Method, apparatus, and computer program for determining, on basis of electrocardiogram data, whether user with hospitalization history is to be readmitted

Abstract

The present disclosure provides a method, computing apparatus, and program for determining, on the basis of electrocardiogram data, whether a user having a hospitalization history is to be readmitted. The method according to an embodiment of the present disclosure is carried out by the computing apparatus comprising at least one processor and comprises the steps of: acquiring electrocardiogram data of a user who has been discharged after hospitalization; acquiring, from the electrocardiogram data, on the basis of a pre-trained neural network model, a first score corresponding to left ventricular systolic dysfunction of the user and a second score corresponding to left ventricular diastolic dysfunction of the user; and predicting the possibility of readmission of the user on the basis of at least one of the first score and the second score.

Description

Method, device, and computer program for determining whether a user with a history of inpatient treatment should be readmitted based on electrocardiogram data

[0001] The present disclosure relates to artificial intelligence technology in the medical field, and more specifically, to a method, apparatus, and computer program for determining whether a user with a history of inpatient treatment should be readmitted based on electrocardiogram data.

[0002] With the recent advancement of Artificial Intelligence (AI), technologies that acquire users' biometric data using wearable devices or mobile electronic devices and analyze their condition in real time based on this data are becoming commonplace. For example, this includes technology that analyzes electrocardiogram (ECG) data acquired through a smartwatch using an AI model to identify a user's condition related to heart disease. Through such biometric data analysis technology utilizing AI, users can continuously monitor for abnormal signs or their health status without visiting a medical institution.

[0003] However, existing AI-based electrocardiogram (ECG) analysis technologies have primarily focused on identifying the presence of specific diseases or abnormalities based on ECG data from a single point in time, which has limitations in reflecting changes in vital signs or the progression of condition deterioration over time. In particular, since cardiovascular and heart failure conditions tend to involve repeated cycles of improvement and deterioration, it is difficult to adequately assess a user's future risk based solely on a judgment from a single point in time.

[0004] Furthermore, conventional technologies focus on anomaly detection or classification accuracy, lacking capabilities to support more practical medical decision-making, such as tracking changes in a user's condition or predicting the likelihood of clinical deterioration. Consequently, there is a growing need for technology capable of quantitatively evaluating changes in a user's condition based on continuously acquired biometric data and more accurately determining potential future risks.

[0005] The present disclosure is conceived in response to the aforementioned background technology and aims to provide a method, apparatus, and computer program for determining whether a user with a history of inpatient treatment should be readmitted based on electrocardiogram data.

[0006] However, the problems to be solved in this disclosure are not limited to those mentioned above, and other unmentioned problems may be clearly understood based on the description below.

[0007] A method for determining whether a user with a history of inpatient treatment will be readmitted based on electrocardiogram data, performed by a computing device including at least one processor according to one embodiment of the present disclosure for solving the above-described problem, comprises the steps of: acquiring electrocardiogram data of a user with a history of inpatient treatment; acquiring a first score corresponding to left ventricular systolic dysfunction of the user and a second score corresponding to left ventricular diastolic dysfunction of the user from the electrocardiogram data based on a previously trained neural network model; and predicting the likelihood of readmission of the user based on at least one of the first score and the second score.

[0008] Additionally, the step of predicting the likelihood of the user's re-hospitalization may include predicting the likelihood of the user's re-hospitalization based on the rate of change during a preset period of at least one of the first score and the second score.

[0009] Additionally, if it is identified that the rate of change of the first score is greater than or equal to the first value or the rate of change of the second score is greater than or equal to the second value during the aforementioned preset period, the method may include a step of predicting that the user has a possibility of re-hospitalization.

[0010] Additionally, the step of predicting the likelihood of re-hospitalization of the user may include the step of setting a first reference value for the first score and a second reference value for the second score, respectively, based on the average value of each of a plurality of first scores and a plurality of second scores obtained during a predetermined first period of a user with a history of inpatient treatment, and the step of predicting the likelihood of re-hospitalization of the user based on at least one comparison result obtained by comparing the first reference value with the first score obtained after the predetermined first period and comparing the second reference value with the second score obtained after the predetermined first period.

[0011] Additionally, the step of predicting the likelihood of re-hospitalization of the user may include predicting the likelihood of re-hospitalization of the user based on at least one of the approach rates of the first score and the second score, respectively, for the first reference value set for the left ventricular systolic dysfunction and the second reference value set for the user's left ventricular diastolic dysfunction.

[0012] In addition, the user may be a patient with a history of hospitalization for heart failure.

[0013] Additionally, the above-mentioned previously trained neural network model may include a first neural network model trained to calculate a score corresponding to the probability of left ventricular systolic dysfunction using electrocardiogram data as input, and a second neural network model trained to calculate a score corresponding to the probability of left ventricular diastolic dysfunction using electrocardiogram data as input.

[0014] Additionally, the method includes the step of obtaining a third score corresponding to myocardial infarction from the electrocardiogram data based on a previously trained neural network model, and the predicting step may include the step of predicting the likelihood of the user's re-hospitalization based on at least one of the first score, the second score, and the third score.

[0015] A computing device for determining whether a user with a history of inpatient treatment is re-hospitalized based on electrocardiogram data according to one embodiment of the present disclosure for solving the above-described problem comprises a memory including executable program codes and at least one core, wherein the device includes one or more processors that, by executing the program codes, acquire electrocardiogram data of a user with a history of inpatient treatment, acquire a first score corresponding to left ventricular systolic dysfunction of the user and a second score corresponding to left ventricular diastolic dysfunction of the user from the electrocardiogram data based on a previously learned neural network model, and predict the possibility of re-hospitalization of the user based on at least one of the first score and the second score.

[0016] A computer program stored on a computer-readable storage medium according to one embodiment of the present disclosure for solving the above-described problem, wherein the computer program, when executed on one or more processors, performs operations to determine whether a user with a history of inpatient treatment is re-hospitalized based on electrocardiogram data, and the operations include an operation of acquiring electrocardiogram data of a user with a history of inpatient treatment, an operation of acquiring a first score corresponding to left ventricular systolic dysfunction of the user and a second score corresponding to left ventricular diastolic dysfunction of the user from the electrocardiogram data based on a previously learned neural network model, and an operation of predicting the possibility of re-hospitalization of the user based on at least one of the first score and the second score.

[0017] According to one embodiment of the present disclosure, by analyzing changes in scores related to left ventricular systolic dysfunction (LVSD) and left ventricular diastolic dysfunction (LVDD) based on electrocardiogram data repeatedly acquired from a user, it is possible to detect the deterioration of a user's condition who has been hospitalized for heart failure at an early stage. Accordingly, it is possible to predict the risk of re-hospitalization and take medical action before the user's condition clinically deteriorates; consequently, this can reduce unnecessary emergency room visits or re-hospitalizations and improve the efficiency of heart failure patient management.

[0018] FIG. 1 is an exemplary diagram of a method for determining whether a user with a history of inpatient treatment will be readmitted based on electrocardiogram data according to one embodiment of the present disclosure.

[0019] FIG. 2 is a block diagram of a computing device that determines whether a user with a history of inpatient treatment is re-hospitalized based on electrocardiogram data according to one embodiment of the present disclosure.

[0020] FIG. 3 is a schematic flowchart of a method for determining whether a user with a history of inpatient treatment will be readmitted based on electrocardiogram data according to one embodiment of the present disclosure.

[0021] FIG. 4 is an example diagram of a method for determining whether a user with a history of inpatient treatment will be readmitted based on electrocardiogram data according to one embodiment of the present disclosure.

[0022] FIG. 5 is a block diagram of a computing device according to another embodiment of the present disclosure.

[0023] Embodiments of the present disclosure are described below with reference to the attached drawings so that those skilled in the art (hereinafter, those skilled in the art) can easily implement them. The embodiments presented in the present disclosure are provided to enable those skilled in the art to use or implement the contents of the present disclosure. Accordingly, various modifications to the embodiments of the present disclosure will be apparent to those skilled in the art. That is, the present disclosure may be embodied in various different forms and is not limited to the embodiments below.

[0024] Throughout the specification of the present disclosure, identical or similar reference numerals refer to identical or similar components. Additionally, to clearly explain the present disclosure, reference numerals in the drawings that are unrelated to the description of the present disclosure may be omitted.

[0025] The term “or” as used in this disclosure is intended to mean an implicit “or” rather than an exclusive “or.” That is, unless otherwise specified in this disclosure or its meaning is not clear from the context, “X uses A or B” should be understood to mean one of the natural implicit substitutions. For example, unless otherwise specified in this disclosure or its meaning is not clear from the context, “X uses A or B” may be interpreted as X using A, X using B, or X using both A and B.

[0026] The term “and / or” as used in this disclosure should be understood to refer to and include all possible combinations of one or more of the enumerated related concepts.

[0027] The terms “comprising” and / or “comprising” as used in this disclosure should be understood to mean the presence of certain features and / or components. However, the terms “comprising” and / or “comprising” should be understood not to exclude the presence or addition of one or more other features, other components and / or combinations thereof.

[0028] Where not otherwise specified in the present disclosure or where it is not clear from the context that the singular form indicates, the singular should generally be interpreted as including “one or more.”

[0029] The term "the N (N is a natural number)" used in this disclosure may be understood as an expression used to distinguish the components of this disclosure from one another according to certain criteria, such as functional perspectives, structural perspectives, or convenience of explanation. For example, components performing different functional roles in this disclosure may be distinguished as a first component or a second component. However, components that are substantially identical within the technical scope of this disclosure but must be distinguished for the convenience of explanation may also be distinguished as a first component or a second component.

[0030] The term “acquisition” as used in this disclosure may be understood to mean not only receiving data through a wired or wireless communication network with an external device or system, but also generating data in an on-device form.

[0031] Meanwhile, the terms "module" or "unit" used in this disclosure may be understood as referring to an independent functional unit that processes computing resources, such as a computer-related entity, firmware, software or a part thereof, hardware or a part thereof, or a combination of software and hardware. In this case, "module" or "unit" may be a unit composed of a single element, or a unit expressed as a combination or set of multiple elements. For example, in a narrow sense, "module" or "unit" may refer to a hardware element of a computing device or a set thereof, an application program that performs a specific function of software, a procedure implemented through software execution, or a set of instructions for program execution. Furthermore, in a broad sense, "module" or "unit" may refer to the computing device itself that constitutes the system, or an application executed on the computing device. However, since the above-described concept is merely an example, the concepts of "module" or "part" may be defined in various ways within the scope understandable to those skilled in the art based on the contents of this disclosure.

[0032] As used in this disclosure, the term "model" may be understood as a system implemented using mathematical concepts and language to solve a specific problem, a set of software units to solve a specific problem, or an abstract model regarding a processing process to solve a specific problem. For example, a neural network "model" may refer to an overall system implemented as a neural network that possesses problem-solving capabilities through learning. In this case, the neural network may possess problem-solving capabilities by optimizing parameters connecting nodes or neurons through learning. A neural network "model" may include a single neural network or a set of neural networks composed of multiple neural networks.

[0033] The term "data" as used in this disclosure may include "image," "signal," etc. The term "image" as used in this disclosure may refer to multidimensional data composed of discrete image elements. In other words, "image" may be understood as a term referring to a digital representation of an object visible to the human eye. For example, "image" may refer to multidimensional data composed of elements corresponding to pixels in a two-dimensional image. "Image" may refer to multidimensional data composed of elements corresponding to voxels in a three-dimensional image.

[0034] As used in this disclosure, "object" refers to an object or entity from which data is collected or analyzed. For example, in the context of medical data processing, "object" may include a person (patient or subject) from whom biosignal or image data is acquired. Furthermore, "object" may be extended to specific objects or structures in cases involving image processing or object recognition, and may include biological objects such as animals in addition to humans. In other words, "object" may encompass various forms of physical or biological objects depending on the purpose of analysis and refers to an entity that acquires data through biosignal measurement devices or sensors.

[0035] The explanation of the foregoing terms is intended to aid in understanding the present disclosure. Accordingly, it should be noted that unless a foregoing term is explicitly stated as a matter limiting the content of the present disclosure, it is not to be used in the sense of limiting the technical concept of the content of the present disclosure.

[0036] FIG. 1 is an exemplary diagram of a method for determining whether a user with a history of inpatient treatment will be readmitted based on electrocardiogram data according to one embodiment of the present disclosure.

[0037] A computing device (100) according to one embodiment of the present disclosure may be a hardware device or part of a hardware device that performs comprehensive processing and computation of data, or it may be a software-based computing environment connected to a communication network. For example, the computing device (100) may be a server that performs intensive data processing functions and is an entity that shares resources, or it may be a client that shares resources through interaction with the server. Additionally, the computing device (100) may be a cloud system in which a plurality of servers and clients interact to comprehensively process data. Since the description above is merely one example regarding the type of computing device (100), the type of computing device (100) may be configured in various ways within a range understandable to a person skilled in the art based on the contents of the present disclosure. For example, the computing device (100) may be implemented as various electronic devices such as a desktop, laptop, smartphone, smart watch, or server. Alternatively, the computing device (100) may be implemented as a biosignal measuring device that directly measures biosignals.

[0038] Referring to FIG. 1, a computing device (100) according to one embodiment of the present disclosure can acquire biometric data (10) from a patient (1), analyze the biometric data (10) using a previously learned neural network model (600), and determine the state of the user (1) based on the analysis results.

[0039] Here, the user (1) may be a patient who has a history of hospitalization treatment at a hospital (and other medical institutions, etc.) due to a disease. Patients with a history of hospitalization treatment include patients who are hospitalized and currently receiving treatment due to a specific disease, and patients who have been discharged after hospitalization. In particular, according to one embodiment of the present disclosure, the user (1) may be a patient who has a history of hospitalization due to heart failure (HF). That is, the user (1) may include users who are hospitalized and currently receiving treatment for heart failure or who have been discharged from the hospital. The state of the user (1) may be a health condition related to the user (1). For example, the state of the user (1) may include the presence or absence of a specific disease, the possibility of the manifestation of a specific disease, etc. In particular, according to one embodiment of the present disclosure, the computing device (100) can predict the possibility of the user (1) being re-hospitalized based on the state of the user (1).

[0040] The bio-data (10) may include various forms of data observed or measured from the subject to reflect the physiological state of the subject. The bio-data (10) may include bio-signals collected continuously as well as single physiological indicators measured at a specific point in time. For example, the bio-data (10) may include electrocardiogram (ECG) data, electroencephalogram (EEG) data, respiration, heart rate, blood pressure, oxygen saturation, body temperature, etc. However, for the convenience of explaining the present disclosure, the bio-data (10) will be described below as electrocardiogram data.

[0041] A computing device (100) can predict the condition of a patient with a history of hospitalization by analyzing biometric data (10) based on a pre-trained neural network model (600). In particular, the computing device (100) can input electrocardiogram data into a pre-trained neural network model (600) to obtain a score regarding the functional state of the user's heart, and can predict the likelihood of re-hospitalization of a user with a history of hospitalization due to a disease based on the obtained score.

[0042] According to one embodiment of the present disclosure, the neural network model (600) may be trained to predict the functional state of the user's heart rather than predicting the disease itself. More specifically, the neural network model (600) may be trained to identify the functional state of the user's heart rather than predicting whether the user has heart failure. Here, the functional state of the heart may include left ventricular systolic dysfunction (LVSD), left ventricular diastolic dysfunction (LVDD), myocardial infarction (MI), etc. Furthermore, the neural network model (600) may be trained to identify pathological states corresponding to cardiac functional abnormalities such as left ventricular systolic dysfunction (LVSD), left ventricular diastolic dysfunction (LVDD), and myocardial infarction (MI), and to calculate a score that quantitatively represents the pathological state.

[0043] Referring again to FIG. 1, the computing device (100) obtains a score corresponding to left ventricular systolic dysfunction and a score corresponding to left ventricular diastolic dysfunction from a neural network model (600), and can predict the likelihood of re-hospitalization of user (1) (i.e., a user with a history of hospitalization for heart failure) based on the obtained scores. In particular, the computing device (100) can repeatedly obtain scores regarding user (1) and predict the likelihood of re-hospitalization of user (1) by identifying the amount of change in the scores or the trend of the scores. Predicting the likelihood of re-hospitalization of user (1) may involve predicting whether the risk of re-hospitalization is inherently increasing due to changes in the functional state of user (1)'s heart, even though the disease of user (1) has not manifested.

[0044] Hereinafter, embodiments of the present disclosure related thereto will be described in detail based on FIGS. 2 to 5.

[0045] FIG. 2 is a block diagram of a computing device that determines whether a user with a history of inpatient treatment is re-hospitalized based on electrocardiogram data according to one embodiment of the present disclosure.

[0046] Referring to FIG. 2, a computing device (100) according to one embodiment of the present disclosure may include a processor (110), a memory (120), and a communication interface (130). However, since FIG. 2 is merely an example, the computing device (100) may include other configurations for implementing a computing environment. Additionally, only some of the disclosed configurations may be included in the computing device (100).

[0047] A processor (110) according to one embodiment of the present disclosure may be understood as a constituent unit comprising hardware and / or software for performing computing operations. For example, the processor (110) may read a computer program and perform data processing for machine learning. The processor (110) may process computational processes such as processing input data for machine learning, extracting features for machine learning, and calculating errors based on backpropagation. A processor (110) for performing such data processing may include a central processing unit (CPU), a general purpose graphics processing unit (GPGPU), a tensor processing unit (TPU), an application specific integrated circuit (ASIC), or a field programmable gate array (FPGA). Since the above-described types of processors (110) are merely examples, the types of processors (110) may be configured in various ways within a range understandable to those skilled in the art based on the contents of the present disclosure.

[0048] The processor (110) is connected to other components of the computing device (100) (e.g., memory (120) and communication interface (130)) to control the overall operation of the computing device (100).

[0049] A memory (120) according to one embodiment of the present disclosure may be understood as a configuration unit comprising hardware and / or software for storing and managing data processed by a computing device (100). That is, the memory (120) may store data of any form generated or determined by a processor (110) and data of any form received by a communication interface (130) of the computing device (100).

[0050] To this end, the memory (120) may include at least one type of storage medium among flash memory type, hard disk type, multimedia card micro type, card type memory, RAM (random access memory), SRAM (static random access memory), ROM (read-only memory), EEPROM (electrically erasable programmable read-only memory), PROM (programmable read-only memory), magnetic memory, magnetic disk, and optical disk. Additionally, the memory (120) may include a database system that controls and manages data in a predetermined system. Since the above-described types of memory (120) are merely examples, the types of memory (120) may be configured in various ways within a range understandable to those skilled in the art based on the contents of the present disclosure.

[0051] The memory (120) can structure and organize data, combinations of data, and program code executable by the processor (110) that are necessary for the processor (110) to perform operations. For example, the memory (120) can store multiple biometric data received through a communication interface and information of an object corresponding to the biometric data.

[0052] Additionally, the memory (120) may store a neural network model trained to identify the functional state of the user's (1) heart (e.g., left ventricular systolic dysfunction, left ventricular diastolic dysfunction, myocardial infarction, etc.), program code that operates to perform learning of the neural network model, program code that operates the neural network model to receive biological data input and perform inference according to the purpose of use of the computing device (100), and output data or processed data generated as the program code is executed.

[0053] The communication interface (130) can be understood as a configuration unit that transmits and receives data through any known form of wired or wireless communication system. For example, the communication interface (130) can perform data transmission and reception using wired or wireless communication systems such as a local area network (LAN), wideband code division multiple access (WCDMA), long term evolution (LTE), wireless broadband internet (WiBro), 5th generation mobile communication (5G), ultrawide-band wireless communication, ZigBee, radio frequency (RF) communication, wireless LAN, wireless fidelity (Wi-Fi), near field communication (NFC), or Bluetooth. Since the communication systems described above are merely examples, wired or wireless communication systems for data transmission and reception of the communication interface (130) can be applied in various ways other than the examples described above. The communication interface (130) can receive data necessary for the processor (110) to perform calculations through wired or wireless communication with any system or any client. Additionally, the communication interface (130) can transmit data generated through the calculations of the processor (110) through wired or wireless communication with any system or any client. For example, the communication interface (130) can receive biometric data and medical data through communication with a database within a hospital environment, a cloud server performing tasks such as standardization of medical data, or a computing device.The communication interface (130) can transmit output data of the neural network model (600) and intermediate data, processed data, etc. derived from the computation process of the processor (110) through communication with the aforementioned database, server, or computing device.

[0054] In particular, the processor (110) can obtain biometric data from an external biometric signal measuring device (e.g., an electrocardiogram measuring device), a computing device (e.g., a smart watch), etc. through a communication interface.

[0055] FIG. 3 is a schematic flowchart of a method for determining whether a user with a history of inpatient treatment is re-hospitalized based on electrocardiogram data according to one embodiment of the present disclosure. FIG. 4 is an example diagram of a method for determining whether a user with a history of inpatient treatment is re-hospitalized based on electrocardiogram data according to one embodiment of the present disclosure.

[0056] According to one embodiment of the present disclosure, the processor can acquire electrocardiogram data of a user (1) who has a history of hospitalization (S310).

[0057] Specifically, the processor can identify a user (1) who has a history of inpatient treatment. The processor can identify the user (1)'s hospital visit history, medical records, surgery records, etc. through the user's (1) identification information (name, resident registration number, patient number, ID, etc.). Additionally, the processor can determine whether the user (1) is a user (1) who has a history of inpatient treatment. The processor can also identify the cause of the user's (1) hospitalization, such as the type of disease or the cause of injury. In particular, the processor can identify a user (1) who has a history of inpatient treatment due to heart failure.

[0058] The processor can acquire electrocardiogram data of a user (1) who has a history of hospitalization. Specifically, the processor (110) can acquire electrocardiogram data (11) obtained by an external biosignal measuring device connected to a computing device (e.g., an external biosignal measuring device linked to the electronic device (100), a smart watch, etc.) through a communication interface (130). Alternatively, the processor (110) may directly acquire the electrocardiogram data (11) of the user (1) through a sensing unit included in the electronic device (100). Specifically, the processor (110) can acquire electrocardiogram data (11) through a plurality of electrodes of a sensing unit attached to the body of the user (1). Meanwhile, the electrocardiogram data (11) may be data in which an electrocardiogram signal measured by a single lead is digitally processed. However, it is not limited to this, and the electrocardiogram data (11) may include data obtained from a 12-lead electrocardiogram signal and a 6-lead electrocardiogram signal.

[0059] When an electrocardiogram signal is measured from a user (1), the processor (110) can generate multiple electrocardiogram data (11) by dividing the electrocardiogram signal into pre-set lengths or by applying a window of pre-set length on the time axis. For example, the electrocardiogram data (11) can be obtained by sampling an electrocardiogram signal measured for 10 seconds using 12 leads at 500 points per second. The processor (110) can be configured into multiple input unit electrocardiogram data (11) by dividing them by time intervals. However, the generation of such electrocardiogram data (11) can be performed by an external biosignal measuring device as described above.

[0060] And, the processor can obtain a first score corresponding to left ventricular systolic dysfunction of the user (1) and a second score corresponding to left ventricular diastolic dysfunction of the user (1) from the electrocardiogram data based on a previously learned neural network model (600) (S320).

[0061] Specifically, the processor (110) inputs electrocardiogram data into a pre-trained neural network model (600) to extract potential feature information related to the functional state of the user's (1) heart from the electrocardiogram data, and obtains a result value that classifies the functional state of the user's (1) heart based on the extracted electrocardiogram data. The functional state of the user's (1) heart can be classified into multiple grades, severity, etc., and the processor (110) can determine the functional state of the user's (1) heart by comparing the result value with multiple reference values ​​set to distinguish each grade (or severity).

[0062] The processor (110) can input electrocardiogram data into a pre-trained neural network model (600) to obtain a score corresponding to the functional state of the user's (1) heart. At this time, the processor (110) can obtain a first score corresponding to left ventricular systolic dysfunction and a second score corresponding to left ventricular diastolic dysfunction. Here, the first score may be a numerical value representing the probability that the functional state of the user's (1) heart corresponds to left ventricular systolic dysfunction or the probability of left ventricular systolic dysfunction occurring. The second score may be a numerical value representing the probability that the functional state of the user's (1) heart corresponds to left ventricular diastolic dysfunction or the probability of left ventricular systolic dysfunction occurring.

[0063] According to one embodiment of the present disclosure, a neural network model that has been trained may include a first neural network model trained to calculate a score corresponding to the possibility of left ventricular systolic dysfunction using electrocardiogram data as input, and a second neural network model trained to calculate a score corresponding to the possibility of left ventricular diastolic dysfunction using electrocardiogram data as input. Accordingly, with reference to FIG. 4, the processor may input electrocardiogram data obtained to the first neural network model (600-1) to obtain a first score, and input electrocardiogram data obtained to the second neural network model (600-2) to obtain a second score.

[0064] A processor (110) can train a first neural network model (600-1) to identify the heart condition related to left ventricular systolic dysfunction of a user (1) based on a training data set containing multiple electrocardiogram data (11) obtained from multiple subjects. At this time, the training data set may include label data in which a label regarding the presence or absence of left ventricular systolic dysfunction (or a risk group for left ventricular systolic dysfunction) is assigned in correspondence with the electrocardiogram data (11) of the input data. The processor (110) inputs the input data (i.e., electrocardiogram data (11)) included in the training data set into the first neural network model (600-1) and can calculate a loss function based on the difference between the output value of the first neural network model (600-1) and the label during the training process. The loss function may be defined as a cross-entropy loss or an objective function that optimizes the balance between precision and recall. Based on the calculated loss function, the processor (110) can adjust the weights of the first neural network model (600-1) through backpropagation. By repeating this process, the processor (110) can improve the classification performance of the functional state of the user (1)'s heart related to left ventricular systolic dysfunction of the first neural network model (600-1), and finally obtain a neural network model (600) trained to predict whether the user (1) has left ventricular systolic dysfunction based on electrocardiogram data (11). The processor (110) can also obtain a first score corresponding to the probability of left ventricular systolic dysfunction from the previously trained first neural network model (600-1).

[0065] Meanwhile, the processor (110) can train a second neural network model (600-2) to identify the heart condition related to left ventricular diastolic dysfunction of the user (1) based on a training data set containing multiple electrocardiogram data (11) obtained from multiple subjects. At this time, the training data set may include label data in which a label regarding the presence or absence of left ventricular diastolic dysfunction (or a risk group for left ventricular diastolic dysfunction) is assigned in correspondence with the electrocardiogram data (11) of the input data. In addition, regarding the learning method, the embodiments of the present disclosure described above may be applied in the same way.

[0066] The neural network model (600) can be implemented as a multi-layer perceptron (MLP), convolutional neural network (CNN), recurrent neural network (RNN), generative adversarial network, etc.

[0067] The neural network model (600) may include a plurality of residual blocks that extract potential features of input biological data (10) and a classifier that classifies specific diseases based on the extracted potential features. Meanwhile, if the electrocardiogram data (11) is 12-lead electrocardiogram data (11) (or 6-lead electrocardiogram data (11)), the processor (110) may preprocess the electrocardiogram signals measured at each of the 12 leads by aligning or combining them with each other, and matching the time intervals of the multiple lead signals, to form a single input tensor that fits the input format of the neural network model (600).

[0068] In addition, the neural network model (600) may include multiple neural network models (600) according to the lead method in which the electrocardiogram data is acquired (e.g., 12 leads, 6 leads, or a single lead). At this time, each neural network model (600) may have a different number of electrocardiogram data combined and input according to the lead method. The processor (110) may select one of the multiple neural network models (600) according to the lead method of the acquired electrocardiogram data, analyze the electrocardiogram data, and acquire first and second scores.

[0069] Meanwhile, the processor (110) can predict the possibility of the user (1) being re-hospitalized based on at least one of the first score and the second score (S330). Specifically, the processor (110) can predict the possibility of the user (1) being re-hospitalized by identifying changes and trends in the first score and the second score. To this end, the first score and the second score previously obtained from the user (1) may be stored in memory. The processor (110) can predict the possibility of the user (1) being re-hospitalized by comparing the user (1)'s first score obtained in the past with the user (1)'s first score obtained in real time. Similarly, the processor (110) can predict the possibility of the user (1) being re-hospitalized by comparing the user (1)'s second score obtained in the past with the user (1)'s second score obtained in real time.

[0070] In particular, according to one embodiment of the present disclosure, the processor (110) can predict the likelihood of the user (1) being re-hospitalized based on the rate of change of at least one of the first score and the second score during a preset period. Specifically, the processor (110) can repeatedly obtain the first score and the second score during a preset period from the user (1) who has a history of hospitalization treatment, and identify the rate of change of the first score and the second score during the preset period.

[0071] At this time, if the processor (110) identifies that the rate of change of the first score is greater than or equal to the first value or the rate of change of the second score is greater than or equal to the second value during the aforementioned preset period, it can predict that the user (1) is highly likely to be hospitalized again.

[0072] Specifically, the processor (110) can repeatedly obtain a first score and a second score for the user (1) at a preset interval (e.g., every 5 hours, every day, or every week). At this time, the processor (110) can calculate the rate of change of each score increasing or decreasing using the scores accumulated over time. For example, the processor (110) can calculate the average rate of change using the first score (LVSD score) obtained over the past week. Similarly, the processor (110) can calculate the average rate of change over the past week for the second score (LVDD score). At this time, if the processor (110) identifies that the calculated rate of change of the first score increases by more than a preset first value, or identifies that the rate of change of the second score increases by more than a preset second value, it determines that the condition of the user's (1) heart is outside the normal range and, based on this, predicts that the user (1) is highly likely to be re-hospitalized.

[0073] Meanwhile, according to one embodiment of the present disclosure, the processor (110) sets a first reference value for the first score and a second reference value for the second score based on the average value of each of a plurality of first scores and a plurality of second scores obtained during a predetermined first period of a user (1) who has a history of hospitalization treatment, compares the first reference value with the first score obtained after the predetermined first period, and can predict the possibility of re-hospitalization of the user (1) based on at least one comparison result obtained by comparing the second reference value with the second score obtained after the predetermined first period.

[0074] Specifically, the processor (110) may calculate an average value for each of the multiple first scores and multiple second scores obtained during a pre-set period (e.g., 7 or 14 days immediately after discharge) after the user (1) is discharged, and set this as the first reference value and the second reference value for the user (1). The processor (110) may compare each score obtained after the reference set period with the reference value to determine whether the score repeatedly increases in a direction deviating from the reference value. For example, if the first score newly obtained after the reference set period is continuously maintained at a higher state than the reference value, or if the second score increases stepwise relative to the reference value, the processor (110) may determine that the user (1)'s cardiac function is on a deteriorating trend and predict that there is a high probability of re-hospitalization.

[0075] Meanwhile, the processor (110) calculates the rate of change of the first score during a preset period above the first reference value, and if the average rate of change of the first score is above the preset first value and the difference between the first score and the first reference value is above the preset third value, it can predict that the user (1) is highly likely to be re-hospitalized.

[0076] According to one embodiment of the present disclosure, the processor (110) can predict the likelihood of the user (1) being re-hospitalized based on at least one of the approach rates of the first score and the second score, respectively, for the first cut-off value set for the left ventricular systolic dysfunction and the second cut-off value set for the left ventricular diastolic dysfunction of the user (1).

[0077] The processor (110) can calculate the speed at which the first score and the second score approach a preset cutoff value for each. The cutoff value may be a reference value for determining the high-risk group of the first score and the second score. The processor (110) can determine whether the speed of convergence to the cutoff value is gradually increasing by comparing the time taken for the user (1) to approach the first cutoff value (or the second cutoff value) within 10% and the time taken to approach it further within 5% thereafter.

[0078] If the processor (110) determines that the speed of access to the cutoff value corresponding to each of the first score or second score of a preset unit time (or unit number of times) is greater than the preset threshold speed, it evaluates that the score is likely to exceed the reference value within a short period of time and predicts that there is a possibility of re-hospitalization for the user (1).

[0079] Meanwhile, according to one embodiment of the present disclosure, the processor (110) can identify the possibility of re-hospitalization of the user (1) by identifying the rate of change of the first score and the second score when the first score and the second score are each greater than or equal to the cutoff value. Specifically, if the processor (110) identifies that the user's first score is greater than or equal to the first cutoff value or the second score is greater than or equal to the second cutoff value, it can determine the possibility of re-hospitalization of the user (1) by calculating the rate of increase of the first score over a preset unit time or the rate of increase of the second score over a preset unit time. If the processor identifies that the rate of increase of the first score over a unit time is greater than or equal to the first value after the first score exceeds the first cutoff, or identifies that the rate of increase of the second score over a unit time is greater than or equal to the second value after the second score exceeds the second cutoff, it can predict that there is a possibility of re-hospitalization of the user (1). For example, the processor (110) determines that there is no possibility of re-hospitalization because the access speed to the first cut-off value of the first score of the user (1) is lower than the preset threshold speed, but if the identified first score is subsequently identified as increasing at a rate greater than the first value after exceeding the first cut-off value, it can determine that there is a possibility of re-hospitalization of the user (1).

[0080] Meanwhile, the processor (110) calculates the number of increases in the first score (or second score) after the cut-off in addition to the increase rate, and if it is identified that the number of increases is greater than or equal to the reference number, it can determine that there is a possibility of the user (1) being re-hospitalized.

[0081] Meanwhile, according to one embodiment of the present disclosure, the processor (110) obtains a third score corresponding to myocardial infarction from the electrocardiogram data based on a previously learned neural network model (600), and can predict the possibility of re-hospitalization of the user (1) based on at least one of the first score, the second score, and the third score.

[0082] The processor (110) may further obtain a third score corresponding to myocardial infarction in relation to the functional state of the user's (1) heart and predict the likelihood of the user's (1) re-hospitalization based on at least one of the first score, the second score, and the third score. To this end, the neural network model (600) may further include a third neural network model trained to calculate the third score.

[0083] The processor (110) can predict the likelihood of the user (1) being re-hospitalized based on the rate of change of the third score during a preset period, the difference from the average value of the third score obtained during a preset period of the third score, and the speed of approach to the cut-off value set for the third score.

[0084] Meanwhile, if the processor (110) predicts that there is a high probability of the user (1) being readmitted, it may provide such information to the medical staff. Additionally, the processor (110) may suggest changes to the user (1)'s treatment plan. For example, if the processor (110) determines that there is a high probability of the user (1) being readmitted after hospitalization, it may provide information regarding the likelihood of readmission of the user (1) to the medical staff's terminal device and provide information suggesting an extension of hospitalization. Alternatively, if the processor (110) determines that there is a high probability of the user (1) being readmitted based on electrocardiogram data obtained from the user (1)'s smartwatch after hospitalization and discharge, it may provide information suggesting a hospital visit and treatment to the user's terminal device.

[0085] FIG. 5 is a block diagram of a computing device (500) according to another embodiment of the present disclosure.

[0086] Referring to FIG. 5, a computing device (500) according to another embodiment of the present disclosure includes a processor (510), memory (520), a communication interface (530), a sensing unit (540), a display (550), a user (1) interface (560), a camera (570), and a speaker (580). Among the configurations shown in FIG. 9, the processor (510), memory (520), and communication interface (530) correspond to the processor (110), memory (120), and communication interface (130) of the computing device (100) shown in FIG. 2, so a detailed description is omitted.

[0087] A sensing unit (540) according to one embodiment of the present disclosure can acquire biometric information about a user (1). For example, the sensing unit (540) may include at least one electrode. In this case, the processor (110) can acquire an electrocardiogram signal of the user (1) through at least one electrode. Additionally, the sensing unit (540) may include an image sensor or a light sensor. In this case, the processor (110) can acquire biometric information such as a photo-blood flow signal through the image sensor (or light sensor). Furthermore, the sensing unit (540) may further include a PPG sensor, a blood pressure sensor, a body temperature sensor, etc., and through this, various biometric information such as photo-blood flow, blood pressure, and body temperature about the user (1) can be acquired.

[0088] The display (550) can display various images. Here, the images include both still images and videos. The display (550) can output guide information based on the diagnosis results of the user (1). The display (550) can be implemented as various types of displays such as LCD (Liquid Crystal Display Panel), OLED (Organic Light Emitting Diodes), LCoS (Liquid Crystal on Silicon), DLP (Digital Light Processing), etc. Additionally, the display (550) may also include a driving circuit, a backlight unit, etc., which can be implemented in forms such as a-si TFT, LTPS (low temperature poly silicon) TFT, OTFT (organic TFT), etc. For example, the display (550) can display judgment results and guide information based on the first and second scores.

[0089] Meanwhile, the display (550) may be implemented as a touch screen by combining it with a touch panel, and the display (550) can perform the function of an input interface that receives touch input from a user (1) as well as an output interface that outputs an image through the touch screen.

[0090] The user interface (560) is a configuration used by the computing device (100) to perform interaction with the user (1), and may include at least one of a touch sensor, a motion sensor, a button, a jog dial, and a switch, but is not limited thereto. The processor (510) can receive diet information, exercise information, activity information, medication information, etc. through the user interface (560).

[0091] The camera (570) captures objects around the user (1) to obtain images of the objects. Specifically, the camera (570) can obtain images of food consumed by the user (1). At this time, the processor (510) can determine the user's (1) diet information based on the images of food consumed by the user (1), or determine the user's (1) medication information based on the images of medication taken by the user (1). In addition, for this purpose, the camera (570) can be implemented with an image sensor having a CMOS structure (CIS, CMOS Image Sensor), an image sensor having a CCD structure (Charge Coupled Device), etc. However, it is not limited thereto, and the camera (570) can be implemented as a camera module of various resolutions capable of capturing a subject. Meanwhile, the camera (570) can be implemented as a depth camera (e.g., IR depth camera, etc.), a stereo camera, or an RGB camera, etc.

[0092] The speaker (580) is configured to output various audio data, for which various processing operations such as decoding, amplification, and noise filtering have been performed by an audio processing unit (not shown). The speaker (580) can output various notification sounds or voice messages. According to one embodiment of the present disclosure, the processor (510) can convert an electrical signal received from an external device into the voice of the user (1) and output it through the speaker (580). For example, the speaker (580) can output the user's (1) guide information related to context information as a voice message.

[0093] The microphone (590) can receive the user's voice. The processor (510) can obtain control information and context information related to the user (1) by analyzing the user's voice or surrounding sounds obtained through the microphone (590).

[0094] Meanwhile, according to one embodiment of the present disclosure, a non-transitory computer-readable medium may be provided that stores a program for determining whether a user with a history of inpatient treatment should be readmitted based on electrocardiogram data. Here, a non-transitory computer-readable medium refers to a medium that stores data semi-permanently and is readable by a device, rather than a medium that stores data for a short period of time, such as a register, cache, or memory. Specifically, the various applications or programs described above may be provided by being stored on a non-transitory computer-readable medium such as a CD, DVD, hard disk, Blu-ray disc, USB, memory card, ROM, etc.

[0095] The various embodiments of the present disclosure described above may be combined with additional embodiments and modified to the extent understandable to those skilled in the art in light of the detailed description above. The embodiments of the present disclosure are illustrative in all respects and should be understood as not restrictive. For example, each component described as a single unit may be implemented in a distributed manner, and components described as distributed may likewise be implemented in a combined form. Accordingly, all modifications or variations derived from the meaning, scope, and equivalents of the claims of the present disclosure should be interpreted as being included within the scope of the present disclosure.

Claims

A method for determining whether a user with a history of inpatient treatment is to be readmitted based on electrocardiogram data, performed by a computing device comprising at least one processor, A step of acquiring electrocardiogram data of a user with a history of inpatient treatment; A step of obtaining a first score corresponding to the user's left ventricular systolic dysfunction and a second score corresponding to the user's left ventricular diastolic dysfunction from the electrocardiogram data based on a previously trained neural network model; and A step of predicting the likelihood of the user's re-hospitalization based on at least one of the first score and the second score; comprising method. In paragraph 1, The step of predicting the likelihood of the above user's re-hospitalization is, A step of predicting the likelihood of the user being re-hospitalized based on the rate of change during a preset period of at least one of the first score and the second score; comprising method. In paragraph 2, A step of predicting that there is a possibility of the user being re-hospitalized if it is identified that the rate of change of the first score is greater than or equal to the first value or the rate of change of the second score is greater than or equal to the second value during the aforementioned preset period; method. In paragraph 1, The step of predicting the likelihood of the above user's re-hospitalization is, A step of setting a first reference value for the first score and a second reference value for the second score, respectively, based on the average value of each of a plurality of first scores and a plurality of second scores obtained during a preset first period of a user with the above-mentioned history of inpatient treatment; and The method comprises the step of predicting the likelihood of the user's re-hospitalization based on at least one comparison result obtained by comparing the first reference value with the first score obtained after the first pre-set period, and comparing the second reference value with the second score obtained after the first pre-set period. method. In paragraph 1, The step of predicting the likelihood of the above user's re-hospitalization is, A step of predicting the likelihood of re-hospitalization of the user based on at least one of the access rates of the first score and the second score, respectively, for the first reference value set for the left ventricular systolic dysfunction and the second reference value set for the user's left ventricular diastolic dysfunction; method. In paragraph 1, The above user, A patient who received inpatient treatment for heart failure, method. In paragraph 1, The above-mentioned previously trained neural network model is, A first neural network model trained to calculate a score corresponding to the probability of left ventricular systolic dysfunction using electrocardiogram data as input, and a second neural network model trained to calculate a score corresponding to the probability of left ventricular diastolic dysfunction using electrocardiogram data as input, method. In paragraph 1, The method includes the step of obtaining a third score corresponding to myocardial infarction from the electrocardiogram data based on a previously trained neural network model; The above-mentioned prediction step is, A step of predicting the likelihood of the user being re-hospitalized based on at least one of the first score, the second score, and the third score; comprising method. In a computing device that determines whether a user with a history of inpatient treatment is re-hospitalized based on electrocardiogram data, Memory containing executable program code; and A processor comprising at least one core, wherein by executing the program code, the processor obtains electrocardiogram data of a user with a history of hospitalization, obtains a first score corresponding to the user's left ventricular systolic dysfunction and a second score corresponding to the user's left ventricular diastolic dysfunction from the electrocardiogram data based on a pre-trained neural network model, and predicts the likelihood of the user's re-hospitalization based on at least one of the first score and the second score. Computing device. A computer program stored on a computer-readable storage medium, wherein the computer program, when executed on one or more processors, performs operations to determine whether a user with a history of inpatient treatment is re-admitted based on electrocardiogram data, and The above operations are, Action of acquiring electrocardiogram data of a user with a history of inpatient treatment; The operation of obtaining a first score corresponding to the user's left ventricular systolic dysfunction and a second score corresponding to the user's left ventricular diastolic dysfunction from the electrocardiogram data based on a previously trained neural network model; and An operation to predict the likelihood of the user being re-hospitalized based on at least one of the first score and the second score; comprising Computer program.

Images