figure 4, it can be seen that the patient's underlying history or condition, such as ischemic heart disease and brain injury, had a very significant impact on the prediction of the artificial intelligence model.

[0083] In addition, it can be confirmed that important preoperative measurements include hemoglobin, creatinine, albumin, CK-MB, ESR, etc.

[0084] In addition, it can be confirmed that among drugs, antithrombotic agents and beta-blocking agents played an important role.

[0085] In short, it can be confirmed that a variable representing the patient's baseline history or recent condition has the greatest influence on the prediction of MACCE occurrence, and in predicting the occurrence of MACCE in the target patient based on a weighted sum operation for each of the prediction values ​​of the first to third models, the highest weight may be assigned to the prediction value of the first model, which is the prediction value reflecting the patient's recent condition.

[0086] FIG. 5 is a diagram exemplarily illustrating the verification results of an artificial intelligence model according to one embodiment of the present invention.

[0087] Figure 5(A) shows the AUC (area under the ROC curve) corresponding to the accuracy of MACCE occurrence prediction for five artificial intelligence models trained on patient data classified according to a predetermined time interval for risk assessment.

[0088] Specifically, each of the above five artificial intelligence models may include a first to third model trained based on patient data corresponding to each of the first to third intervals, and the five artificial intelligence models were used to predict MACCE within one year from the date of discharge of the target patient after surgery.

[0089] Referring to Figure 5, it can be seen that the AUC of the AI ​​model including Adaboost (i.e., corresponding to the prediction accuracy of the model) is 0.786, the AUC of the AI ​​model including DecisionTree is 0.663, the AUC of the AI ​​model including Gradient Boosting Machine is 0.826, the AUC of the AI ​​model including Lasso Logistic Regression is 0.813, the AUC of the AI ​​model including Random Forest is 0.817, and the AUC according to RCRI is 0.704.

[0090] Here, it can be seen that four out of five artificial intelligence models show higher MACCE occurrence prediction performance compared to the conventional RCRI.

[0091] In addition, referring to FIGS. 5(B) and FIGS. 5(C), an AUPRC (area under the precision-recall curve) representing the precision and recall of the artificial intelligence model and a correction plot evaluating the agreement between the predicted probability and the actual result are shown, and from this, it can be confirmed that the prediction performance of the artificial intelligence model trained with patient data classified according to a predetermined time interval for risk assessment is high.

[0092] Combinations of each block of the block diagram attached to the present invention and each step of the flowchart may be performed by computer program instructions. Since these computer program instructions may be loaded into an encoding processor of a general-purpose computer, a special-purpose computer, or other programmable data processing equipment, the instructions performed through the encoding processor of the computer or other programmable data processing equipment create means for performing the functions described in each block of the block diagram or each step of the flowchart. Since these computer program instructions may also be stored in computer-available or computer-readable memory that can be directed toward the computer or other programmable data processing equipment to implement the function in a specific way, the instructions stored in computer-available or computer-readable memory may also produce a manufactured item containing instruction means for performing the function described in each block of the block diagram or each step of the flowchart. Since computer program instructions can be loaded onto a computer or other programmable data processing equipment, instructions that execute a computer or other programmable data processing equipment by performing a series of operation steps on the computer or other programmable data processing equipment to create a process executed by the computer can also provide steps for executing the functions described in each block of the block diagram and each step of the flowchart.

[0093] Additionally, each block or each step may represent a module, segment, or part of code containing one or more executable instructions for executing a specified logical function(s). Also, it should be noted that in some alternative embodiments, the functions mentioned in the blocks or steps may occur out of order. For example, two blocks or steps described in succession may actually be performed substantially simultaneously, or the blocks or steps may sometimes be performed in reverse order according to the corresponding function.

[0094] The above description is merely an illustrative explanation of the technical concept of the present invention, and those skilled in the art to which the present invention pertains will be able to make various modifications and variations within the scope of the essential quality of the present invention. Accordingly, the embodiments disclosed in the present invention are intended to explain, not limit, the technical concept of the present invention, and the scope of the technical concept of the present invention is not limited by such embodiments. The scope of protection of the present invention shall be interpreted by the claims below, and all technical concepts within the equivalent scope shall be interpreted as being included within the scope of rights of the present invention.

Claims

1. A method for predicting the occurrence of MACCE (major cardiac and cerebrovascular events) in target patients using an artificial intelligence model, A step of acquiring target patient data from an electronic health record (EHR); and The method includes the step of inputting the above-mentioned target patient data into at least one artificial intelligence model to predict the occurrence of MACCE in the target patient, and The above-mentioned at least one artificial intelligence model is, learned based on patient data classified according to a specified time interval for risk assessment, MACCE occurrence prediction method.

2. In Paragraph 1, The predetermined time interval regarding the above risk assessment is, A first period corresponding to 3 to 30 days prior to the patient's admission; A second period corresponding to 3 days to 3 months prior to the hospitalization of the above-mentioned patient; and including a third period corresponding to 3 days to 1 year prior to the hospitalization of the above patient MACCE occurrence prediction method.

3. In Paragraph 2, The above-mentioned at least one artificial intelligence model is, A first model trained based on patient data corresponding to the first section, a second model trained based on patient data corresponding to the second section, and a third model trained based on patient data corresponding to the third section. MACCE occurrence prediction method.

4. In Paragraph 3, The step of predicting the occurrence of MACCE in the above-mentioned patient is, A step of determining which time interval among the first to third intervals the target patient data is included in; and A step of predicting the occurrence of MACCE in a target patient using a model determined by referring to a time interval containing the target patient data among the first to third models. MACCE occurrence prediction method.

5. In Paragraph 4, If it is determined that the above target patient data is included in two or more time intervals among the first to third intervals, The step of predicting the occurrence of MACCE in the above-mentioned patient is, A step of predicting the occurrence of MACCE in the target patient by applying different weights to predicted values ​​derived from models corresponding to each of the two or more time intervals. MACCE occurrence prediction method.

6. In Paragraph 3, The step of predicting the occurrence of MACCE in the above-mentioned patient is, The method includes the step of predicting the occurrence of MACCE in the target patient based on a weighted sum operation of the predicted values ​​derived by inputting the target patient data into each of the first to third models. MACCE occurrence prediction method.

7. In Paragraph 6, Characterized by the fact that in the above weighted sum operation, the weight for the predicted value derived through the first model is set to the highest value. MACCE occurrence prediction method.

8. A device for predicting the occurrence of MACCE (major cardiac and cerebrovascular events) in target patients using an artificial intelligence model, Memory where the MACCE occurrence prediction program is stored; and It includes a processor that loads the MACCE occurrence prediction program from the memory and executes the MACCE occurrence prediction program. The above processor is, Acquire target patient data from electronic health records (EHR), and The above target patient data is input into at least one artificial intelligence model to predict the occurrence of MACCE in the above target patient, The above-mentioned at least one artificial intelligence model is, learned based on patient data classified according to a specified time interval for risk assessment, MACCE occurrence prediction device.

9. In Paragraph 8, The predetermined time interval regarding the above risk assessment is, A first period corresponding to 3 to 30 days prior to the patient's admission; A second period corresponding to 3 days to 3 months prior to the hospitalization of the above-mentioned patient; and including a third period corresponding to 3 days to 1 year prior to the hospitalization of the above patient MACCE occurrence prediction device.

10. In Paragraph 9, The above-mentioned at least one artificial intelligence model is, A first model trained based on patient data corresponding to the first section, a second model trained based on patient data corresponding to the second section, and a third model trained based on patient data corresponding to the third section. MACCE occurrence prediction device.

11. In Paragraph 10, The above processor is, Determining which time interval among the first to third intervals the above target patient data is included in, Predicting the occurrence of MACCE in the target patient using a model determined by referring to the time interval containing the target patient data among the first to third models above. MACCE occurrence prediction device.

12. In Paragraph 11, If it is determined that the above target patient data is included in two or more time intervals among the first to third intervals, The above processor is, Predicting the occurrence of MACCE in the target patient by applying different weights to the predicted values ​​derived from the models corresponding to each of the two or more time intervals mentioned above. MACCE occurrence prediction method.

13. In Paragraph 10, The above processor is, Predicting the occurrence of MACCE in the target patient based on a weighted sum operation of the predicted values ​​derived by inputting the target patient data into each of the first to third models. MACCE occurrence prediction device.

14. In Paragraph 13, Characterized by the fact that in the above weighted sum operation, the weight for the predicted value derived through the first model is set to the highest value. MACCE occurrence prediction device.

15. A non-transient computer-readable recording medium storing a computer program, When the above computer program is executed by a processor, A step of acquiring target patient data from an electronic health record (EHR); and The method includes the step of inputting the target patient data into at least one artificial intelligence model to predict the occurrence of MACCE (major cardiac and cerebrovascular events) in the target patient, and The above-mentioned at least one artificial intelligence model is, learned based on patient data classified according to a specified time interval for risk assessment, Instructions for the processor to perform a MACCE occurrence prediction method Non-transient computer-readable recording medium.

Images