Estimation device and estimation method

Abstract

To accurately estimate whether a patient has a disease accompanied by Cheyne-Stokes respiration.SOLUTION: An estimation device for estimating whether a patient has a disease accompanied by Cheyne-Stokes respiration includes: a detection unit that detects a plurality of candidate events occurring during a measurement period of time-series data on the basis of variations in the time-series data of pulse intervals of a patient; an identification unit that identifies at least one of the plurality of candidate events as a base event by performing a plurality of determination processes; and an estimation unit that estimates whether the patient has a disease accompanied by Cheyne-Stokes respiration on the basis of the base event. The plurality of determination processes include a first type of determination process for determining whether a first candidate event among the plurality of candidate events satisfies a first condition that is not based on a determination result of a second candidate event among the plurality of candidate events, and a second type of determination process for determining whether the first candidate event satisfies a second condition based on a determination result of the second candidate event.SELECTED DRAWING: Figure 3

Description

[Technical Field]

[0001] The present invention relates to an estimation device and an estimation method. [Background technology]

[0002] Some heart failure patients with sleep-disordered breathing may have Cheyne-Stokes respiration, a central abnormal breathing disorder characterized by periodic repetition of hyperpnea and hypopnea or apnea. Patent Document 1 describes a technique for detecting the occurrence of an apnea state based on the patient's beat interval data. Non-Patent Document 1 describes a technique for estimating whether a patient has Cheyne-Stokes respiration based on the patient's beat interval data. [Prior art documents] [Patent documents]

[0003] [Patent Document 1] Japanese Patent Publication No. 2020-168214 [Non-patent literature]

[0004] [Non-Patent Document 1] Yoshioka E, et al., "Discrimination between primary-central and primary-obstructive respiratory events by a pattern of cyclic variation of heart rate", 42nd Annual International Conferences of the IEEE Engineering in Medicine and Biology Society, July 20, 2020 Summary of the Invention [Problem to be solved by the invention]

[0005] The technology in Non-Patent Document 1 estimates a patient's disease based only on individual fluctuations in heart rate interval data. While this estimation method can estimate apnea, it is considered difficult to estimate the type of breathing, such as central or obstructive. One aspect of the present invention aims to provide a technology for accurately estimating whether a patient has a disease accompanied by Cheyne-Stokes respiration. [Means for solving the problem]

[0006] According to some embodiments, there is provided an estimation device for estimating whether a patient has a disease accompanied by Cheyne-Stokes respiration, the estimation device comprising: a detection means for detecting a plurality of candidate events occurring during a measurement period of time series data based on fluctuations in the time series data of beat intervals of the patient; an identification means for identifying at least one of the plurality of candidate events as an underlying event by performing a plurality of determination processes; and an estimation means for estimating whether the patient has a disease accompanied by Cheyne-Stokes respiration based on the underlying event, wherein the plurality of determination processes include a first type determination process for determining whether a first candidate event of the plurality of candidate events satisfies a first condition that is not based on a determination result of a second candidate event of the plurality of candidate events, and a second type determination process for determining whether the first candidate event satisfies a second condition that is based on a determination result of the second candidate event. [Effects of the Invention]

[0007] According to some embodiments, it is possible to accurately estimate whether a patient has a disorder associated with Cheyne-Stokes breathing. [Brief explanation of the drawings]

[0008] [Figure 1] FIG. 2 is a block diagram illustrating an example of a hardware configuration according to some embodiments. [Figure 2] FIG. 1 is a schematic diagram illustrating the principle of disease prediction in some embodiments. [Figure 3] FIG. 10 is a flow diagram illustrating an example of the operation of disease estimation according to some embodiments. [Figure 4] 1 is a schematic diagram illustrating examples of candidate events according to some embodiments; [Figure 5] Schematic diagram illustrating an example of an independent positive condition of some embodiments. [Figure 6] Schematic diagram illustrating an example of a positive linkage condition according to some embodiments. [Figure 7] Schematic diagram illustrating an example of a continuity condition according to some embodiments. DETAILED DESCRIPTION OF THE INVENTION

[0009] Hereinafter, the embodiments will be described in detail with reference to the accompanying drawings. Note that the following embodiments do not limit the scope of the invention as claimed, and not all combinations of features described in the embodiments are necessarily essential to the invention. Two or more of the features described in the embodiments may be combined in any desired manner. Furthermore, the same reference numerals are used to designate identical or similar components, and redundant descriptions will be omitted.

[0010] Referring to FIG. 1 , an example hardware configuration of a computer 100 according to some embodiments will be described. As described in detail below, the computer 100 is used to estimate whether a patient has a disease accompanied by Cheyne-Stokes respiration (CSR). Therefore, the computer 100 may be referred to as an estimation device or a disease estimation device. Hereinafter, a disease accompanied by CSR will be referred to as a CSR disease. The computer 100 may be, for example, a server computer or a personal computer (e.g., a desktop or laptop computer). The computer 100 may be a computer resource located in a cloud environment. The computer 100 may also be a dedicated computer for performing medical-related processing, including estimating whether a patient has CSR disease.

[0011] The computer 100 may include the hardware devices shown in Fig. 1. The processor 101 controls the overall operation of the computer 100. The processor 101 may be configured, for example, by a central processing unit (CPU), a graphics processing unit (GPU), or a combination thereof. The processor 101 may be a single processor or a collection of multiple processors communicatively connected to each other.

[0012] The memory 102 stores programs and data used in the processing of the computer 100. The memory 102 may be configured, for example, by a combination of random access memory (RAM) and read-only memory (ROM).

[0013] The input device 103 is a device for obtaining instructions from a user of the computer 100. The input device 103 may be configured, for example, by a combination of one or more of a keyboard, buttons, a touchpad, and a microphone. The display device 104 is a device for visually presenting information to a user of the computer 100. The display device 104 may be, for example, a dot-matrix display such as a liquid crystal display. The computer 100 may have a device (e.g., a touch screen) in which the input device 103 and the display device 104 are integrated. The input device 103 and the display device 104 may be external to the computer. In this case, the computer 100 may have an interface for communicating with the external input device 103 and display device 104.

[0014] The communication device 105 is a device for communicating with devices external to the computer 100. When the computer 100 performs wired communication, the communication device 105 may be a network interface card (NIC) having a connector for connecting a cable. When the computer 100 performs wireless communication, the communication device 105 may be a wireless communication module including an antenna and a baseband processing circuit.

[0015] The secondary storage device 106 is a device for non-volatilely storing programs and data used in the processing of the computer 100. The secondary storage device 106 is configured by, for example, a hard disk drive (HDD) or a solid state drive (SSD).

[0016] The computer 100 may be capable of communicating with the measuring device 110. The measuring device 110 may be a device capable of measuring a patient's heartbeat, such as a heartbeat or pulse. For example, the measuring device 110 may be a device for measuring the interval between heartbeats of the patient (heartbeat data of an electrocardiogram) or the interval between the heart pumping blood throughout the body (pulse wave pulse data). In some embodiments, the measuring device 110 may be an electrocardiograph for measuring data representing the electrical activity of the patient's heart (i.e., electrocardiogram data). The electrocardiograph may be a Holter-type electrocardiograph or another type of electrocardiograph, such as a stationary electrocardiograph. The computer 100 may be a device separate from the measuring device 110. Alternatively, the computer 100 and the measuring device 110 may be integrated into one device. The measuring device 110 may include an SpO2 sensor capable of detecting pulse wave data.

[0017] The measuring device 110 stores the measured data in its own memory 111. The data measured by the measuring device 110 is referred to as measurement data 112. The measurement data 112 may be heart rate data or pulse data. When the measuring device 110 is an electrocardiograph, the measurement data 112 may be called electrocardiogram data. The computer 100 can read the measurement data 112 from the measuring device 110 using the communication device 105. The computer 100 may obtain the measurement data 112 directly from the measuring device 110, or may obtain the measurement data 112 via another device. For example, the computer 100 may obtain the measurement data 112 stored in an external server.

[0018] The principle of the estimation operation performed by the computer 100 will be described with reference to FIG. 2. Graph 200 in FIG. 2(a) shows time series data of the respiratory volume during sleep of a patient suffering from central sleep apnea syndrome. Central sleep apnea syndrome is an example of a disease in which CSR occurs during sleep. The horizontal axis of graph 200 represents time, and the vertical axis of graph 200 represents the pressure within the patient's mouth and nose. In graph 200, as indicated by arrow 201, the respiratory volume that had stopped gradually increases. Thereafter, the respiratory volume gradually decreases, and breathing stops again. Respiration with this waveform is called CSR.

[0019] Graph 210 in FIG. 2(b) shows time series data of the heart rate interval (RRI) during sleep of a patient suffering from central sleep apnea syndrome. The horizontal axis of graph 210 represents time, and the vertical axis of graph 210 represents the RRI. During period 211 of graph 210, as the patient's respiratory volume gradually increases, the RRI gradually decreases (i.e., the heart rate gradually increases). Thereafter, the RRI increases (i.e., the heart rate decreases).

[0020] Graph 220 in Figure 2(c) shows time-series data on the respiratory rate of a patient suffering from obstructive sleep apnea syndrome (OSA). The horizontal axis of graph 220 represents time, and the vertical axis of graph 220 represents the pressure within the patient's mouth and nose. In graph 220, as indicated by arrow 221, the respiratory rate, which had stopped, suddenly increases. Thereafter, the respiratory rate gradually decreases, and breathing stops again.

[0021] Graph 230 in FIG. 2(d) shows time-series data of RRI during sleep of a patient suffering from obstructive sleep apnea syndrome. The horizontal axis of graph 230 represents time, and the vertical axis of graph 230 represents RRI. During period 231 of graph 230, as the patient's respiratory volume increases abruptly, the RRI decreases abruptly (i.e., the heart rate increases abruptly). Thereafter, the RRI increases (i.e., the heart rate decreases).

[0022] As described above, the waveform of the patient's RRI (particularly, the manner in which the RRI decreases) differs depending on whether the patient has a CSR disease (e.g., central sleep apnea syndrome). Therefore, in some embodiments, the computer 100 estimates whether the patient has a CSR disease based on the time-series data of the patient's RRI.

[0023] An example of a method for predicting whether a patient has CSR disease will be described with reference to FIG. 3 . Each step of the method of FIG. 3 may be performed by the processor 101 executing a program loaded into the memory 102. Alternatively, some or all of the steps of the method of FIG. 3 may be performed by a dedicated integrated circuit such as an ASIC. The method of FIG. 3 may be initiated, for example, in response to an instruction from a user of the computer 100. The user of the computer 100 may be the patient himself / herself who is the subject of the prediction. Alternatively, the user of the computer 100 may be a person other than the patient (e.g., a doctor). For example, a doctor may cause the computer 100 to execute the method of FIG. 3 to help determine the patient's treatment plan. Various thresholds used in the following methods may be preset and stored in the computer 100 (e.g., the secondary storage device 106) as part of the program. These thresholds may be updateable by a user (e.g., a patient or a doctor).

[0024] In S301, the computer 100 acquires time series data of the patient's beat intervals. The beat intervals may be heart beat intervals or pulse intervals. In the following description, heart beat intervals are used as the beat intervals. The same description applies when pulse intervals are used as the beat intervals. The time series data of the heart beat intervals is referred to as RRI data. The RRI data may be acquired while the patient is in bed, and particularly while the patient is sleeping. The measurement period of the RRI data may correspond to the patient's bedtime period or the patient's sleeping period.

[0025] The computer 100 may generate the RRI data based on the measurement data 112 read from the measurement device 110. Alternatively, the RRI data may be generated by the measurement device 110, and the computer 100 may read the RRI data generated by the measurement device 110.

[0026] In S302, the computer 100 detects multiple candidate events that occurred during the measurement period of the RRI data based on fluctuations in the patient's RRI data. A candidate event may be an event that may occur in the RRI data due to CSR. In some embodiments, a candidate event may be associated with any minimum point in a waveform based on the RRI data (e.g., a waveform representing a low-frequency component). For example, a candidate event may be a change in the low-frequency component of the RRI data from a decrease to an increase. Such a change may also be called a dip.

[0027] An example of processing for detecting candidate events will be described with reference to FIG. 4. Graph 400 in FIG. 4 represents a portion of the RRI data acquired in S301. The horizontal axis of graph 410 represents time, and the vertical axis of graph 410 represents RRI. The computer 100 removes high-frequency components from the RRI data to make it easier to recognize feature points. Graph 410 in FIG. 4 represents a portion of the RRI data from which the high-frequency components have been removed. The processing using the low-pass filter may be performed, for example, by averaging the envelopes of the maximum and minimum values ​​of graph 400.

[0028] The computer 100 may detect a change in the RRI from a decrease to an increase in the graph 410 as a candidate event. Furthermore, the computer 100 may identify a period associated with the candidate event as the candidate period. For example, the computer 100 may identify at least a portion of the period between two adjacent maximum points as the candidate period. The graph 410 reaches a maximum value y4 at time t11, a minimum value y1 at a subsequent time t13, and a maximum value y3 at a subsequent time t15. Therefore, the computer 100 detects a candidate event from time t11 to t15. The computer 100 may calculate y2 (=0.8×(y3+y4) / 2) by multiplying the average value of the two maximum values ​​by a predetermined coefficient (e.g., 0.8), and identify the period in which the value of the graph 410 is equal to or less than y2 (i.e., from time t12 to t14) as the candidate period 411. Similarly, the computer 100 may identify a candidate period 412 in relation to a candidate event that occurred after the candidate period 411, and may identify a candidate period 413 in relation to a candidate event that occurred after that. Instead of this method of identifying a candidate period, the computer 100 may identify a period between two local maximum points (for example, times t11 to t15) as a candidate period.

[0029] The computer 100 may detect only events that satisfy certain conditions as candidate events. For example, the computer 100 may detect an event that occurred during at least a portion of a period between two local maxima (e.g., the period from time t12 to t14 as described above) as a candidate event if the width (i.e., t14-t12) of this period is greater than a threshold width and the depth (i.e., y2-y1) of this period is greater than a threshold depth. Alternatively or additionally, the conditions for detecting an event as a candidate event may include a condition that low-frequency components of the RRI data are dominant during at least a portion of the period between the two local maxima. By determining such a condition, a decrease (dip) in the RRI data that is thought to have occurred without being caused by the patient's respiratory arrest can be excluded from the candidate events.

[0030] Returning to the explanation of FIG. 3, in S303, computer 100 selects one unprocessed candidate event. "Unprocessed" means that the processes of S304 to S309, which will be described below, have not been performed. The candidate event selected in S303 is referred to as a selected event. Computer 100 may select the multiple candidate events detected in S302 in order of occurrence, starting with the event that occurred first.

[0031] In S304, the computer 100 determines whether the selected event satisfies the independent positive condition. If the selected event is determined to satisfy the independent positive condition ("YES" in S304), the computer 100 transitions the process to S306, and otherwise ("NO" in S304), the computer 100 transitions the process to S305.

[0032] The independent positive condition may be a condition that is not based on the determination result of a candidate event other than the selected event. For example, the independent positive condition may be a condition that is based only on the selected event. If the selected event is determined to satisfy the independent positive condition, the computer 100 identifies the selected event as a positive event. A positive event may be a candidate event that suggests that the patient has CSR disease (i.e., is positive).

[0033] For example, in S304, the computer 100 may determine whether the RRI decreases in the selected event so as to satisfy a specific condition. The specific condition may include the time from the start of the decrease in the RRI to the end of the decrease being longer than a threshold time. Alternatively or in addition, the specific condition may include the absolute value of the rate of change in the RRI data from the start of the decrease in the RRI to the end of the decrease (i.e., the slope of the graph) being greater than a threshold rate. If the computer 100 determines that the RRI decreases in the selected event so as to satisfy a specific condition (e.g., the time from the start of the decrease to the end of the decrease is longer than the threshold time), the computer 100 may transition the process to S306, and otherwise transition the process to S305.

[0034] A specific example of the determination process in S304 will be described with reference to FIG. 5. Graph 500 in FIG. 5 represents a portion of RRI data from which high-frequency components have been removed. A candidate event is detected during a candidate period 501 from time t21 to t24. Graph 500 reaches a minimum value at time t23. During the period in which graph 500 decreases (i.e., at time t21), the slope reaches its minimum (i.e., the absolute value of the slope reaches its maximum) at time t22. In this case, computer 100 identifies the period from time t22 to time t23 as a start-peak period (SP period) 502. As described with reference to FIG. 2, the SP period 502 can be lengthened if the patient has CSR. Therefore, computer 100 may identify the selected event as a positive event if the length of SP period 502 is longer than a threshold time (e.g., 16 seconds).

[0035] 3, if the selected event is determined not to be a positive event in S304, the computer 100 executes the process of S305. In S305, the computer 100 determines whether the selected event satisfies the positive linkage condition. If the selected event is determined to satisfy the positive linkage condition ("YES" in S305), the computer 100 transitions the process to S306, and otherwise ("NO" in S305), the computer 100 transitions the process to S307.

[0036] The linked positive condition may be a condition used to determine whether a selected event is a positive event based on the determination results of a candidate event other than the selected event. For example, even if the length of the SP period related to the selected event is shorter than the threshold time (e.g., 16 seconds) in S304, if a candidate event near the selected event is a positive event, it is highly likely that the patient also performed CSR in the selected event. Therefore, even if the selected event was not identified as a positive event in S304, the computer 100 identifies the selected event as a positive event if the linked positive condition is satisfied in S305.

[0037] The positive linkage condition may be, for example, that all of the following conditions are satisfied: (1) a candidate event immediately preceding the selected event (hereinafter, the immediately preceding event) is identified as a positive event; (2) the similarity between the immediately preceding event and the selected event is greater than a threshold similarity; and (3) the RRI in the selected event decreases to satisfy a specific condition (e.g., the time from the start of the decrease to the end of the decrease is longer than a threshold time (e.g., 10 seconds)). The threshold time used in condition (3) is shorter than the threshold time used in S304. Alternatively, the positive linkage condition may be that (1) and (2) are satisfied, or that (3) is not satisfied. In this example, the positive linkage condition is based on the determination result of the immediately preceding event. The positive linkage condition may be based on the determination result of another candidate event instead of or in addition to the immediately preceding event. For example, the positive linkage condition may be based on a candidate event immediately following the selected event, or on an average value of RRI data for multiple candidate events immediately preceding or following the selected event.

[0038] An example of a method for determining the similarity between a previous event and a selected event will be described with reference to FIG. 6. Graph 600 in FIG. 6 shows a portion of RRI data from which high-frequency components have been removed. A candidate event detected in candidate period 603 from time t35 to t38 is set as the selected event. Therefore, a candidate event detected in candidate period 601 from time t31 to t34 is set as the previous event. Because the length of SP period 602 (i.e., from time t32 to t33) of the previous event is longer than the threshold time (e.g., 16 seconds), the previous event is identified as a positive event in S304. On the other hand, because the length of SP period 604 (i.e., from time t36 to t37) of the selected event is shorter than the threshold time (e.g., 16 seconds), the previous event is not identified as a positive event in S304.

[0039] Therefore, in S305, the computer 100 calculates the similarity between the immediately preceding event and the selected event. For example, the computer 100 may calculate the similarity based on a correlation coefficient between a portion of the RRI data related to the immediately preceding event and a portion of the RRI data related to the selected event. Specifically, the similarity may be calculated based on a correlation coefficient between a portion of the RRI data related to the SP period 602 of the immediately preceding event and a portion of the RRI data related to the SP period 604 of the selected event. Because the SP period 604 is shorter than the SP period 602, the computer 100 may calculate the correlation coefficient between the SP period 604 and a portion of the SP period 602. For example, the computer 100 may calculate the correlation coefficient between the SP period 604 and the period 605 at the end of the SP period 602, and determine this correlation coefficient as the similarity between the immediately preceding event and the selected event. The computer 100 may then compare this similarity with a threshold similarity (e.g., 0.8).

[0040] Alternatively, the computer 100 may calculate a correlation coefficient between the SP period 604 and each of the multiple periods included in the SP period 602 (e.g., period 605, period 606, and period 607), and determine a representative value (e.g., the maximum or average value) of these correlation coefficients as the similarity between the immediately preceding event and the selected event. The multiple periods included in the SP period 602 are different from one another. For example, the computer 100 may determine the multiple periods so that the start times are spaced apart by a predetermined interval.

[0041] The computer 100 may calculate the correlation coefficient using the RRI data acquired in S301 as is, or may calculate the correlation coefficient using the RRI data after removing high frequency components.

[0042] 3, if it is determined in S304 that the independent positive condition is satisfied or if it is determined in S305 that the linked positive condition is satisfied, the computer 100 identifies the selected event as a positive event in S306. On the other hand, if it is determined in S304 that the independent positive condition is not satisfied and if it is determined in S305 that the linked positive condition is not satisfied, the computer 100 identifies the selected event as a negative event in S307. A negative event may be a candidate event that suggests that the patient does not have a CSR disease (i.e., is negative).

[0043] If the selected event is identified as a positive event, in S308, the computer 100 determines whether the selected event satisfies the continuity condition. If the selected event is determined to satisfy the continuity condition ("YES" in S308), the computer 100 transitions the process to S309, and otherwise ("NO" in S308), the computer 100 transitions the process to S310.

[0044] The continuity condition may be a condition used to determine whether a selected event is a supporting event based on the determination results of candidate events other than the selected event. The supporting event may be a candidate event used as a basis for determining whether a patient has CSR disease. For example, the continuity condition may include a threshold number (e.g., three) or more consecutive candidate events identified as positive events.

[0045] An example of the continuity condition will be described with reference to Fig. 7. The upper part of Fig. 7 shows multiple candidate events 701-710 detected in the RRI data. In Fig. 7, the candidate events are represented by "C". As a result of the determination processes of S304 and S305, it is assumed that candidate events 701-704, 706, and 708-710 are identified as positive events, and candidate events 705 and 707 are identified as negative events. In Fig. 7, positive events are represented by "P" and negative events are represented by "N".

[0046] Once a CSR occurs, it tends to occur continuously. Therefore, the computer 100 identifies each of the four consecutive candidate events 701 to 704 as a basis event, and each of the three consecutive candidate events 708 to 710 as a basis event. In FIG. 7, the basis events are represented by "(G)." On the other hand, although the candidate event 706 has been identified as a positive event, it is not consecutive to any other positive events. Therefore, the computer 100 does not identify the candidate event 706 as a basis event.

[0047] The continuity condition may include a threshold time interval (e.g., 20 seconds) between two adjacent positive events in a series of positive events, which prevents continuity from being determined based on two or more weakly related positive events.

[0048] 3, when the computer 100 processes the candidate events sequentially in chronological order, the computer 100 cannot identify the candidate events 701 and 702 as basis events at the stage of processing these candidate events. Therefore, the computer 100 may identify the candidate events 701 and 702 as basis events when processing the candidate event 703.

[0049] As described above, the computer 100 identifies at least one of the multiple candidate events detected in S302 as a basis event by performing multiple determination processes in S303 to S310. Then, in S311, the computer 100 estimates whether the patient has a CSR disease based on the identified basis event. In S311, the computer 100 may further estimate the severity of the CSR disease based on the identified basis event.

[0050] For example, the computer 100 may estimate whether a patient has a CSR disease based on the ratio of the duration of the evidence event to the measurement period of the RRI data. The duration of the evidence event may be the sum of the time from the beginning to the end of a series of consecutively detected evidence events. For example, in the example shown in FIG. 7, the duration of the evidence event may be the sum of the length of period 711 and the length of period 712. Period 711 starts with candidate event 701 (e.g., the beginning of the candidate period) and ends with candidate event 704 (e.g., the end of the candidate period). The same applies to period 712.

[0051] The computer 100 may estimate that the patient has a CSR disease when the rate of evidence events is equal to or greater than a threshold value (e.g., 50%). Furthermore, the computer 100 may estimate the severity of the CSR disease based on the rate of evidence events. For example, the computer 100 may classify the severity of the CSR disease into multiple categories (mild, moderate, severe). For example, the computer 100 may estimate the severity of the CSR disease as mild when the rate of evidence events is equal to or greater than 50% and less than 60%, as moderate when the rate of evidence events is equal to or greater than 60% and less than 80%, and as severe when the rate of evidence events is equal to or greater than 90%.

[0052] In S312, the computer 100 outputs the estimation result of S311. For example, the computer 100 may display the estimation result on the display device 104. Based on this display, a user of the computer 100 (e.g., a patient or a doctor) can recognize the estimation result. In addition to or instead of displaying the estimation result, the computer 100 may store the estimation result in the secondary storage device 106 or transmit it to another device.

[0053] As described above, the computer 100 performs a determination process (e.g., S304) of whether one candidate event satisfies a condition not based on the determination results of other candidate events, and a determination process (e.g., S307, S308) of whether one candidate event satisfies a condition based on the determination results of other candidate events, thereby enabling the computer 100 to accurately estimate whether a patient has CSR disease.

[0054] 3, computer 100 may omit S305. Even in this case, it is possible to prevent the patient from being estimated as a false negative. Alternatively, computer 100 may omit S308. Even in this case, it is possible to prevent the patient from being estimated as a false positive.

[0055] The invention is not limited to the above-described embodiment, and various modifications and variations are possible within the scope of the gist of the invention. [Explanation of symbols]

[0056] 100 Computer, 110 Measuring device

Claims

1. 1. An estimation device for estimating whether a patient has a disease accompanied by Cheyne-Stokes respiration, comprising: a detection means for detecting, based on fluctuations in the time-series data of the patient's beat intervals, a plurality of candidate events occurring during a measurement period of the time-series data; an identification means for identifying at least one of the plurality of candidate events as a basis event by performing a plurality of determination processes; and an estimation means for estimating whether the patient has a disease accompanied by Cheyne-Stokes respiration based on the evidence event; The plurality of determination processes include: a first type determination process for determining whether a first candidate event among the plurality of candidate events satisfies a first condition that is not based on a determination result of a second candidate event among the plurality of candidate events; a second type of determination process for determining whether the first candidate event satisfies a second condition based on the determination result of the second candidate event; An estimation device comprising:

2. The estimation device according to claim 1 , wherein the plurality of determination processes include, as the first type of determination process, a first determination process that includes determining whether the beat interval in the first candidate event decreases so as to satisfy a third condition.

3. The estimation device according to claim 2 , wherein the third condition includes a condition in which a time from the start of the decrease in the beat interval to the end of the decrease is longer than a threshold time.

4. The plurality of determination processes are the second type determination processes, determining whether the beat interval in the second candidate event decreases to satisfy a third condition; determining whether a similarity between the first candidate event and the second candidate event is greater than a threshold similarity; The estimation device according to claim 1 , further comprising a second determination process including:

5. 5. The estimation device according to claim 4, wherein the similarity between the first candidate event and the second candidate event is based on a correlation coefficient between a first portion of the time series data that is associated with the first candidate event and a second portion of the time series data that is associated with the second candidate event.

6. 6. The estimation device of claim 5, wherein the similarity between the first candidate event and the second candidate event is further based on a correlation coefficient between the first portion of the time series data and a third portion of the time series data that is related to the second candidate event and different from the second portion.

7. The estimation device according to claim 1 , wherein the plurality of determination processes include, as the second type determination process, a third determination process that determines whether a threshold number or more of candidate events determined to satisfy a fourth condition are detected consecutively.

8. The estimation device according to claim 7 , wherein the fourth condition includes that the beat interval decreases so as to satisfy a third condition.

9. The estimation device according to claim 8 , wherein the fourth condition further includes a condition that a similarity to a candidate event determined to decrease in the beat interval so as to satisfy the third condition is greater than a threshold similarity.

10. The identification means Among the plurality of candidate events, The beat interval is decreased so as to satisfy a third condition; and and a similarity between the candidate event and the beat interval determined to decrease so as to satisfy the third condition is greater than a threshold similarity, as a positive event. The estimation device according to claim 1 , wherein a threshold number or more of the positive events are identified as the evidence event.

11. The estimation device according to claim 1 , wherein each of the plurality of candidate events is associated with a minimum point of a waveform based on the time-series data.

12. The estimation device according to claim 1 , wherein the estimation means estimates whether the patient has the disease based on a ratio of a duration of the evidence event to the measurement period.

13. The estimation device according to claim 1 , wherein the estimation means further estimates the severity of the disease based on the underlying event.

14. A program for causing a computer to function as each of the means of the estimation device according to any one of claims 1 to 13.

15. 1. A method for predicting whether a patient has a disease associated with Cheyne-Stokes respiration, comprising: a detection step of detecting a plurality of candidate events occurring during a measurement period of the time series data based on fluctuations in the time series data of the patient's beat intervals; an identification step of identifying at least one of the plurality of candidate events as a basis event by performing a plurality of determination processes; and estimating whether the patient has a Cheyne-Stokes respiration disorder based on the evidence event; The plurality of determination processes include: a first type determination process for determining whether a first candidate event among the plurality of candidate events satisfies a first condition that is not based on a determination result of a second candidate event among the plurality of candidate events; a second type of determination process for determining whether the first candidate event satisfies a second condition based on the determination result of the second candidate event; An estimation method, including:

Images