Assistance apparatus, assistance system, and program for electrocardiogram analysis

Abstract

This assistance system comprises an electrocardiograph, an electrocardiogram analysis apparatus, and an assistance apparatus. The electrocardiograph is configured to acquire an electrocardiogram of a subject. The electrocardiogram analysis apparatus is configured to receive the electrocardiogram, divide the electrocardiogram into a plurality of sections, and extract sections other than a specific section from the plurality of sections as candidate sections. The assistance apparatus is configured to communicate with and connect to the electrocardiogram analysis apparatus. The assistance apparatus comprises a computing device and a display apparatus that is controlled by the computing device. The functions of the computing device are described in detail in a specification.

Description

Support device, support system, and program for electrocardiogram analysis

[0001] One embodiment of the present invention relates to an electrocardiogram analysis support device and support system, and a method for analyzing an electrocardiogram. Alternatively, one embodiment of the present invention relates to a program for supporting electrocardiogram analysis.

[0002] Electrocardiograms (ECGs) are widely used as a means of assessing a patient's health condition. However, depending on the type of disease, long-term (e.g., one day) measurement is required. Furthermore, even after long-term measurement, waveforms due to the disease may not clearly appear in the ECG. For this reason, systems have recently been developed that use supervised learning models to automatically determine whether or not an ECG contains signs of heart disease and automatically determine the probability of its onset (see Patent Documents 1 and 2).

[0003] Patent No. 6865329 Patent No. 7002168

[0004] An object of one embodiment of the present invention is to provide a novel assistance device for assisting electrocardiogram analysis, an electrocardiogram analysis system including the assistance device, a program for assisting electrocardiogram analysis using the assistance device, and an electrocardiogram analysis method using the electrocardiogram analysis system.Alternatively, an object of one embodiment of the present invention is to provide an assistance device for electrocardiogram analysis that can determine a disease or its symptoms with high accuracy, an electrocardiogram analysis system including the assistance device, a program for assisting electrocardiogram analysis using the assistance device, and an electrocardiogram analysis method using the electrocardiogram analysis system.

[0005] One embodiment of the present invention is an assistance system for electrocardiogram analysis. The assistance system includes an electrocardiograph, an electrocardiogram analyzer, and an assistance device. The electrocardiograph is configured to acquire an electrocardiogram of a subject. The electrocardiogram analyzer is configured to receive the electrocardiogram, divide the electrocardiogram into a plurality of intervals, and extract intervals from the plurality of intervals other than a specific interval as candidate intervals. The assistance device is configured to communicate with the electrocardiogram analyzer. The assistance device includes a computing device and a display device controlled by the computing device. The computing device is configured to (1) accept a first user command to upload the electrocardiogram to the electrocardiogram analyzer, (2) display the candidate intervals on the screen of the display device, (3) accept input of a second user command to select a non-analysis interval from the candidate intervals, and (4) send an instruction to the electrocardiogram analyzer to analyze the electrocardiogram using an analysis interval, which is an interval from the candidate interval not selected by the second user command. The specific interval includes an interval showing a waveform caused by a disease.

[0006] One embodiment of the present invention is an electrocardiogram analysis support device. The electrocardiogram analysis support device includes a computing device and a display device controlled by the computing device. The computing device is configured to (1) receive a first user command to cause an electrocardiogram analyzer, communicatively connected to the computing device and configured to divide an electrocardiogram into multiple intervals and extract intervals other than a specific interval from the multiple intervals as candidate intervals, to upload the electrocardiogram of the subject to the computing device; (2) display the candidate intervals on the screen of the display device; (3) receive a second user command to select a non-analysis interval from the candidate intervals; and (4) send an instruction to the electrocardiogram analyzer to analyze the electrocardiogram of the subject using an analysis interval, which is an interval from the candidate intervals not selected by the second user command. The specific interval includes an interval showing a waveform attributable to a disease.

[0007] One embodiment of the present invention is a program for supporting electrocardiogram analysis performed by an electrocardiogram analyzer using a computing device. The electrocardiogram analyzer is communicatively connected to the computing device and configured to divide a subject's electrocardiogram into multiple intervals and extract intervals from the multiple intervals other than a specific interval as candidate intervals. The program is configured to cause the computing device to (1) accept a first user command to upload the subject's electrocardiogram to the electrocardiogram analyzer, (2) display the candidate intervals on a screen of a display device of the computing device, (3) accept a second user command to select a non-analysis interval from the candidate intervals, and (4) send an instruction to the electrocardiogram analyzer to analyze the electrocardiogram using analysis intervals that are the candidate intervals not selected by the second user command.

[0008] One embodiment of the present invention is a computer-readable storage medium on which the above program is recorded.

[0009] One embodiment of the present invention is a method for analyzing an electrocardiogram, which includes (1) acquiring an electrocardiogram of a subject, (2) dividing the electrocardiogram into a plurality of intervals, (3) extracting intervals other than a specific interval from the plurality of intervals as candidate intervals, (4) selecting non-analysis intervals from the candidate intervals, and (5) analyzing the electrocardiogram using analysis intervals, which are intervals from the candidate intervals that have not been selected as non-analysis intervals.

[0010] 1 is a block diagram of a support system for electrocardiogram analysis according to one embodiment of the present invention. FIG. 2 is a flow chart of electrocardiogram analysis using the support system for electrocardiogram analysis according to one embodiment of the present invention. FIG. 3 is a schematic diagram explaining operations performed in the support system for electrocardiogram analysis according to one embodiment of the present invention. FIG. 4 is an example of a display displayed on a display device of a support device according to one embodiment of the present invention. FIG. 5 is an example of a display displayed on a display device of a support device according to one embodiment of the present invention. FIG. 6 is an example of a display displayed on a display device of a support device according to one embodiment of the present invention. FIG. 7 is an example of a display displayed on a display device of a support device according to one embodiment of the present invention.

[0011] Hereinafter, various embodiments of the present invention will be described with reference to the drawings, etc. However, the present invention can be embodied in various forms without departing from the spirit of the present invention, and should not be construed as being limited to the description of the embodiments exemplified below.

[0012] In order to clarify the description, the drawings may show the width, thickness, shape, etc. of each part schematically compared to the actual embodiment, but these are merely examples and do not limit the interpretation of the present invention. In this specification and each drawing, elements having the same functions as those explained in the previous drawings may be assigned the same reference numerals, and duplicate explanations may be omitted.

[0013] 1. Configuration of the Support System for Electrocardiogram Analysis FIG. 1 shows a block diagram of a support system 100 for electrocardiogram analysis according to one embodiment of the present invention. As shown in FIG. 1, the support system 100 includes a support device 110, an electrocardiogram analysis device 140, and an electrocardiograph 150. The support system 100 may further include a database 160. There is no limit to the number of electrocardiographs 150, electrocardiogram analysis devices 140, and support devices 110 included in the support system 100. For example, the support system 100 may include multiple electrocardiographs 150, multiple support devices 110, and multiple databases 160. The support system 100 may also include multiple electrocardiogram analysis devices 140. The support device 110 is controlled by a program (described below) that operates according to user commands. The support system 100 may further include a heart rate monitor 170 for measuring the user's heart rate, an acceleration sensor 180 for detecting the user's condition, and the like.

[0014] The assistance device 110 and the electrocardiogram analysis device 140 are communicatively connected to each other via a network 102. An example of the network 102 is a wide area network such as the Internet. If the assistance system 100 includes a database 160, the database 160 may also be communicatively connected to the assistance device 110 and the electrocardiogram analysis device 140 via the network 102. The electrocardiograph 150 may also be communicatively connected to the assistance device 110, the electrocardiogram analysis device 140, and / or the database 160 via the network 102, or may not be configured to communicate via the network 102. In the latter case, electrocardiogram data acquired by the electrocardiograph 150 may be stored in the assistance device 110 and / or the database 160 via a storage medium such as a flash memory.

[0015] The electrocardiograph 150 is a device that acquires an electrocardiogram of the subject, and a known electrocardiograph can be used as the electrocardiograph 150. When acquiring an electrocardiogram over a long period of time, a Holter monitor or an event recorder, which places less strain on the subject when measuring, may be used as the electrocardiograph 150. The electrocardiograph 150 acquires potential changes caused by the flow of electricity within the heart, and provides an electrocardiogram as a plot of potential changes against measurement time.

[0016] The electrocardiogram analysis device 140 is a device that analyzes electrocardiogram data acquired by the electrocardiograph 150. The electrocardiogram analysis device 140 is a device equipped with computing and communication functions, and may be a personal computer or an application server. The functions and operations of the electrocardiogram analysis device 140 will be described later. Based on the analysis results obtained by the electrocardiogram analysis device 140, it is possible to determine whether a subject has various diseases or test abnormalities. The diseases and test abnormalities that the electrocardiogram analysis device 140 can detect are not limited to cardiac diseases, and the electrocardiogram analysis device 140 can detect various diseases and test abnormalities. For example, the electrocardiogram analysis device 140 can detect not only cardiac diseases and test abnormalities such as atrial fibrillation, heart failure, ventricular arrhythmia, valvular disease, and cardiomyopathy, but also test abnormalities that may indicate various diseases such as hyponatremia, diabetes, and hyperkalemia from the electrocardiogram data of the subject.

[0017] The assistance device 110 is a terminal used by a user such as a doctor or other medical professional. Although not shown, the assistance device 110 includes a computing device and a display device controlled by the computing device. The computing device is a device with calculation and communication functions, and may be, for example, a desktop personal computer or a portable communication terminal such as a smartphone or tablet. Input interfaces such as a keyboard, mouse, and touch panel are connected to or incorporated into the computing device, and the user uses these input interfaces to input user commands into the computing device. The display device is connected to or incorporated into the computing device and functions as an interface for the user. Details of the functions and operations of the assistance device 110 will be described later.

[0018] The database 160 may be a storage device connected to or installed in a communication computing terminal such as a personal computer, or may be a so-called file server specialized in storing and managing electronic data.

[0019] The heart rate monitor 170 and the acceleration sensor 180 are attached to the subject, and thereby it is possible to grasp the subject's state (for example, whether the subject is sleeping, at rest, or exercising, including walking or running). The heart rate monitor 170 and the acceleration sensor 180 can also be connected to the electrocardiogram analysis device 140 via the network 102. Furthermore, information from the heart rate monitor 170 and the acceleration sensor 180 can be linked to electrocardiogram data in the electrocardiogram analysis device 140, and may be stored in the electrocardiogram analysis device 140 and / or the database 160 together with the electrocardiogram data.

[0020] 2. Electrocardiogram Analysis Below, we will explain electrocardiogram analysis using the assistance system 100, along with the functions of the assistance device 110 and the electrocardiogram analysis device 140. A flowchart of the electrocardiogram analysis is shown in Figure 2. In this electrocardiogram analysis, after acquiring the electrocardiogram of the subject, two analyses (primary analysis and secondary analysis) are performed using a supervised learning model, and further, preprocessing for the secondary analysis is performed between the primary and secondary analyses. This will be described in detail below.

[0021] (1) Acquisition of an Electrocardiogram First, an electrocardiogram of the subject is acquired using the electrocardiograph 150. The measurement time for the electrocardiogram may be, for example, 10 seconds or more, 1 hour or more, 12 hours or more, 1 day or more, or 3 days or more. Performing measurement over a long period of time can increase the probability of detecting disease or abnormalities. There is no upper limit to the measurement time, and it may be, for example, 1 week, 3 days, 1 day, or 12 hours. The acquired electrocardiogram is stored in the assistance device 110 and / or database 160 from the electrocardiograph 150 via communication over the network 102 or via a storage medium. At the same time, information obtained from the heart rate monitor 170 and acceleration sensor 180 may also be stored in the assistance device 110, the electrocardiogram analysis device 140, and / or the database 160 via communication over the network 102 or via a storage medium.

[0022] (2) Primary Analysis The electrocardiogram is subjected to a primary analysis by the electrocardiogram analyzer 140. Specifically, first, the assistance device 110 uploads the electrocardiogram data stored in the assistance device 110 and / or the database 160 to the electrocardiogram analyzer 140 via the network 102 in accordance with a user command received via the input interface. Note that if the electrocardiogram data is stored in the electrocardiogram analyzer 140, this step does not need to be performed.

[0023] The electrocardiogram analysis device 140 divides the transmitted or stored electrocardiogram into multiple intervals. The duration of each interval can be set arbitrarily as long as it is shorter than the electrocardiogram acquisition time, and may be selected from a range of 5 to 120 seconds, for example, and is typically 30 seconds. As described above, since an electrocardiogram is acquired over a relatively long period of time, each interval corresponds to a portion of the entire electrocardiogram, and each electrocardiogram is composed of multiple intervals, as schematically shown in FIG. 3 . For example, if an electrocardiogram acquired over a 24-hour period is divided into 30-second intervals, one electrocardiogram will consist of 2,880 intervals.

[0024] Then, the electrocardiogram analyzer 140 performs an analysis (primary analysis) on each interval. In the primary analysis, the presence or absence of one or more diseases or test abnormalities is determined by utilizing a supervised learning model. Specifically, a large number of normal electrocardiograms, i.e., electrocardiograms of a large number of people who do not have the diseases or test abnormalities specified by the user as test items, are acquired, each electrocardiogram is divided into multiple intervals, and data for each interval is created as input data (features). Similarly, a large number of people who have the diseases or test abnormalities specified by the user as test items are acquired, each electrocardiogram is divided into multiple intervals, and data for each interval is created as input data. This input data is used to train a supervised learning model.

[0025] In the primary analysis, the electrocardiogram of the subject to be examined is used to perform analysis using a supervised learning model. The output data at this time is the judgment result for each interval, which is divided into two categories: whether or not each interval is an interval in which a disease or test abnormality selected by the user as a test item appears. As described above, there are no particular restrictions on the diseases or test abnormalities that are the detection targets; input data can be created for various diseases or test abnormalities in addition to cardiac diseases, and the supervised learning model can be trained accordingly. Alternatively, the diseases or test abnormalities selected as test items in the primary analysis can be cardiac diseases or test abnormalities.

[0026] Examples of supervised learning model algorithms include linear regression models, random forests, decision trees, gradient boosting, and deep learning, with deep learning being particularly preferred as it is expected to achieve high accuracy. When using deep learning, convolutional neural networks, recurrent neural networks, Vision Transformer models, and the like may be used as appropriate.

[0027] If the primary analysis detects a section in which a disease or test abnormality specified by the user appears, the user can recheck the electrocardiogram, determine whether the results of the primary analysis are correct, and take appropriate action based on that determination. However, primary analysis does not always accurately detect sections in which a disease or test abnormality appears. It may also erroneously detect normal sections as sections in which a disease or test abnormality appears. Furthermore, the primary analysis trains a learning model using electrocardiograms of individuals showing the disease or test abnormality specified by the user as test items and electrocardiograms of normal individuals without these diseases or test abnormalities as input data. However, depending on the type of disease or test abnormality, its symptoms may not be clearly visible on the electrocardiogram. For example, even if an electrocardiogram of a subject with a certain disease or test abnormality is acquired over a long period of time, the electrocardiogram may not include any sections showing waveforms due to the disease or test abnormality, and may instead consist of sections that are essentially recognizable as sinus waveforms. In such cases, the primary analysis may overlook the disease or test abnormality. Furthermore, there may be cases where it is necessary to diagnose the presence or absence of diseases or test abnormalities not listed in the test items in the primary analysis. For this reason, in an electrocardiogram analysis method according to one embodiment of the present invention, a secondary analysis (described later) is performed using the electrocardiogram data after the primary analysis.

[0028] (3) Preprocessing Unlike primary analysis, secondary analysis uses the following two types of data to create input data. One type is data for each interval obtained by dividing electrocardiogram data of a large number of individuals who do not have the disease or test abnormality that the user is testing. In contrast, the other type of electrocardiogram data is data for intervals in which the disease or test abnormality does not appear among the multiple intervals obtained by dividing electrocardiogram data of a large number of individuals who have the disease or test abnormality. For example, if the disease or test abnormality is paroxysmal atrial fibrillation, ventricular tachycardia, supraventricular tachycardia, or atrial flutter, the interval in which no waveform indicating the paroxysmal arrhythmia appears is used. In other words, secondary analysis uses, as input data, electrocardiogram data in which it is difficult for the user to determine the presence of a disease or test abnormality from the subject's electrocardiogram.

[0029] By using a supervised learning model trained using the input data, diseases or test abnormalities can be detected with high accuracy even when it is difficult for a user to determine the presence of a disease or test abnormality from a subject's electrocardiogram. For example, in the secondary analysis, a section indicating sinus rhythm, a section indicating premature contractions, a section indicating bundle branch block, a section indicating QRS width variation, a section indicating reduced heart rate variability, a section indicating T-wave amplitude variation, and the like are used within the analysis interval. Even if the analysis interval is a section indicating sinus rhythm, analysis using the supervised learning model can detect diseases or test abnormalities such as atrial fibrillation, heart failure, and ventricular arrhythmia. If the analysis interval includes a section indicating premature contractions, diseases or test abnormalities such as ventricular arrhythmia and valvular disease can be detected. On the other hand, if the analysis interval includes a section indicating bundle branch block, a section indicating QRS width variation, a section indicating reduced heart rate variability, or a section indicating T-wave amplitude variation, diseases or test abnormalities such as cardiomyopathy, hyponatremia, diabetes, and hyperkalemia can be detected.

[0030] Therefore, if a section in which a disease or test abnormality appears is used in the secondary analysis, it becomes difficult to accurately detect the disease or test abnormality. Furthermore, there are sections in the electrocardiogram data that contain waveforms that are inappropriate for the secondary analysis, such as sections that contain artifacts. Therefore, in an electrocardiogram analysis method according to one embodiment of the present invention, prior to the secondary analysis, preprocessing is performed to identify and exclude sections that are inappropriate for the secondary analysis (specific sections) from the multiple sections used in the primary analysis, and to extract the remaining sections as candidate sections for the secondary analysis.

[0031] Specifically, the electrocardiogram analysis device 140 identifies sections in which a disease or test abnormality appears in the primary analysis as specific sections. Furthermore, the electrocardiogram analysis device 140 may be configured to determine whether each section contains noise unrelated to the disease or test abnormality and identify sections containing noise as specific sections. Examples of noise include electrostatic induction, electromagnetic induction, AC noise due to leakage current, electromyograms unrelated to the subject's disease, baseline drift, and noise due to static electricity. There are no restrictions on the method for determining the presence or absence of noise; known methods and algorithms may be applied. For example, a supervised learning model trained on input data from electrocardiograms containing and not containing noise may be used to determine whether each section comprising the subject's electrocardiogram contains noise. Sections identified as specific sections are excluded from the multiple sections, and the remaining sections are extracted as candidate sections for user judgment on the assistance device 110, described below. For this reason, each section is associated with another section in preprocessing. For example, each section of the subject's electrocardiogram is given an identifier in chronological order to identify it, and each identifier is assigned a flag to identify whether it has been extracted as a candidate section and / or whether it has not been extracted (i.e., whether it has been identified as a specific section).

[0032] Next, sections containing waveforms unsuitable for secondary analysis are excluded. However, this process requires the expertise of the user, who is a specialist. For this reason, the assistance system 100 provides an opportunity to exclude sections unsuitable for secondary analysis from the extracted candidate sections based on the user's expert knowledge. Specifically, the electrocardiogram analysis device 140 displays multiple candidate sections on the screen of the display device of the assistance device 110 in accordance with instructions transmitted from the assistance device 110. The user selects, from the displayed candidate sections, candidate sections that are unsuitable for detecting diseases or test abnormalities in secondary analysis, such as candidate sections containing artifacts, and excludes these candidate sections as non-analysis sections from the candidate sections. In this way, candidate sections suitable for secondary analysis are selected as analysis sections.

[0033] The display method can be set arbitrarily. FIG. 4 shows an example of a display on the screen of the display device of the assistance device 110. In the example shown in FIG. 4, the first 10 candidate intervals among the multiple intervals constituting the electrocardiogram are displayed in chronological order in the upper region 112. The number of candidate intervals (n) displayed here is not limited and can be changed as appropriate depending on the size of the display device, etc. Alternatively, the program may be configured so that the user can set the number of candidate intervals to be displayed. For example, one to twenty candidate intervals, typically ten, may be displayed on one screen. The candidate intervals may be displayed in a single row, or may be displayed across multiple rows as shown in FIG. 4. Preferably, the candidate intervals are displayed across multiple rows and columns. Displaying the candidate intervals in this manner allows easy comparison of the candidate intervals, thereby more efficiently identifying intervals unsuitable for secondary analysis.

[0034] As shown in FIG. 4 , an identifier or a corresponding number assigned to each section may be displayed on or near each section. Furthermore, multiple candidate sections (here, the 11th to 13th sections) following these multiple candidate sections may be displayed in chronological order in a lower area 114 on the screen. An area 116 indicating the subject's attributes (such as name, age, gender, electrocardiogram acquisition date and time, and subject identification information) may also be provided. Furthermore, icons 118 for receiving input to display subsequent or preceding candidate sections not displayed on the screen may be provided on the screen. When candidate sections are displayed across multiple pages, an area 120 for receiving input to directly move to any page may also be provided. In the example shown in FIG. 4 , all candidate sections from the 1st to the 13th section are displayed, but as described above, specific sections are not selected as candidate sections. Therefore, not all consecutive sections are necessarily displayed on a single screen, and the numbers of adjacent candidate sections may not be consecutive, as shown in FIG. 5 . As will be described in detail later, an icon 122 is provided on the screen for receiving a user command to cause the electrocardiogram analyzer 140 to detect a disease or test abnormality.

[0035] FIG. 6 shows an enlarged view of one candidate section. As shown in FIG. 6 , an icon 124 for receiving a user command to enlarge a candidate section or an icon 126 for receiving a user command to exclude a candidate section as a non-analysis section can be placed on or near each candidate section. When the user inputs a user command via the icon 124, the assistance device 110 enlarges the candidate section on the display screen. The enlargement method can also be set arbitrarily. For example, as shown in FIG. 7 , the candidate section to be enlarged may be enlarged across one or more lines so that it overlaps with other candidate sections that are not enlarged. Alternatively, as shown in FIG. 8 , the user may select an enlargement target region 128 on the screen to enlarge that region. By enlarging each candidate section in this manner, the user can observe each candidate section in more detail and determine whether it is an analysis section suitable for secondary analysis by the electrocardiogram analyzer 140.

[0036] If the user determines that a candidate section is not suitable for secondary analysis by the electrocardiogram analyzer 140, the user selects it as a non-analysis section and uses the icon 126 to exclude it from the candidate sections (see FIG. 6 ). Upon receiving input via the icon 126, the assistance device 110, in accordance with program instructions, erases the candidate section from the screen and displays the subsequent candidate sections. For example, if sections 7 to 9 of the candidate sections shown in FIG. 4 are excluded, these candidate sections are erased from the screen, and the subsequent sections (sections 11 to 13) are displayed in chronological order in the upper area 112 (see FIG. 9 ). Furthermore, the subsequent sections (sections 14 to 16) are displayed in chronological order in the lower area 114. This allows the user to observe the candidate sections in a short time without feeling stressed. At the same time, the assistance device 110 transmits the identifiers of the sections selected as non-analysis sections to the electrocardiogram analyzer 140. The electrocardiogram analyzer 140 sets a flag for the identifier of the section selected as the non-analysis section to indicate that it has been excluded from the candidate sections by the user. On the other hand, the candidate sections not selected by the user remain as analysis sections. Through the above process, the candidate sections to be used in the secondary analysis are selected.

[0037] Typically, the number of candidate intervals selected as non-analysis intervals is overwhelmingly small compared to the number of candidate intervals remaining as analysis intervals. Therefore, the burden on the user of selecting a non-analysis interval from the candidate intervals is extremely small compared to the operation of selecting an analysis interval from the candidate intervals displayed on the screen. Therefore, by having the user select a non-analysis interval from the candidate intervals, analysis intervals can be quickly selected and collected without imposing a heavy burden on the user.

[0038] As an optional display method, the support device 110 may be configured to display all sections on the screen of the display device in accordance with program instructions. The display method in this case can also be set arbitrarily. For example, as shown in FIG. 10 , a tab different from the tab displaying the candidate sections may be provided, and all sections may be displayed on this tab. In this case, a display for distinguishing between sections extracted as candidate sections by the primary analysis and specific excluded sections may be placed on or near each section. This provides the user with an opportunity to verify the results of the primary analysis, allowing the user to determine whether sections suitable for secondary analysis were mistakenly excluded in the primary analysis.

[0039] Furthermore, as an optional display method, the assistance device 110 may be configured to display multiple candidate sections on the screen in chronological order from candidate sections arbitrarily selected by the user. For example, the user may select the candidate section to be displayed first in consideration of the subject's condition. For example, the user may select the candidate section to be displayed first based on information acquired by the heart rate monitor 170 or the acceleration sensor 180. Specifically, one of the candidate sections acquired while the subject is sleeping, resting, or exercising may be displayed first, and subsequent candidate sections may be displayed in chronological order from this candidate section.

[0040] (4) Secondary Analysis Secondary analysis is then performed using the analysis interval. That is, after selecting the non-analysis interval, the user operates the icon 122. Upon receiving a user command via the icon 122, the assistance device 110 transmits an instruction to the electrocardiogram analyzer 140 to perform disease or test abnormality detection using the analysis interval. The electrocardiogram analyzer 140 performs disease or test abnormality detection in accordance with this instruction. The various machine learning algorithms described for the primary analysis can also be used in secondary analysis. The output data is a judgment result for each interval, which is divided into two categories: whether or not each interval is an interval in which a disease or test abnormality selected by the user as a test item has occurred. As with the primary analysis, the secondary analysis may be performed for each analysis interval, for all analysis intervals simultaneously, or for multiple analysis intervals (e.g., 10 analysis intervals).

[0041] (5) Display of Analysis Results The results of the secondary analysis are displayed on the screen of the display device of the assistance device 110. The display method at this time may be set arbitrarily. For example, as shown in FIG. 11 , in addition to an area 130 indicating the subject's attributes, the entire or part of the electrocardiogram may be displayed. For example, a section showing a waveform suggesting a disease or test abnormality may be displayed. In addition, various information to assist the user's medical treatment, such as the date and time of electrocardiogram measurement, the date and time of electrocardiogram analysis, the probability of disease or test abnormality onset, degree, or score, may be displayed. In addition, past electrocardiogram analysis results of the subject may also be displayed.

[0042] In the electrocardiogram analysis described above, the user operates various icons to input user commands, but there are no restrictions on the method of inputting the user commands, and the user may select a user command from a pull-down menu, or may input the user command by operating a keyboard or by voice. Furthermore, the primary analysis and secondary analysis described above may be performed using the same electrocardiogram analyzer, or may be performed using different electrocardiogram analyzers.

[0043] As described above, in electrocardiogram analysis using the support system 100, the electrocardiogram analysis device 140 divides the electrocardiogram into multiple intervals, and in preprocessing after the primary analysis, some of the multiple intervals are extracted as candidate intervals suitable for secondary analysis. Furthermore, the candidate intervals are provided for verification by the user, allowing the electrocardiogram analysis device 140 to select only candidate intervals suitable for detecting various diseases and test abnormalities as analysis intervals. Therefore, in the secondary analysis performed by the electrocardiogram analysis device 140, electrocardiogram analysis can be performed using analysis intervals that have less noise and are suitable for detecting diseases and test abnormalities. As a result, various diseases, their symptoms, and test abnormalities can be detected more accurately.

[0044] 3. Program One embodiment of the present invention is a program for assisting electrocardiogram analysis performed by the electrocardiogram analyzer 140 using the assistance device 110. Another embodiment of the present invention is a computer-readable storage medium on which this program is recorded. This program may be installed in the assistance device 110, or may be a program configured to be run over a network by a user using a browser installed in the assistance device 110.

[0045] The program is configured to cause the computing device of the assistance device 110 to execute the instructions for the electrocardiogram analysis described above. Accordingly, the program is configured to accept a user command to upload the electrocardiogram of the subject to the electrocardiogram analysis device 140. Upon accepting this user command, the program causes the computing device to upload the electrocardiogram stored in the assistance device 110 or the database 160 to the electrocardiogram analysis device 140 via the network 102.

[0046] The program further causes the computing device to display all sections of the electrocardiogram or multiple candidate sections extracted by the electrocardiogram analysis device 140 on the screen of the display device of the assistance device 110. The program may be configured to transmit all sections of the electrocardiogram or the candidate sections to the assistance device 110. The program is configured to cause the assistance device 110 to accept user commands such as a user command to enlarge and display the candidate sections, a user command to select the candidate sections as non-analysis sections, and a user command to cause the electrocardiogram analysis device 140 to detect diseases or test abnormalities using the analysis sections. Upon receiving these user commands, the program is configured to cause the computing device of the assistance device 110 to perform corresponding operations, such as enlarging and displaying the candidate sections, sending identifiers of the candidate sections selected as non-analysis sections to the electrocardiogram analysis device 140 and erasing the candidate sections selected as non-analysis sections from the screen, and sending instructions to the electrocardiogram analysis device 140 to analyze the electrocardiogram using the analysis sections to detect diseases or test abnormalities. The program may further be configured to receive a user command to cause the computing device of the assistance device 110 to display all of the multiple sections of the electrocardiogram on the screen, and to cause the computing device of the assistance device 110 to display all of the sections on the screen. At this time, the program may cause the computing device of the assistance device 110 to execute a display for identifying sections that were not extracted as candidate sections by the electrocardiogram analysis device 140.

[0047] Examples of the computer-readable storage medium include magnetic media such as hard disks, flexible disks, and magnetic tapes, optical media such as CD-ROMs and DVDs, magneto-optical media such as floptical disks, and hardware devices configured to store or execute the program, such as ROMs, RAMs, flash memories, etc. The program instructions include not only machine code, such as that generated by a compiler, but also high-level language code executed by the server using an interpreter, etc. The program may be configured to be installable into the computing device of the support apparatus 110 from a computer-readable storage medium. Alternatively, the program may be downloadable to the computing device via the network 102.

[0048] The above-described embodiments of the present invention can be combined as appropriate as long as they are not mutually inconsistent. Furthermore, even if a person skilled in the art appropriately adds or deletes components or modifies designs, or adds or omits processes or modifies conditions based on the embodiments, such combinations are included within the scope of the present invention as long as they include the gist of the present invention.

[0049] Even if there are other effects and advantages different from those brought about by the aspects of each of the above-mentioned embodiments, those that are clear from the description in this specification or that can be easily predicted by a person skilled in the art are naturally understood to be brought about by the present invention.

[0050] 100: Support system, 102: Network, 110: Support device, 112: Upper region, 114: Lower region, 116: Region, 118: Icon, 120: Region, 122: Icon, 124: Icon, 126: Icon, 128: Region to be enlarged, 130: Region, 140: Electrocardiogram analyzer, 150: Electrocardiograph, 160: Database, 170: Heart rate monitor, 180: Acceleration sensor

Claims

1. An assistance system for electrocardiogram analysis comprising: an electrocardiograph configured to acquire an electrocardiogram of a subject; an electrocardiogram analysis device configured to receive the electrocardiogram, divide the electrocardiogram into a plurality of intervals, and extract, from the plurality of intervals, intervals other than specific intervals as candidate intervals; and an assistance device configured to communicate with the electrocardiogram analysis device, wherein the assistance device comprises a computing device and a display device controlled by the computing device, the computing device being configured to: accept a first user command to upload the electrocardiogram to the electrocardiogram analysis device; display the candidate intervals on the screen of the display device; accept input of a second user command to select a non-analysis interval from the candidate intervals; and send an instruction to the electrocardiogram analysis device to analyze the electrocardiogram using analysis intervals that are intervals from the candidate intervals not selected by the second user command, wherein the specific intervals include intervals that show waveforms caused by a disease.

2. The support system of claim 1, wherein the electrocardiogram analysis device is configured to analyze the multiple sections using a supervised learning model created by machine learning using electrocardiogram data in which the disease is present and electrocardiogram data in which the disease is not present as training data.

3. The support system of claim 1, wherein the electrocardiogram analysis device is configured to analyze the analysis interval using a supervised learning model created by machine learning using as training data electrocardiogram data of people who have developed the disease, but in which the disease is not present, and electrocardiogram data of people who have not developed the disease.

4. The assistance system according to claim 3, wherein the electrocardiogram analyzer is further configured to determine, based on the analysis, whether the electrocardiogram shows signs of the disease or abnormalities that suggest the disease.

5. The assistance system of claim 1, wherein the computing device is configured to display on the screen the candidate sections acquired from an arbitrarily selected point in time onward and the subsequent candidate sections in chronological order.

6. The assistance system of claim 1, wherein the computing device is further configured to display all of the plurality of sections on the screen.

7. The assistance system according to claim 6, wherein the computing device is further configured to display on the screen a display for identifying the specific section.

8. The assistance system of claim 1, wherein the computing device is further configured to: display the n candidate sections on the screen in chronological order; delete the candidate section selected as the non-analyzed section from the screen in accordance with the second user command; and display one of the candidate sections subsequent to the n candidate sections on the screen, wherein n is selected from an integer greater than or equal to 1 and less than or equal to 20.

9. An electrocardiogram analysis support device comprising: a computing device; and a display device controlled by the computing device, wherein the computing device is configured to receive a first user command to cause an electrocardiogram analysis device, communicatively connected to the computing device and configured to divide an electrocardiogram into a plurality of intervals and extract, as candidate intervals, intervals other than specific intervals, to upload the electrocardiogram of the subject to the computing device; display the candidate intervals on a screen of the display device; receive input of a second user command to select non-analysis intervals from the candidate intervals; and send an instruction to the electrocardiogram analysis device to analyze the electrocardiogram of the subject using analysis intervals, which are intervals among the candidate intervals not selected by the second user command, wherein the specific intervals include intervals showing waveforms caused by a disease.

10. The assistance device of claim 9, wherein the computing device is configured to display on the screen the candidate sections acquired after an arbitrarily selected point in time and the subsequent candidate sections in chronological order.

11. The assistance apparatus of claim 9, wherein the computing device is further configured to display all of the plurality of segments on the screen.

12. The assistance apparatus according to claim 11, wherein the computing device is further configured to display a message for identifying the specific section on the screen.

13. The assistance device described in claim 9, wherein the computing device is further configured to: display n candidate sections on the screen in chronological order; delete from the screen the candidate section selected as the non-analyzed section in accordance with the second user command; and display one of the candidate sections subsequent to the n candidate sections on the screen, wherein n is selected from integers greater than or equal to 1 and less than or equal to 20.

14. A program for using a computing device to assist electrocardiogram analysis performed in an electrocardiogram analyzer, the electrocardiogram analyzer being communicatively connected to the computing device and configured to divide a subject's electrocardiogram into a plurality of intervals and extract the intervals other than a specific interval from the plurality of intervals as candidate intervals, the program being configured to cause the computing device to: accept a first user command to upload the subject's electrocardiogram to the electrocardiogram analyzer; display the candidate intervals on a screen of a display device of the computing device; accept input of a second user command to select a non-analysis interval from the candidate intervals; and send an instruction to the electrocardiogram analyzer to analyze the electrocardiogram using an analysis interval that is an interval from the candidate intervals not selected by the second user command.

15. The program of claim 14, further configured to cause the computing device to display on the screen the candidate sections acquired after an arbitrarily selected point in time and subsequent candidate sections in chronological order.

16. The program of claim 14, further configured to cause the computing device to display all of the plurality of intervals on the screen.

17. The program of claim 16, further configured to cause the computing device to display a display for identifying the specific section on the screen.

18. The program of claim 16, further configured to cause the computing device to: display n candidate sections on the screen in chronological order; delete from the screen the candidate section selected as the non-analyzed section in accordance with the second user command; and display on the screen one of the candidate sections subsequent to the n candidate sections, wherein n is selected from an integer greater than or equal to 1 and less than or equal to 20.

Images