Index estimation system, estimation formula creation system, index estimation method, estimation formula creation method, and index presentation system

Abstract

This index estimation system, in order to improve estimation accuracy of an index related to a blood component of a subject, has an estimation unit that estimates an index related to a blood component of the subject by using a pulse wave feature amount indicating a feature of a pulse wave of the subject and an elapsed time from a predetermined action of the subject at a point in time when the pulse wave was acquired. The estimation unit estimates the index using an estimation formula.

Description

Index estimation system, estimation formula creation system, index estimation method, estimation formula creation method, and index presentation system

[0001] The present disclosure relates to an index estimation system that estimates an index related to a blood component of a subject.

[0002] Blood glucose levels are known as one of the indicators of physical health. Although it is possible to understand the state of blood, including blood glucose levels, drawing blood is painful, and doing it on a daily basis places a significant burden on patients.

[0003] For example, Patent Document 1 describes an electronic device that estimates a subject's blood glucose level using a detected pulse wave of the subject and an estimation formula created based on the subject's blood glucose level when fasting and the pulse wave and blood glucose level after a meal.

[0004] Japanese Patent Application Publication No. 2020-199308

[0005] An index estimation system according to one aspect of the present disclosure includes an estimation unit that estimates a first index related to a blood component of a subject using a pulse wave feature that indicates characteristics of the subject's pulse wave and the elapsed time since the subject performed a predetermined action prior to acquiring the pulse wave, and the estimation unit estimates the first index using an estimation formula.

[0006] An estimation formula creation system according to one aspect of the present disclosure includes an acquisition unit that acquires an index related to blood components obtained by drawing blood from at least one subject, a pulse wave feature that indicates characteristics of the subject's pulse wave detected at the same time as the blood drawing, and the elapsed time from a predetermined action of the subject to the blood drawing; and an estimation formula creation unit that creates an estimation formula that estimates the index from the pulse wave feature and the elapsed time by treating the index, the pulse wave feature, and the elapsed time as a set of data and using multiple sets of data as learning data through machine learning.

[0007] An index estimation method according to one aspect of the present disclosure includes an acquisition step of acquiring pulse wave features indicating characteristics of a subject's pulse wave and the elapsed time since a predetermined action, and an estimation step of estimating a first index related to the subject's blood glucose using the pulse wave features and the elapsed time, wherein the estimation step uses a set of data including a second index that is the same as or related to the first index and is obtained by a method different from the method used to estimate the first index, the pulse wave features detected at the same time as the acquisition of the second index, and the elapsed time from the predetermined action to the acquisition of the second index, and estimates the first index using an estimation formula created using a plurality of the set of data acquired at multiple times.

[0008] An estimation formula creation method according to one aspect of the present disclosure includes an acquisition step of acquiring an index related to blood components obtained by drawing blood from at least one subject, a pulse wave feature amount indicating characteristics of the subject's pulse wave detected at the same time as the blood drawing, and the elapsed time from a predetermined action of the subject to the blood drawing; and an estimation formula creation step of creating an estimation formula that estimates the index from the pulse wave feature amount and the elapsed time by treating the index, the pulse wave feature amount, and the elapsed time as a set of data and using machine learning to learn a plurality of the set of data.

[0009] An index presentation system according to one aspect of the present disclosure includes an estimation unit that estimates an index related to blood components of a subject using pulse wave feature quantities that indicate characteristics of the subject's pulse wave and the elapsed time since the subject performed a predetermined action, and a presentation unit that presents the estimation result by the estimation unit to the subject.

[0010] 1 is a schematic diagram showing an example of an index estimation system according to an embodiment of the present disclosure; FIG. 1 is a functional block diagram showing a configuration of a main part of the index estimation system; FIG. 2 is a diagram for explaining an example of a pulse wave feature amount; FIG. 3 is a diagram for explaining an example of a pulse wave feature amount; FIG. 4 is a diagram for explaining an example of a pulse wave feature amount; FIG. 5 is a diagram for explaining an example of a pulse wave feature amount; FIG. 6 is a diagram for explaining an example of a pulse wave feature amount; FIG. 7 is a diagram showing an example of an estimation device; FIG. 8 is a diagram showing an example of learning data used in an estimation formula creation device; FIG. 1 is a functional block diagram showing the configuration of the main parts of an estimation device according to another embodiment of the present disclosure. FIG. 2 is a functional block diagram showing the configuration of the main parts of an estimation device and an estimation formula creation device according to yet another embodiment of the present disclosure. FIG. 3 is a flowchart showing the flow of a process of proposing and updating the re-creation of an estimation formula in the estimation device. FIG. 4 is a graph showing the relationship between an index and a blood glucose level. FIG. 5 is a functional block diagram showing the configuration of the main parts of an estimation device and an estimation formula creation device according to yet another embodiment of the present disclosure. FIG. 6 is a flowchart showing the flow of a process of estimating blood component indexes in the estimation device. FIG. 7 is a functional block diagram showing the configuration of the main parts of an estimation device and an estimation formula creation device according to yet another embodiment of the present disclosure.

[0011] [Embodiment 1] [Overview] An index estimation system 1 according to this embodiment estimates indexes related to blood components of a subject. Examples of indexes related to blood components include indexes related to blood glucose, indexes related to lipids, indexes related to blood proteins, and indexes related to amino acids. Conventionally, such indexes related to blood components (hereinafter also referred to as blood component indexes) have required measurement by drawing blood, which causes stress for the subject. The index estimation system 1 according to this embodiment can estimate the blood component indexes of a subject without drawing blood, thereby reducing stress for the subject.

[0012] Blood component indexes are indexes that indicate components in the blood, such as an index related to blood glucose, an index related to lipids, or an index related to blood protein. An index related to blood glucose may be, for example, a blood glucose level, or an index related to glucose metabolism. An index related to lipids may be, for example, a triglyceride level, or an index related to lipid metabolism. An index related to blood protein may be, for example, a total protein level, or an index related to protein metabolism.

[0013] In this embodiment, the blood component index estimated by the estimation device 10 is also referred to as the first index, and the blood component index measured by blood sampling or the like from the subject or measured person is also referred to as the second index.

[0014] Fig. 1 shows an example of an index estimation system 1. As shown in Fig. 1, the index estimation system 1 includes, by way of example, an estimation device 10, a detection device 20, an estimation formula creation device 30, and a presentation device 40. As will be described in detail later, the index estimation system 1 does not necessarily have to include all four of these devices.

[0015] In the index estimation system 1, a wearable detection device 20 is used to measure the pulse wave of a subject. An estimation device 10 estimates a first index, which is an index of blood components of the subject, using the pulse wave measured by the detection device 20 and an estimation formula 300 created by an estimation formula creation device 30. A presentation device 40 presents the first index estimated by the estimation device 10, etc. The estimation formula creation device 30 uses training data 50 to create the estimation formula 300 used by the estimation device 10. As will be described in detail later, the training data 50 includes the elapsed time from a predetermined action. By performing estimation using the estimation formula 300 created by machine learning using the training data 50 including the elapsed time, the accuracy of estimation of the first index can be improved.

[0016] Furthermore, since the first index is estimated from the measured pulse wave using the wearable detection device 20, the first index can be estimated without causing stress to the subject.

[0017] [Details of Each Device] Next, the estimation device 10, detection device 20, estimation formula creation device 30, and presentation device 40 included in the index estimation system 1 will be described with reference to Fig. 2. Fig. 2 is a functional block diagram showing the configuration of the main parts of the index estimation system 1.

[0018] Since the first index is estimated by the estimation device 10 and the second index is measured by blood sampling, etc., it can be said that the second index is obtained by a method different from the method used to estimate the first index. Furthermore, the first index and the second index both indicate blood component indexes, and can be said to be the same or related indexes.

[0019] [Estimation Device 10 ] As shown in FIG. 2 , the estimation device 10 includes a first acquisition unit 101 , an estimation unit 102 , a first output unit 103 , and a storage unit 104 .

[0020] The first acquisition unit 101 acquires, from the detection device 20, a pulse wave signal of the subject detected by the detection device 20. The first acquisition unit 101 may acquire the pulse wave of the subject detected by the detection device 20 in real time. That is, the time when the first acquisition unit 101 acquires the pulse wave signal may be the time when the detection device 20 detects the pulse wave of the subject. The first acquisition unit 101 may acquire the pulse wave of the subject stored in the detection device 20. In this case, the pulse wave of the subject stored in the detection device 20 may be stored together with information about the time when the pulse wave was detected. That is, the first acquisition unit 101 may acquire the pulse wave signal of the subject together with information about the time when the pulse wave was detected.

[0021] The estimation unit 102 derives pulse wave features from the subject's pulse wave signal acquired by the first acquisition unit 101, and estimates the subject's first index using the derived pulse wave features, the elapsed time since the subject's specified behavior, and the estimation formula 300 stored in the memory unit 104.

[0022] The subject may input the fact that he or she has performed the predetermined behavior into the estimation device 10, and the estimation unit 102 may use the elapsed time from the time of input. Alternatively, the estimation unit 102 may determine the elapsed time from the predetermined behavior of the subject from a determination result by the determination unit 108, which will be described later. The predetermined behavior is a behavior in which the amount of fluctuation in blood components per unit time exceeds a threshold, such as eating, exercising, or bathing.

[0023] The pulse wave feature amount refers to, for example, at least one of the following:

[0024] (1) The ratio of the amplitude of the reflected wave to the maximum amplitude of the pulse wave, or the ratio of the amplitude of the pulse wave after a predetermined time has elapsed from the time of the maximum amplitude to the maximum amplitude of the pulse wave. (2) The time from a predetermined time determined based on the pulse wave to the appearance of the reflected wave, or the ratio of the amplitude of the reflected wave to the amplitude at a predetermined time determined based on the pulse wave. The time determined based on the pulse wave may be determined based on a velocity pulse wave obtained by subtracting the pulse wave, or an acceleration pulse wave obtained by subtracting the pulse wave twice. The difference may be a derivative. Furthermore, the time determined based on the pulse wave may be, for example, the time when the maximum amplitude is reached.

[0025] (3) Pulse rate within a predetermined time period (4) Time from the rise of the pulse wave to the peak (5) Area of ​​the AC component of the pulse wave (6) Ratio of the AC component to the DC component of the pulse wave (7) Amount of change in the ratio of the amplitude of the reflected wave to the maximum amplitude of the pulse wave before and after a predetermined action, or Amount of change in the ratio of the amplitude of the reflected wave to the amplitude at a predetermined time point determined based on the pulse waves before and after a predetermined action Examples of pulse wave feature quantities are shown in Figures 3 to 6. Figure 3 is a diagram for explaining an example of the ratio of the amplitude of the reflected wave to the maximum amplitude of the pulse wave, or the ratio of the amplitude after a predetermined time has elapsed from the time of the maximum amplitude to the maximum amplitude of the pulse wave, as described in (1) above.

[0026] FIG. 3 shows the pulse wave PW1. P0 in FIG. 3 indicates the maximum amplitude of the pulse wave PW1. P1 indicates the amplitude 100 milliseconds after the time of the maximum amplitude. P2 indicates the amplitude of the reflected wave. P3 indicates the amplitude 120 milliseconds after the time of the maximum amplitude. Therefore, P2 / P0 indicates the ratio of the amplitude of the reflected wave to the maximum amplitude of the pulse wave. Furthermore, P1 / P0 indicates the ratio of the amplitude 100 milliseconds after the time of the maximum amplitude to the maximum amplitude of the pulse wave. Furthermore, P3 / P0 indicates the ratio of the amplitude 120 milliseconds after the time of the maximum amplitude to the maximum amplitude of the pulse wave.

[0027] FIG. 4 is a diagram illustrating an example of the time from a predetermined time point determined based on the pulse wave to the appearance of a reflected wave, as described above in (2), and the ratio of the amplitude of the reflected wave to the amplitude at the predetermined time point determined based on the pulse wave. In FIG. 4 , 4001 indicates a pulse wave PW2. 4002 indicates a velocity pulse wave PW21 obtained by subtracting the pulse wave PW2 once, and 4003 indicates an acceleration pulse wave PW22 obtained by subtracting the pulse wave PW2 twice. The time point RW at which the reflected wave appears may be the time point at which the acceleration pulse wave PW22 reaches its second maximum value. The predetermined time point determined based on the pulse wave may be, for example, the time point at which the velocity pulse wave PW21 has maximum amplitude. In this case, the time from the time point at which the velocity pulse wave PW21 has maximum amplitude to the appearance of the reflected wave is the time dt1 from point a1 on the velocity pulse wave PW21 to the time point RW of the reflected wave.

[0028] The predetermined time point determined based on the pulse wave may be, for example, the time point at which the acceleration pulse wave PW22 reaches its first maximum value. In this case, the time from the time point at which the acceleration pulse wave PW22 reaches its first maximum value to the time point at which the reflected wave appears is the time dt2 from the time point a2 on the acceleration pulse wave PW22 to the time point RW of the reflected wave.

[0029] Furthermore, the predetermined time point determined based on the pulse wave may be, for example, the time point when the ejection wave appears. The time point when the ejection wave appears may be the time point when the acceleration pulse wave PW22 reaches its first minimum value. In this case, the time from the time point when the acceleration pulse wave PW22 reaches its first minimum value to the time point when the reflected wave appears is the time dt3 from point a3 on the acceleration pulse wave PW22 to time point RW of the reflected wave.

[0030] 5 is a diagram illustrating the time from the rising edge of the pulse wave to the peak, as described above in (4). In the pulse wave PW3 shown in FIG. 5, the time from the rising edge to the peak refers to the time UT from the rising start point b1 of the pulse wave PW3 to the peak point b2.

[0031] FIG. 6 is a diagram illustrating the area of ​​the AC component of the pulse wave (5) and the ratio of the AC component to the DC component of the pulse wave (6). In the pulse wave PW4 shown in FIG. 6, the fluctuating portion is the AC component and the fixed portion is the DC component. The area of ​​the AC component of the pulse wave PW4 refers to the area SA of the AC component over a predetermined period. Furthermore, the ratio of the AC component to the DC component of the pulse wave PW4 refers to the area SA of the AC component over a predetermined period relative to the area SD of the DC component. Therefore, SA / SD indicates the ratio of the AC component to the DC component of the pulse wave PW4.

[0032] FIG. 7 is a diagram illustrating the change in the ratio of the amplitude of the reflected wave to the maximum amplitude of the pulse wave before and after the predetermined behavior described above in (7). Reference numeral 701 in FIG. 7 shows a pulse wave PW5 of a subject before a meal, which is a predetermined behavior. Reference numeral 702 in FIG. 7 shows a pulse wave PW6 of the subject 75 minutes after a meal. The ratio of the amplitude of the reflected wave to the maximum amplitude of the pulse wave may be P2 / P0, as described in (1) above. The change in the ratio before and after the predetermined behavior may be the change in P2 / P0 before and after a meal. For example, if P2 / P0 = 105.56 for pulse wave PW5 and P2 / P0 = 96.53 for pulse wave PW6, the change is 105.56 - 96.53 = +9.03. Furthermore, the change in the ratio of the amplitude of the reflected wave to the amplitude at a predetermined time point determined based on the pulse wave before and after the predetermined action in (7) above may be determined by subtracting the ratio of the amplitude of the reflected wave to the amplitude at a predetermined time point determined based on the pulse wave after the predetermined action from the ratio of the amplitude of the reflected wave to the amplitude at a predetermined time point determined based on the pulse wave before the predetermined action. The ratio of the amplitude of the reflected wave to the amplitude at a predetermined time point determined based on the pulse wave may be determined by dividing the amplitude of the reflected wave by the amplitude at the predetermined time point determined based on the pulse wave. The amplitude at a predetermined time point determined based on the pulse wave may be, for example, the amplitude at the time when the ejection wave appears, or the amplitude of the pulse wave at the time when the first minimum value of the accelerated pulse wave is reached.

[0033] The first output unit 103 outputs the estimation result from the estimation unit 102 to the presentation device 40 .

[0034] [Detection Device 20] As shown in FIG. 2, the detection device 20 includes a drive unit 201, a light source unit 202, a light receiving unit 203, and a conversion unit 204.

[0035] The driving unit 201 controls the light emitted from the light source unit 202 .

[0036] The light source unit 202 is configured by, for example, a light emitting diode, and emits light (for example, red light or infrared light) toward the inside of the subject's body under the control of the drive unit 201. The light source unit 202 emits light toward, for example, the blood in the subject's blood vessels.

[0037] The light receiving unit 203 receives light that is emitted from the light source unit 202 and is reflected, transmitted, or scattered inside the subject's body. The light receiving unit 203 is, for example, a photodiode.

[0038] The converter 204 performs AD conversion on the signal indicating the change in light received by the light receiver 203 and outputs the result as a pulse wave signal to the estimation device 10. Since the change in light received by the light receiver 203 indicates a change in the blood volume of the subject, the signal AD converted by the converter 204 indicates the volume pulse wave of the subject. Therefore, the signal output from the converter 204 can be called a pulse wave signal.

[0039] Fig. 8 shows examples of the detection device 20. As shown in Fig. 8, the detection device 20 may be a wearable device. In the example shown in Fig. 8, detection device 20A and detection device 20B are shown as ring-type detection devices 20 that can be worn on the subject's finger. Also, detection device 20C and detection device 20D are shown as watch-type detection devices 20 that can be worn on the subject's wrist.

[0040] Detection device 20A has a structure in which light source unit 202 and light receiving unit 203 are located on the pad of the finger when worn. Detection device 20B has a structure in which light source unit 202 and light receiving unit 203 are located on the side of the finger when worn. Detection device 20C has a structure in which light source unit 202 and light receiving unit 203 are located on the inside of the subject's wrist, i.e., on the side where the wrist is palmar flexed, when worn. Detection device 20D has a structure in which light source unit 202 and light receiving unit 203 are located on the outside of the subject's wrist, i.e., on the side where the wrist is dorsiflexed, when worn.

[0041] [Estimation Formula Creation Device 30 ] As shown in FIG. 2 , the estimation formula creation device 30 includes a second acquisition unit 301 , an estimation formula creation unit 302 , and a second output unit 303 .

[0042] The second acquisition unit 301 acquires element data included in the training data 50 used to create the estimation formula 300. FIG. 9 shows an example of the element data included in the training data 50. As shown in FIG. 9, the element data included in the training data 50 include a pulse wave feature amount, a second index which is a blood component index, and elapsed time. The training data 50 is a set of data including the pulse wave feature amount, the second index, and the elapsed time.

[0043] The timing of detecting the pulse wave from which the pulse wave feature amount is derived and the timing of measuring the second index may be the same, but they are not limited to being the same timing and may be detected and measured at different times as long as there is no medically significant difference.

[0044] The second index included in the training data 50 may be calculated by a method different from the method of estimation by the estimation unit 102 using the estimation formula 300. For example, the second index included in the training data 50 may be calculated by analyzing a body fluid such as blood, interstitial fluid, or saliva. If the second index is a blood glucose level, the blood glucose level may be measured using a blood glucose meter "MediSafeFit" manufactured by Terumo Corporation. If the second index is a triglyceride level, the triglyceride level may be measured using a triglyceride meter "Cobas B 101" manufactured by Roche.

[0045] The second acquisition unit 301 may acquire the pulse wave feature amount, the second index, and the elapsed time separately. In this case, the second acquisition unit 301 may be divided into three units: a second acquisition unit 301A that acquires the pulse wave feature amount, a second acquisition unit 301B that acquires the second index, and a second acquisition unit 301C that acquires the elapsed time.

[0046] The training data 50 may be composed of data of subjects other than the subject, or may include data of the subject as described below. Furthermore, the training data 50 may be composed of data of a single subject, or may be composed of data of multiple subjects. If the training data 50 is composed of data of multiple subjects, the results estimated by the estimation formula 300 can be more accurate.

[0047] Furthermore, the group data may be measured at multiple timings. The multiple timings may include, for example, many group data sets within one hour of a predetermined behavior. The multiple timings may span multiple predetermined behaviors. For example, the multiple timings may include at least one time before a predetermined behavior and at least one time after the predetermined behavior. For example, in the case where a specific predetermined behavior A is performed, the multiple timings may include at least one time before behavior A and at least one time after behavior A. The group data may always include data sets before the predetermined behavior.

[0048] For example, if the predetermined behavior is eating, the accuracy of estimating the first index can be improved by including many sets of data within one hour of starting the meal, that is, by increasing the frequency of acquiring data within one hour of starting the meal, because the amount of fluctuation in blood component indices is large from about 30 minutes to two hours after starting the meal.

[0049] The element data included in the training data 50 may include, in addition to the element data described above, information indicating the attributes of the subject or examinee, such as age and sex.

[0050] The estimation formula creation unit 302 creates the estimation formula 300 through machine learning using the learning data 50 acquired by the second acquisition unit 301. More specifically, the estimation formula creation unit 302 performs regression analysis using the pulse wave feature amount and the elapsed time as explanatory variables and the second index as a response variable, and creates the estimation formula 300 that estimates the first index from the pulse wave feature amount and the elapsed time.

[0051] The second output unit 303 outputs the estimation equation 300 created by the estimation equation creation unit 302 to the estimating device 10 .

[0052] 2 , the presentation device 40 includes a third acquisition unit 401 and a presentation unit 402. The third acquisition unit 401 acquires an estimation result from the estimation device 10. The presentation unit 402 presents the estimation result acquired by the third acquisition unit 401. The presentation unit 402 may be a display device that displays data, or may be a speaker that outputs audio or the like. Alternatively, the presentation unit 402 may be both.

[0053] Furthermore, the presentation unit 402 may present the elapsed time from a predetermined behavior along with the estimation result by the estimation device 10. Here, when there are multiple estimation results for one predetermined behavior, the presentation unit 402 may present the elapsed time for at least one estimation result together with the estimation result. For example, the presentation unit 402 may present the elapsed time for the estimation result with the largest and / or smallest value together with the estimation result. The presentation unit 402 does not need to present estimation results other than the estimation result with the largest and / or smallest value together with the elapsed time. This makes it easier for the subject to notice results that are of interest to him or her.

[0054] [Flow of Estimation Formula Creation Process] Next, the flow of the process of creating the estimation formula 300 will be described with reference to Fig. 10. Fig. 10 is a flowchart showing the flow of the process of creating the estimation formula 300.

[0055] The process of creating the estimated formula 300 first involves acquiring the training data 50. The training data 50 may be acquired by the estimated formula creation device 30, or may be acquired by a device different from the estimated formula creation device 30, and the acquired training data 50 may then be acquired by the estimated formula creation device 30.

[0056] In the following, a case where the learning data 50 is acquired in the estimation equation generation device 30 will be described as an example.

[0057] 10 , in the process of creating the estimation formula 300, first, a set of data is obtained that includes the subject's pulse wave and second index measured at a certain point in time, for example, before a predetermined action, and the elapsed time. In this case, since the predetermined action has not yet occurred, the elapsed time is 0 minutes (S101). The pulse wave, second index, and elapsed time may be obtained separately and then combined into a single set of data.

[0058] Next, after the subject begins a predetermined action, counting of the elapsed time begins (S102). The estimation formula creation device 30 sets a measurement time tn (n is the number of measurements) (S103). When the elapsed time reaches the measurement time tn (YES in S104), the subject's pulse wave, the second index, and the elapsed time are acquired (S105, acquisition step). If the elapsed time does not exceed the maximum measurement time tmax (NO in S106), the process proceeds to step S107, where n = n + 1 is set, and the process returns to step S103 to set the next measurement time tn. When the elapsed time reaches the next measurement time tn (YES in S104), the subject's pulse wave, the second index, and the elapsed time are acquired (S105). In other words, steps S103 to S105 are repeated until the elapsed time reaches the maximum measurement time tmax.

[0059] An example of the measurement time tn will now be described with reference to Fig. 11. Fig. 11 is a diagram showing an example of the measurement time tn for measuring the subject's pulse wave and the second index, which is used in the learning data 50. Here, the start time of a meal, which is a predetermined behavior, is set to t0 = 0 minutes, and the subject's pulse wave and the second index are measured at t1 = 45 minutes, t2 = 60 minutes, t3 = 75 minutes, t4 = 90 minutes, t5 = 120 minutes, t6 = 180 minutes, t7 = 240 minutes, t8 = 300 minutes, and t9 = 360 minutes.

[0060] If the elapsed time exceeds the maximum measurement time tmax (YES in S106), the estimation formula creation unit 302 derives pulse wave feature values ​​for each acquired pulse wave (S108). The estimation formula creation unit 302 then acquires the pulse wave feature values, second index, and elapsed time of the pulse waves acquired during the measurement time as a set of data (S109). At this time, the estimation formula creation unit 302 may acquire the pulse wave feature values, second index, and elapsed time of the pulse waves acquired at any measurement time tn as a set of data, and may acquire the set of data acquired up to the maximum measurement time tmax as each set of data. Next, the estimation formula creation device 30 determines whether the number of set data acquired in step S109 is greater than a predetermined number N (S110). If it determines that the number of set data is equal to or less than the predetermined number N (NO in S110), the process returns to step S101. At this time, data acquisition for a different subject may be started, or data acquisition for the same subject performing different predetermined actions may be performed. If it is determined that the number of group data is equal to or greater than the predetermined number N (YES in S110), a regression analysis is performed using the pulse wave feature amount and elapsed time as explanatory variables and the measured second index as a response variable to create an estimated formula 300 (S111, estimated formula creation step).

[0061] Furthermore, the group data used for the learning data 50 may be a set of four data items: (1) a pulse wave at elapsed time 0, (2) a pulse wave feature amount at time tn, (3) a second index at time tn, and (4) elapsed time at time tn. In other words, the group data may always include a pulse wave at elapsed time 0.

[0062] [Examples of Training Data and Verification of Search Formula] When creating the estimation formula 300 using the training data 50, the estimation formula creation device 30 may perform leave-one-out cross-validation (LOOCV) to verify the accuracy of the training data 50.

[0063] Fig. 12 shows a functional block diagram of an estimation formula creation device 30A. In the estimation formula creation device 30A shown in Fig. 12, an estimation formula creation unit 302 includes a verification unit 321. The verification unit 321 performs leave-one-out cross validation on the training data 50 to verify the accuracy of the training data 50. The estimation formula creation unit 302 creates an estimation formula 300 using training data 50 whose accuracy is guaranteed based on the verification results of the verification unit 321. This makes it possible to create an estimation formula 300 using training data 50 with high accuracy.

[0064] [Example of Changes in Blood Glucose Levels After a Meal] FIGS. 13 to 16 show examples of training data 50. FIG. 13 is a graph showing an example of changes in blood glucose levels after a subject starts a meal. The example shown in FIG. 13 shows changes in blood glucose levels over six days when the subject eats two chicken and mixed rice balls. The horizontal axis of FIG. 13 represents time (minutes), and the vertical axis represents blood glucose levels (mg / dL) or the index AI (%). The index here represents the ratio of the amplitude of the reflected wave to the maximum amplitude of the pulse wave. In FIG. 13, the graph showing data values ​​with circles represents blood glucose levels measured by blood sampling. The graph showing data values ​​with triangles represents an index derived from the measured pulse wave. As shown in FIG. 13, the blood glucose level begins to rise after the start of a meal, reaches its maximum between 50 and 100 minutes, and then begins to decline. It can also be seen that the index AI changes with changes in blood glucose levels.

[0065] Similar to FIG. 13 , FIG. 14 is a graph showing an example of changes in blood glucose levels after a subject starts a meal. The example shown in FIG. 14 shows changes in blood glucose levels over six days when the subject ate two tuna mayonnaise rice balls and a mixed sandwich. The horizontal axis of FIG. 14 represents time (minutes), and the vertical axis represents blood glucose levels (mg / dL) or the index AI (%). The index here represents the ratio of the amplitude of the reflected wave to the maximum amplitude of the pulse wave. In FIG. 14 , the graph showing data values ​​represented by circles represents blood glucose levels measured by blood sampling. The graph showing data values ​​represented by triangles represents an index derived from the measured pulse wave. As shown in FIG. 14 , the blood glucose level begins to rise after the start of the meal, reaches its maximum between 50 and 100 minutes, and then begins to decline. It can also be seen that the index AI changes with changes in blood glucose levels.

[0066] FIG. 15 is a graph showing an example of changes in blood glucose levels after a subject starts a meal, when the subject eats a meal followed by another meal at a later time. The example shown in FIG. 15 shows changes in blood glucose levels when the subject eats a tuna mayonnaise rice ball and then eats a mixed sandwich 90 minutes later. The horizontal axis of FIG. 15 represents time (minutes), and the vertical axis represents blood glucose level (mg / dL) or index AI (%). The index here represents the ratio of the amplitude of the reflected wave to the maximum amplitude of the pulse wave. In FIG. 15 , the graph showing data values ​​represented by circles represents blood glucose levels measured by blood sampling. The graph showing data values ​​represented by triangles represents an index derived from the measured pulse wave. As shown in FIG. 15 , blood glucose levels begin to rise after the start of the meal, reach a maximum around 50 minutes later, then decline, but rise again around 90 minutes later, and then decline gradually. It can also be seen that the index AI changes with changes in blood glucose levels.

[0067] FIG. 16 is a graph showing an example of changes in blood glucose levels after a subject starts a meal when the subject eats a meal and then eats another meal after a certain time interval. Similar to FIG. 15 , the example shown in FIG. 16 shows changes in blood glucose levels when the subject eats a tuna mayonnaise rice ball and then eats a mixed sandwich 90 minutes later. The horizontal axis of FIG. 16 represents time (minutes), and the vertical axis represents blood glucose level (mg / dL) or index AI (%). The index here represents the ratio of the amplitude of the reflected wave to the maximum amplitude of the pulse wave. In FIG. 16 , the graph showing data values ​​represented by circles represents blood glucose levels measured by blood sampling. The graph showing data values ​​represented by triangles represents an index derived from the measured pulse wave. In the example shown in FIG. 16 , the blood glucose level begins to rise after the start of the meal, reaches a peak around 50 minutes later, then drops temporarily, but rises again. Thereafter, it drops rapidly. Furthermore, while the blood glucose level changes, the index AI remains almost unchanged.

[0068] FIG. 17 shows the results of leave-one-out cross-validation (LOOCV) performed on the estimation formula 300 created by performing a regression analysis using the data shown in FIGS. 13 to 16 as training data 50. The graph in FIG. 17 shows measured values ​​on the horizontal axis and estimated values ​​on the vertical axis. As shown in FIG. 17, when a line 801 is assumed to have a slope of 1, the validation results of the estimation formula 300 are approximately aligned with the line 801, indicating that the data shown in FIGS. 13 to 16 are highly accurate as training data.

[0069] [Flow of First Index Estimation Processing] Next, the flow of the first index estimation processing in the estimation device 10 will be described with reference to Fig. 18. Fig. 18 is a flowchart showing the flow of the first index estimation processing in the estimation device 10.

[0070] 18 , the estimation device 10 starts counting the elapsed time when the subject starts a predetermined behavior (S201). The start of the predetermined behavior may be input by the subject. Alternatively, the start time of the predetermined behavior may be determined from a determination result by the determination unit 105 (described later).

[0071] Next, the estimation device 10 acquires the subject's pulse wave detected by the detection device 20 and acquires the elapsed time since the predetermined behavior (S202, acquisition step). Then, the estimation device 10 estimates the subject's first index from the pulse wave feature quantity of the acquired pulse wave and the elapsed time since the predetermined behavior using the estimation formula 300 (S203, estimation step). The estimation device 10 then outputs the estimation result (S204). The above is the flow of the process for estimating the first index by the estimation device 10.

[0072] 19 , in the above-described embodiment, the index estimation system 1 has been described as including an estimation device 10, a detection device 20, an estimation formula creation device 30, and a presentation device 40. However, the index estimation system 1 does not necessarily have to include all of these four devices.

[0073] For example, as shown in 902 in FIG. 19, a system including the estimation device 10 among the four devices may be used as an index estimation system 1A.

[0074] Furthermore, as shown in 903 in FIG. 19, a system including the estimation device 10 and the detection device 20 among the four devices may be used as an index estimation system 1B.

[0075] 19 , a system including the estimation device 10 and the estimation formula creation device 30 among the four devices may be referred to as an index estimation system 1C. Since the index estimation system 1C includes the estimation formula creation device 30, it can also be called an estimation formula creation system.

[0076] 20 , a system including the estimation device 10 and the presentation device 40 among the four devices may be referred to as an index estimation system 1D. Since the index estimation system 1D includes the presentation device 40, it can also be called an index presentation system.

[0077] 20 , a system including the estimation device 10, the detection device 20, and the estimation formula creation device 30 among the four devices may be referred to as an index estimation system 1E. Since the index estimation system 1E includes the estimation formula creation device 30, it can also be called an estimation formula creation system.

[0078] 20 , a system including the estimation device 10, the detection device 20, and the presentation device 40 among the four devices may be referred to as an index estimation system 1F. Since the index estimation system 1F includes the presentation device 40, it can also be called an index presentation system.

[0079] Furthermore, each device included in the index estimation system 1 (1A to 1F) may be in a separate housing, or multiple devices may be included in a single housing.

[0080] [Embodiment 2] Another embodiment of the present disclosure will be described below. For convenience of explanation, the same reference numerals will be used to designate components having the same functions as those described in the above embodiment, and the description thereof will not be repeated.

[0081] 21 is a functional block diagram showing the configuration of a main part of an estimation device 10A according to this embodiment. As shown in FIG. 21, this embodiment differs from the estimation device 10 described above in that the estimation device 10A includes a determination unit 105.

[0082] The determination unit 105 determines whether the subject has started a predetermined behavior. If it is determined that the subject has started the predetermined behavior, it notifies the estimation unit 102 of this fact. This allows the estimation unit 102 to recognize the time when the subject started the predetermined behavior and to grasp the elapsed time since the predetermined behavior.

[0083] Furthermore, when the determination unit 105 determines that the subject has performed a predetermined behavior, the estimation unit 102 may perform estimation using the subject's pulse wave feature amount at a time when a predetermined time has elapsed since the predetermined behavior. The estimation unit 102 may perform estimation multiple times for one predetermined behavior by the subject. When the predetermined behavior is eating, the estimation unit 102 may perform estimation multiple times within three hours after the subject starts eating, for example, 45 minutes, 60 minutes, 75 minutes, 90 minutes, 120 minutes, 180 minutes, 240 minutes, 300 minutes, and 360 minutes after the subject starts eating.

[0084] The judgment made by the judgment unit 105 may be made using the behavioral history of the subject, which may include, for example, at least one of input by the subject, hand movements of the subject, changes in the subject's pulse rate, changes in the subject's body temperature, and changes in the subject's pulse wave features.

[0085] In addition, the judgment unit 105 may make a judgment using a trained model obtained by machine learning in which at least one of the subject's hand movements, changes in the subject's pulse rate, changes in the subject's body temperature, and changes in the subject's pulse wave features is used as an explanatory variable, and whether or not a specified action has been taken is used as an objective variable.

[0086] [Embodiment 3] Another embodiment of the present disclosure will be described below. For convenience of explanation, the same reference numerals will be used to designate components having the same functions as those described in the above embodiment, and the description thereof will not be repeated.

[0087] 22 is a functional block diagram showing the configuration of a main part of an estimation device 10B according to this embodiment. As shown in FIG. 22, this embodiment differs from the estimation device 10 and the estimation device 10A described above in that the estimation device 10B includes a proposal unit 106.

[0088] In this embodiment, the first acquisition unit 101 acquires a second index, which is a blood component index, obtained by drawing blood from the subject. The proposal unit 106 acquires the second index from the first acquisition unit 101 and also acquires the estimation result from the estimation unit 102. Then, if there is a difference between the second index and the estimation result that is equal to or greater than a threshold, the proposal unit 106 proposes the re-creation of the estimation formula 300. This makes it possible to propose the re-creation of the estimation formula when the estimation result deviates from the second index, thereby preventing a decrease in the estimation accuracy using the estimation formula. The threshold may be, for example, ±15% of the second index. In this case, the timing of drawing blood from the subject and the timing of acquiring the subject's pulse wave data and elapsed time used for estimation by the estimation unit 102 may be the same. Alternatively, the timing of drawing blood from the subject and the timing of acquiring the subject's pulse wave data and elapsed time used for estimation by the estimation unit 102 may be different from each other as long as no medically significant difference occurs.

[0089] Furthermore, when the proposing unit 106 proposes the re-creation of the estimation formula, the above-mentioned second index, the pulse wave data and elapsed time used for the estimation by the estimating unit 102, may be transmitted as additional data to the estimation formula creating device 30, and the estimation formula creating device 30 may then re-create the estimation formula 300.

[0090] In this case, the estimation formula creation device 30 may be configured to recreate the estimation formula 300 using training data 50 including the additional data, and if the error between the estimation result obtained using the recreated estimation formula 300X and the second index does not exceed a threshold, update the estimation formula 300. Furthermore, when recreating the estimation formula 300 using the additional data, the estimation formula creation device 30 may be configured to perform the leave-one-out cross validation described above, and recreate and update the estimation formula 300 if the estimation error does not exceed a threshold.

[0091]

[0111] Next, the flow of the process of proposing and updating the estimated formula 300 to be recreated in the estimating device 10B and the estimated formula creation device 30A will be described with reference to Fig. 23. Fig. 23 is a sequence diagram showing the flow of the process of proposing and updating the estimated formula 300 to be recreated.

[0092] 23 , the estimation device 10B acquires the pulse wave of the subject detected by the detection device 20, a second index that is an index of blood components of the subject obtained by blood sampling, and the elapsed time since a predetermined action (S301). Next, the estimation device 10B estimates the first index of the subject from the pulse wave feature amount of the acquired pulse wave and the elapsed time since the predetermined action using the estimation formula 300 (S302).

[0093] Then, the estimation device 10B compares the second index, which is the acquired measurement result, with the first index, which is the estimation result, and if the difference is greater than or equal to a threshold (e.g., ±15 percent) (YES in S303), it suggests recreating the estimation formula 300 (S304).

[0094] On the other hand, the acquired second index and first index are compared, and if the difference is smaller than the threshold value (NO in S303), the user determines whether to proceed to step S306 (S305).

[0095] After the re-creation of the estimation formula 300 is suggested in step S304, or after the user determines in step S305 that the estimation formula 300 needs to be re-created, the estimation device 10B transmits the second index, the pulse wave data used in the estimation, and the elapsed time as additional data to the estimation formula creation device 30A.

[0096] The estimation-formula creation device 30A adds the acquired additional data to the learning data 50 used to create the already-created estimation formula 300 to create a new estimation formula 300X (S306). Then, using the newly created estimation formula 300X, it estimates a first index from the pulse wave data and elapsed time included in the additional data. If the error between the first index and the second index, which are the estimation results, is smaller than a threshold (YES in S307), the estimation-formula creation device 30A updates the estimation formula from the already-created estimation formula 300 to the newly created estimation formula 300X (S308). On the other hand, if the error between the estimation result and the second index is equal to or greater than the threshold (NO in S307), the estimation-formula creation device 30A does not update the estimation formula 300 and ends the process.

[0097] In step S306, instead of creating a new estimated formula 300X, the above-described leave-one-out cross validation may be performed on the entire training data 50 including the additional data, and if the error in the validation result is equal to or less than a threshold, the estimated formula may be updated. The above is the flow of the process of proposing and updating the re-creation of the estimated formula 300 in the estimating device 10B and the estimation formula creation device 30A.

[0098] Fourth Embodiment Another embodiment of the present disclosure will be described below. For convenience of explanation, the same reference numerals will be used to designate components having the same functions as those described in the above embodiment, and the description thereof will not be repeated.

[0099] FIG. 24 is a graph showing the relationship between the index AI, which is one of the pulse wave feature quantities, and the blood glucose level of a subject after a meal. In the graph shown in FIG. 24, the vertical axis represents the index AI, and the horizontal axis represents the blood glucose level. FIG. 24 shows the index AI and blood glucose levels immediately after a meal (2401), 45 minutes later (2402), 60 minutes later (2403), 75 minutes later (2404), 90 minutes later (2405), 115 minutes later (2406), 120 minutes later (2407), 135 minutes later (2408), 150 minutes later (2409), 165 minutes later (2410), 180 minutes later (2411), 240 minutes later (2412), 300 minutes later (2413), and 360 minutes later (2414). As shown in FIG. 24, the relationship between the index AI and blood glucose level differs between before the blood glucose level peak and after the blood glucose level peak. That is, until the blood glucose level peaks, the blood glucose level rises over time and the index AI decreases (see arrow 2451), but after the blood glucose level peaks, the blood glucose level decreases over time (see arrow 2452) and then returns to a value close to the index AI before the meal (see arrow 2453).

[0100] Therefore, in this embodiment, estimation is performed using different estimation formulas before and after a predetermined timing. The predetermined timing is described as, for example, the peak of blood glucose levels, but it may also be a predetermined width before and after the peak of blood glucose levels. FIG. 25 is a functional block diagram showing the main configuration of an estimation device 10C according to this embodiment. As shown in FIG. 25 , in this embodiment, the estimation formula 300 includes a first estimation formula 300A and a second estimation formula 300B. The estimation unit 102 then estimates the first index of the subject using the first estimation formula 300A and the second estimation formula 300B.

[0101] Specifically, when the elapsed time is within a first hour, the estimation unit 102 performs estimation using the first estimation formula 300A, and when the elapsed time exceeds the first hour, the estimation unit 102 performs estimation using the second estimation formula 300B. The first time is the time at which a person's blood glucose level peaks after a meal, and may be, for example, the following time:

[0102] (1) 60 minutes from the end of the meal. (2) A time set according to the meal contents. (3) A time set by the subject as the time when the blood glucose level will peak. (4) A time set according to the subject's attributes. The subject's attributes include gender, age, fasting blood glucose level, and HbA1c, which is an index related to blood glucose level.

[0103] As described above, the relationship between the pulse wave feature amount and the blood glucose level, as indicated by the index AI, can change significantly around the blood glucose level peak. According to the configuration of this embodiment, the first time period can be set as the time when the blood glucose level peaks, and different estimation formulas, the first estimation formula 300A and the second estimation formula 300B, can be used before and after the blood glucose level peak. This allows the use of an estimation formula that corresponds to the relationship between the pulse wave feature amount and the first index, thereby improving the accuracy of the estimation result of the first index.

[0104] The first estimation formula 300A and the second estimation formula 300B are formulas including a plurality of parameters, including a first parameter indicating the elasticity of blood vessels. As an example, the first estimation formula 300A and the second estimation formula 300B can be expressed as follows: Estimated value = a1 × first parameter + a2 × second parameter + a3 × third parameter + ... + an × nth parameter, where a1, a2, ... an are weights, and a1 + a2 + ... + an = 1. In addition, the parameters include a pulse wave feature amount (index AI) indicating the elasticity of blood vessels, an index related to cardiac function, an index related to the autonomic nervous system, an index related to the attributes of the subject, and the like.

[0105] In the first estimation formula 300A, when the first parameter indicates an index of vascular elasticity, the weighting a1 of the first index is higher than the weightings of the other parameters (a2, a3, ...). That is, a1 > a2, a3, ..., an. Until a first hour has elapsed since the subject ate a meal, there is a high correlation between the blood glucose level and the pulse wave feature, which is an index of vascular elasticity. Therefore, by using the first estimation formula 300A, it is possible to obtain a highly accurate estimation result of the first index until the first hour has elapsed since the subject ate a meal.

[0106] In the second estimation formula 300B, when the first parameter indicates an index of vascular elasticity, the weighting a1 of the first index is lower than the weightings of the other parameters (a2, a3, ...). That is, a1<a2, a3, ..., an. After a first hour has elapsed since the subject ate a meal, the correlation between the blood glucose level and the pulse wave feature, which is an index of vascular elasticity, is low. Therefore, by using the second estimation formula 300B, the estimation result of the first index can be made more accurate after a first hour has elapsed since the subject ate a meal.

[0107] [Estimation Formula Creation Device] The estimation formula creation device 30 creates a first estimation formula 300A and a second estimation formula 300B as estimation formulas 300. The first estimation formula 300A is created using group data whose elapsed time is within a first hour. The second estimation formula 300B is created using group data whose elapsed time exceeds the first hour.

[0108] [Flow of First Index Estimation Processing] Next, the flow of the first index estimation processing in the estimation device 10C will be described with reference to Fig. 26. Fig. 26 is a flowchart showing the flow of the first index estimation processing in the estimation device 10C. Fig. 26 shows a detailed flow of the first index estimation processing in step S203 in the flow of the first index estimation processing shown in Fig. 18.

[0109] As shown in Fig. 26, in the first index estimation process, the estimation unit 102 first determines whether the elapsed time is within a first time period (S2031). If the elapsed time is within the first time period (YES in S2031), the estimation unit 102 performs estimation using the first estimation formula 300A (S2032). On the other hand, if the elapsed time is longer than the first time period (NO in S2031), the estimation unit 102 performs estimation using the second estimation formula 300B (S2033). Then, the process proceeds to step S204 in Fig. 18. The above is a detailed flow of the estimation process in the estimation device 10C.

[0110] [Flow of Estimation Formula Creation Process] The flow of the process of creating the first estimated formula 300A and the second estimated formula 300B by the estimation formula creation device 30 is as follows. In step S111 of the estimation formula creation process shown in FIG. 10 , the estimation formula creation unit 302 performs the following process. That is, the estimation formula creation unit 302 creates the first estimated formula 300A using group data whose elapsed time is within the first hour, and creates the second estimated formula 300B using group data whose elapsed time exceeds the first hour. This makes it possible to create an estimation formula depending on whether the elapsed time is within the first hour.

[0111] [Embodiment 5] Another embodiment of the present disclosure will be described below. For convenience of explanation, the same reference numerals will be used to designate components having the same functions as those described in the above embodiment, and the description thereof will not be repeated.

[0112] 27 is a functional block diagram showing the configuration of the main parts of an estimation device 10D according to this embodiment. As shown in FIG. 27, this embodiment differs from the above-described embodiments in that an evaluation unit 107 is included.

[0113] The evaluation unit 107 derives pulse wave feature values ​​from the subject's pulse wave signal acquired by the first acquisition unit 101, and evaluates the state of the subject's autonomic nerves using the derived pulse wave feature values. For example, the evaluation unit 107 evaluates whether the subject's autonomic nerves are in a sympathetic-dominant state or a parasympathetic-dominant state. Since it is possible to evaluate the state of the subject's autonomic nerves from the pulse wave feature values ​​using known technology, detailed description thereof will be omitted.

[0114] The estimation unit 102 estimates the first index of the subject using different estimation formulas 300 depending on the evaluation by the evaluation unit 107. More specifically, when the evaluation by the evaluation unit 107 indicates sympathetic dominance, the estimation unit 102 performs the estimation using a first-1 estimation formula 300A-1 or a second-1 estimation formula 300B-1. Furthermore, when the evaluation by the evaluation unit 107 indicates parasympathetic dominance, the estimation unit 102 performs the estimation using a first-2 estimation formula 300A-2 or a second-2 estimation formula 300B-2.

[0115] The first-1 estimated formula 300A-1 and the first-2 estimated formula 300A-2 are estimated formulas included in the first estimated formula 300A. The second-1 estimated formula 300B-1 and the second-2 estimated formula 300B-2 are estimated formulas included in the second estimated formula 300B.

[0116] The first-1 estimated formula 300A and the second-1 estimated formula 300B-1 are obtained by correcting the pulse wave feature quantity, which is one of the parameters, to a lower value. The first-2 estimated formula 300A-2 and the second-2 estimated formula 300B-2 are obtained by correcting the pulse wave feature quantity, which is one of the parameters, to a higher value.

[0117] When the autonomic nervous system is sympathetically dominant, blood vessels constrict, and the pulse wave feature quantity tends to take on high values. Conversely, when the autonomic nervous system is parasympathetically dominant, blood vessels dilate, and the pulse wave feature quantity tends to take on low values. According to the configuration of this embodiment, when the autonomic nervous system is sympathetically dominant, the first-1 estimation formula 300A-1 or the second-1 estimation formula 300B-1 is used, and when the autonomic nervous system is parasympathetically dominant, the first-2 estimation formula 300A-2 or the second-2 estimation formula 300B-2 is used. Therefore, whether the autonomic nervous state is sympathetically dominant or parasympathetically dominant, the first index can be estimated using an appropriate estimation formula, and the first index can be estimated with high accuracy.

[0118] In the estimation formula creation device 30, the second acquisition unit 301 acquires element data including an evaluation of the subject's autonomic nerves. The estimation formula creation unit 302 then uses the element data in which the evaluation of the autonomic nerves indicates a sympathetic nerve dominant state to create the first-first estimation formula 300A-1 and the second-first estimation formula 300B-1. The estimation formula creation unit 302 also uses the element data in which the evaluation of the autonomic nerves indicates a parasympathetic nerve dominant state to create the first-second estimation formula 300A-2 and the second-second estimation formula 300B-2. This allows different estimation formulas to be created depending on the state of the autonomic nerves.

[0119] [Example of implementation by software] The functions of the index estimation system 1 (hereinafter referred to as the "system") can be realized by a program that causes a computer to function as the system, and a program that causes a computer to function as each control block of the system (particularly the estimation unit 102 and the estimation equation creation unit 302).

[0120] In this case, the system includes a computer having at least one control device (e.g., a processor) and at least one storage device (e.g., a memory) as hardware for executing the program. The functions described in each of the above embodiments are realized by executing the program using the control device and storage device.

[0121] The program may be non-transitory and may be recorded on one or more computer-readable recording media. The recording media may or may not be included in the device. In the latter case, the program may be supplied to the device via any wired or wireless transmission medium.

[0122] In addition, some or all of the functions of each of the control blocks can be realized by logic circuits. For example, integrated circuits in which logic circuits that function as each of the control blocks are formed are also included in the scope of the present disclosure. In addition, the functions of each of the control blocks can also be realized by, for example, a quantum computer.

[0123] Furthermore, each process described in each of the above embodiments may be executed by AI (Artificial Intelligence). In this case, the AI ​​may run on the control device or on another device (for example, an edge computer or a cloud server).

[0124] [Summary] The index estimation system according to aspect 1 of the present disclosure includes an estimation unit that estimates a first index related to a blood component of a subject using a pulse wave feature amount indicating characteristics of the subject's pulse wave and the elapsed time since the subject performed a predetermined action prior to acquisition of the pulse wave, and the estimation unit estimates the first index using an estimation formula. With this configuration, it is possible to estimate an index taking into account the elapsed time since the predetermined action, thereby improving the accuracy of the index estimation.

[0125] In the index estimation system according to Aspect 2 of the present disclosure, in Aspect 1, the estimation unit estimates the first index using an estimation formula created using, as a set of data, a second index that is the same as or related to the first index and is obtained by a method different from the method used to estimate the first index, the pulse wave feature value detected at the same time as the acquisition of the second index, and the elapsed time from the predetermined behavior to the acquisition of the second index. With this configuration, the index can be estimated using an estimation formula created using the elapsed time from the predetermined behavior as one piece of data, thereby improving the accuracy of index estimation.

[0126] In the index estimation system according to Aspect 3 of the present disclosure, in Aspect 1 or 2, the estimation unit uses the estimation formula created using a plurality of group data sets acquired at a plurality of time points. With this configuration, since a plurality of group data sets is used, the accuracy of the index can be improved.

[0127] The index estimation system according to Aspect 4 of the present disclosure is the system of any one of Aspects 1 to 3, further comprising a determination unit that determines whether the subject has performed the predetermined behavior using an input from the subject or a behavioral history of the subject, and the estimation unit performs the estimation by using the time that has elapsed since the determination unit determined that the subject performed the predetermined behavior as the elapsed time since the predetermined behavior. With this configuration, the elapsed time since the predetermined behavior can be automatically acquired.

[0128] In the index estimation system according to Aspect 5 of the present disclosure, in Aspect 4, the subject's behavior history includes at least one of the subject's hand movements, pulse rate changes, body temperature changes, and pulse wave feature value changes. With this configuration, the subject's hand movements, pulse rate changes, body temperature changes, and pulse wave feature value changes can be used to automatically obtain the elapsed time since a predetermined behavior.

[0129] In the index estimation system according to Aspect 6 of the present disclosure, in any one of Aspects 1 to 5, when the determination unit determines that the subject has performed the predetermined behavior, the estimation unit performs the estimation using the pulse wave feature amount of the subject at a time point when a predetermined time has elapsed since the time of the predetermined behavior. With this configuration, when the predetermined behavior is performed, it is possible to automatically estimate an index related to a blood component.

[0130] In the index estimation system according to Aspect 7 of the present disclosure, in any one of Aspects 1 to 6, the predetermined behavior is a recent behavior in which the amount of fluctuation in the blood constituents of the subject per unit time exceeds a threshold. With this configuration, it is possible to estimate an index related to the blood constituents of the subject after a behavior in which the amount of fluctuation in the blood constituents increases. An example of a behavior in which the amount of fluctuation in the blood constituents per unit time exceeds a threshold is eating.

[0131] In an index estimation system according to an eighth aspect of the present disclosure, in any one of the first to seventh aspects, the pulse wave feature amount is at least one of the ratio of the amplitude of the reflected wave to the maximum amplitude of the pulse wave, the time from a predetermined time point determined based on the pulse wave to the appearance of the reflected wave, the ratio of the amplitude of the reflected wave to the amplitude at a predetermined time point determined based on the pulse wave, the pulse rate within a predetermined time, the shape of the rising edge of the pulse wave, the area of ​​the AC component of the pulse wave, the ratio of the AC component to the DC component of the pulse wave, the amount of change in the ratio of the amplitude of the reflected wave to the maximum amplitude of the pulse wave before and after a predetermined action, and the amount of change in the ratio of the amplitude of the reflected wave to the amplitude at a predetermined time point determined based on the pulse wave before and after a predetermined action.

[0132] An index estimation system according to a ninth aspect of the present disclosure is any of the first to eighth aspects, further comprising an acquisition unit that acquires the second index, and a proposal unit that proposes re-creation of the estimation formula when a difference between the second index and the first index estimated by the estimation unit is equal to or greater than a threshold. With this configuration, when the estimated first index deviates from the second index, it is possible to propose re-creation of the estimation formula.

[0133] An estimation formula creation system according to a tenth aspect of the present disclosure includes an acquisition unit that acquires an index related to a blood component obtained by drawing blood from at least one subject, a pulse wave feature value indicating a characteristic of the subject's pulse wave detected at the same time as the blood drawing, and an elapsed time from a predetermined action of the subject to the blood drawing, and an estimation formula creation unit that creates an estimation formula for estimating the index from the pulse wave feature value and the elapsed time by grouping the index, the pulse wave feature value, and the elapsed time into a set of data and using a plurality of the group of data as learning data through machine learning. According to the above configuration, since the learning data including the elapsed time is used, a highly accurate estimation formula for estimating the index from the pulse wave can be created.

[0134] In an estimation formula creation system according to Aspect 11 of the present disclosure, in accordance with Aspect 10, the estimation formula creation unit creates the estimation formula using at least one set of data including the index, the pulse wave feature value, and the elapsed time based on measurements taken before the predetermined action, and at least one set of data including the index, the pulse wave feature value, and the elapsed time based on measurements taken after the predetermined action. With this configuration, the data taken before the predetermined action also serves as learning data, thereby improving the accuracy of estimation.

[0135] In the estimation formula creation system according to Aspect 12 of the present disclosure, in Aspects 10 or 11, the estimation formula creation unit creates the estimation formula using the group data of a subject to be estimated in addition to the group data of a plurality of people. With this configuration, the data of the subject to be estimated is also included in the training data, thereby improving the accuracy of the estimation.

[0136] The estimation formula creation system according to Aspect 13 of the present disclosure is the system of any of Aspects 10 to 12, further comprising a proposing unit that, when a predetermined condition is met, proposes adding the training data and having the estimation formula creation unit create the estimation formula. According to the configuration, when a predetermined condition is met, the system can propose additional training. An example of the predetermined condition is that a predetermined time has elapsed since the previous training.

[0137] In the index estimation system according to Aspect 14 of the present disclosure, in any one of Aspects 1 to 9, the estimation formula is created by machine learning. With the above configuration, it is possible to use the estimation formula created by machine learning.

[0138] An index estimation system according to Aspect 15 of the present disclosure is any one of Aspects 1 to 9, further including the estimation formula creation system. According to the above configuration, the estimation formula creation and estimation processing can be performed in the same system.

[0139] The index estimation system according to Aspect 16 of the present disclosure is any one of Aspects 1 to 9, further including a detection device that detects the pulse wave of the subject and outputs a pulse wave feature quantity indicating a feature of the detected pulse wave. According to the above configuration, the pulse wave of the subject can be detected by the same system.

[0140] A seventeenth aspect of the present disclosure provides an index estimation method including: an acquisition step of acquiring pulse wave feature quantities indicating characteristics of a subject's pulse wave and the elapsed time since a predetermined action; and an estimation step of estimating a first index related to the subject's blood glucose using the pulse wave feature quantities and the elapsed time, wherein the estimation step includes a set of data including a second index identical to or related to the first index, obtained by a different method from the method used to estimate the first index, the pulse wave feature quantities detected at the same time as the acquisition of the second index, and the elapsed time from the predetermined action to the acquisition of the second index, and the first index is estimated using an estimation formula created using a plurality of the set of data acquired at multiple times. This method allows for estimation of an index that takes into account the elapsed time since the predetermined action, thereby improving the accuracy of the index estimation.

[0141] An estimation formula creation method according to aspect 18 of the present disclosure includes: an acquisition step of acquiring an index related to a blood component obtained by drawing blood from at least one subject, a pulse wave feature value indicating characteristics of the subject's pulse wave detected at the same time as the blood drawing, and the elapsed time from a predetermined action of the subject to the blood drawing; and an estimation formula creation step of creating an estimation formula that estimates the index from the pulse wave feature value and the elapsed time by grouping the index, the pulse wave feature value, and the elapsed time into a set of data and using a plurality of the set of data as learning data through machine learning. According to the method, since the learning data including the elapsed time is used, a highly accurate estimation formula that estimates the index from the pulse wave can be created.

[0142] An index presentation system according to a nineteenth aspect of the present disclosure includes an estimation unit that estimates an index related to a blood component of a subject using a pulse wave feature amount indicating a feature of the pulse wave of the subject and an elapsed time since a predetermined action of the subject, and a presentation unit that presents the estimation result by the estimation unit to the subject.

[0143] According to Aspect 20 of the present disclosure, in the index presentation system of Aspect 19, the presentation unit presents the elapsed time together with the estimation result. With this configuration, the elapsed time can be presented to the subject in addition to the estimation result.

[0144] In an index estimation system according to Aspect 21 of the present disclosure, in any one of Aspects 1 to 9, the estimation formula used by the estimation unit includes a first estimation formula and a second estimation formula different from the first estimation formula, and the estimation unit performs the estimation using the first estimation formula when the elapsed time is within a first hour, and performs the estimation using the second estimation formula when the elapsed time exceeds the first hour. The inventors of the present application discovered that, regarding the relationship between a pulse wave feature value and a first index, the pulse wave feature value that affects the first index differs when the elapsed time exceeds a certain point. According to the above configuration, different estimation formulas are used to estimate the first index when the elapsed time is up to the first hour and when the elapsed time exceeds the first hour. This allows estimation to be performed using an estimation formula tailored to the pulse wave feature value that affects the first index. Therefore, the estimation result can be highly accurate regardless of whether the elapsed time is before or after the first hour.

[0145] According to Aspect 22 of the present disclosure, in the index estimation system of Aspect 21, the first estimation formula includes a plurality of parameters including a first parameter indicating the elasticity of blood vessels, and the weighting of the first parameter is higher than the weighting of other parameters. According to the above configuration, the first estimation formula highly weights the first parameter indicating the elasticity of blood vessels, and therefore, by using the first estimation formula when the correlation between the first index and the first parameter is high, the estimation result of the first index can be made highly accurate.

[0146] An index estimation system according to Aspect 23 of the present disclosure is the system of Aspects 21 or 22, wherein the second estimation formula includes a plurality of parameters including a first parameter indicating vascular elasticity, and the weighting of the first parameter is lower than the weighting of the other parameters. According to the configuration, the second estimation formula has a low weighting of the first parameter indicating vascular elasticity, and therefore, by using the second estimation formula when the correlation between the first index and the first parameter is low, the estimation result of the first index can be made highly accurate.

[0147] In the index estimation system according to Aspect 24 of the present disclosure, in any one of Aspects 21 to 23, the first time is the time when a person's blood glucose level peaks after a meal. The inventors of the present application discovered that the relationship between the pulse wave feature amount and the blood glucose level changes at the boundary of the blood glucose peak. With the above configuration, the first time can be set to the time when the blood glucose level peaks, so different estimation formulas can be used before and after the blood glucose level peak. This makes it possible to use an estimation formula that corresponds to the relationship between the pulse wave feature amount and the first index, thereby improving the accuracy of the estimation result of the first index.

[0148] In the index estimation system according to Aspect 25 of the present disclosure, in Aspect 24, the first time period is either (1) 60 minutes from the end of the meal when the predetermined behavior is eating, (2) a time set according to the meal content when the predetermined behavior is eating, (3) a time set by the subject as the time when the blood glucose level peaks, or (4) a time set according to the subject's attributes. According to the above configuration, a specific time period can be set as the first time period. The subject's attributes include gender, age, fasting blood glucose level, and HbA1c, an index related to blood glucose levels.

[0149] An index estimation system according to aspect 26 of the present disclosure, in any of aspects 21 to 25, includes an evaluation unit that evaluates the state of the subject's autonomic nervous system from the pulse wave feature amount, wherein the first estimation formula includes a 1-1 estimation formula and a 1-2 estimation formula, and the second estimation formula includes a 2-1 estimation formula and a 2-2 estimation formula. When the evaluation unit determines that the sympathetic nervous system is dominant, the estimation unit uses the 1-1 estimation formula or the 2-1 estimation formula to perform the estimation. When the evaluation unit determines that the parasympathetic nervous system is dominant, the estimation unit uses the 1-2 estimation formula or the 2-2 estimation formula to perform the estimation. The inventors of the present application discovered that the state of the vascular endothelium changes depending on the state of the autonomic nervous system, resulting in changes in the pulse wave feature amount. According to the above configuration, different estimation formulas can be used depending on the state of the autonomic nervous system, allowing for highly accurate estimation of the first index regardless of the state of the autonomic nervous system.

[0150] In the index estimation system according to Aspect 27 of the present disclosure, in Aspect 26, the pulse wave feature quantity is corrected to be lower in the first estimation formula and the second estimation formula. When the autonomic nervous system is in a sympathetic dominant state, blood vessels constrict, and the pulse wave feature quantity tends to take a high value. With this configuration, the pulse wave feature quantity is corrected to be lower in the first estimation formula and the second estimation formula, so that the first index can be accurately estimated in a sympathetic dominant state.

[0151] An index estimation system according to Aspect 28 of the present disclosure is the same as Aspects 26 or 27, wherein the first-second estimation formula and the second-second estimation formula are modified so that the pulse wave feature value is higher. When the autonomic nervous system is in a parasympathetic-dominant state, blood vessels dilate, and the pulse wave feature value tends to take a low value. With this configuration, the first-second estimation formula and the second-second estimation formula modify the pulse wave feature value higher, allowing the first index to be accurately estimated when the autonomic nervous system is in a parasympathetic-dominant state.

[0152] Aspect 29 of the present disclosure relates to an estimation formula creation system according to any one of aspects 10 to 13, wherein the estimation formula creation unit creates a first estimation formula and a second estimation formula different from the first estimation formula, the first estimation formula being created using the set of data in which the elapsed time is within a first hour, and the second estimation formula being created using the set of data in which the elapsed time exceeds the first hour. According to this configuration, different estimation formulas are created for when the elapsed time is up to the first hour and when it exceeds the first hour. This allows the estimation system to perform estimation using an estimation formula tailored to the pulse wave feature that affects the first index. Therefore, the estimation result can be highly accurate regardless of whether the time is before or after the first hour.

[0153] The estimation formula creation system according to aspect 30 of the present disclosure is the same as in aspect 29, except that the acquisition unit acquires an evaluation of the subject's autonomic nervous system, and the estimation formula creation unit creates a first estimation formula, a first estimation formula, and a second estimation formula, a second estimation formula. The first estimation formula and the second estimation formula are created when the subject's autonomic nervous system is evaluated as sympathetic dominant, and the first estimation formula and the second estimation formula are created when the subject's autonomic nervous system is evaluated as parasympathetic dominant. According to the above configuration, different estimation formulas are created depending on the state of the autonomic nervous system. This allows the estimation system to perform estimation using an estimation formula corresponding to the state of the autonomic nervous system. Therefore, the estimation results can be highly accurate regardless of the state of the autonomic nervous system.

[0154] In the index estimation method according to aspect 31 of the present disclosure, in aspect 17, the estimation formula used in the estimation step includes a first estimation formula and a second estimation formula different from the first estimation formula, and in the estimation step, if the elapsed time is within a first hour, the estimation is performed using the first estimation formula, and if the elapsed time exceeds the first hour, the estimation is performed using the second estimation formula.

[0155] In the estimation formula creation method according to aspect 32 of the present disclosure, in aspect 18, in the estimation formula creation step, a first estimation formula and a second estimation formula different from the first estimation formula are created as the estimation formulas, and the first estimation formula is created using the group data whose elapsed time is within a first hour, and the second estimation formula is created using the group data whose elapsed time exceeds the first hour.

[0156] The invention according to the present disclosure has been described above based on the drawings and examples. However, the invention according to the present disclosure is not limited to the above-described embodiments. In other words, the invention according to the present disclosure can be modified in various ways within the scope of the present disclosure, and embodiments obtained by appropriately combining the technical means disclosed in different embodiments are also included in the technical scope of the invention according to the present disclosure. In other words, it should be noted that a person skilled in the art can easily make various modifications or corrections based on the present disclosure. It should also be noted that these modifications or corrections are included in the scope of the present disclosure.

[0157] 1, 1A, 1B, 1C, 1D, 1E, 1F Index estimation system 10, 10A, 10B, 10C Estimation device 101 First acquisition unit 102 Estimation unit 103 First output unit 104 Memory unit 105 Determination unit 106 Proposal unit 20 Detection device 201 Drive unit 202 Light source unit 203 Light receiving unit 204 Conversion unit 30 Estimation formula creation device 301 Second acquisition unit 302 Estimation formula creation unit 303 Second output unit 40 Presentation device 401 Third acquisition unit 402 Presentation unit 50 Learning data

Claims

1. An index estimation system having an estimation unit that estimates a first index related to blood components of a subject using pulse wave feature quantities that indicate characteristics of the subject's pulse wave and the elapsed time since the subject performed a predetermined action prior to acquiring the pulse wave, wherein the estimation unit estimates the first index using an estimation formula.

2. The index estimation system of claim 1, wherein the estimation unit estimates the first index using an estimation formula created using, as a set of data, a second index that is the same as or related to the first index and is obtained by a method different from the method used to estimate the first index, the pulse wave feature detected at the same time as the acquisition of the second index, and the elapsed time from the specified behavior to the acquisition of the second index.

3. The index estimation system according to claim 2, wherein the estimation unit uses the estimation formula created using a plurality of the group data acquired at a plurality of points in time.

4. An index estimation system according to any one of claims 1 to 3, comprising a determination unit that determines whether the subject has performed the specified behavior using input from the subject or the subject's behavioral history, and wherein the estimation unit performs the estimation using the elapsed time from the time when the determination unit determined that the subject performed the specified behavior as the elapsed time from the specified behavior.

5. An index estimation system as described in claim 4, wherein the subject's behavioral history includes at least one of the subject's hand movements, pulse rate changes, body temperature changes, and pulse wave feature changes.

6. An index estimation system as described in claim 4 or 5, wherein, when the determination unit determines that the subject has performed the specified behavior, the estimation unit performs the estimation using the subject's pulse wave feature values ​​at a time when a specified time has elapsed since the time of the specified behavior.

7. An index estimation system according to any one of claims 1 to 6, wherein the predetermined behavior is the most recent behavior in which the fluctuation amount per unit time of the blood components of the subject exceeds a threshold value.

8. An index estimation system according to any one of claims 1 to 7, wherein the pulse wave feature amount is at least one of: the ratio of the amplitude of the reflected wave to the maximum amplitude of the pulse wave; the time from a predetermined time point determined based on the pulse wave until the reflected wave appears; the ratio of the amplitude of the reflected wave to the amplitude at a predetermined time point determined based on the pulse wave; the pulse rate within a predetermined time; the shape of the rising edge of the pulse wave; the area of ​​the AC component of the pulse wave; the ratio of the AC component to the DC component of the pulse wave; the amount of change in the ratio of the amplitude of the reflected wave to the maximum amplitude of the pulse wave before and after a predetermined action; and the amount of change in the ratio of the amplitude of the reflected wave to the amplitude at a predetermined time point determined based on the pulse wave before and after a predetermined action.

9. An index estimation system as described in claim 2 or 3, comprising: an acquisition unit that acquires the second index; and a proposal unit that suggests recreating the estimation formula when there is a difference between the second index and the first index estimated by the estimation unit that is equal to or greater than a threshold value.

10. An estimation formula creation system comprising: an acquisition unit that acquires an index related to blood components obtained by drawing blood from at least one subject, a pulse wave feature that indicates the characteristics of the subject's pulse wave detected at the same time as the blood is drawn, and the elapsed time from a specified action of the subject to the blood being drawn; and an estimation formula creation unit that creates an estimation formula that estimates the index from the pulse wave feature and the elapsed time by treating the index, the pulse wave feature, and the elapsed time as one set of data and using machine learning to learn multiple sets of data.

11. The estimation formula creation system described in claim 10, wherein the estimation formula creation unit creates the estimation formula using at least one set of data including the index, the pulse wave feature, and the elapsed time based on measurements before the specified behavior, and at least one set of data including the index, the pulse wave feature, and the elapsed time based on measurements after the specified behavior.

12. An estimation formula creation system as described in claim 10 or 11, wherein the estimation formula creation unit creates the estimation formula using the group data of a subject to be estimated in addition to the group data of multiple people.

13. An estimation formula creation system according to any one of claims 10 to 12, comprising a proposal unit that, when a predetermined condition is met, proposes adding the learning data and having the estimation formula creation unit create the estimation formula.

14. An index estimation system according to any one of claims 1 to 9, wherein the estimation formula is created by machine learning.

15. An index estimation system according to any one of claims 1 to 9, further comprising an estimation formula creation system according to claim 10.

16. An index estimation system according to any one of claims 1 to 9, further comprising a detection device that detects the pulse wave of the subject and outputs pulse wave feature quantities that indicate the characteristics of the detected pulse wave.

17. An index estimation method comprising: an acquisition step of acquiring pulse wave features indicating characteristics of the subject's pulse wave and the elapsed time from a predetermined action; and an estimation step of estimating a first index related to the subject's blood glucose using the pulse wave features and the elapsed time, wherein in the estimation step, a second index that is the same as or related to the first index obtained by a method different from the method used to estimate the first index, the pulse wave features detected at the same time as the acquisition of the second index, and the elapsed time from the predetermined action to the acquisition of the second index are treated as one set of data, and the first index is estimated using an estimation formula created using multiple sets of data acquired at multiple times.

18. A method for creating an estimation formula, comprising: an acquisition step of acquiring an index relating to blood components obtained by drawing blood from at least one subject, a pulse wave feature indicating characteristics of the subject's pulse wave detected at the same time as the blood drawing, and the elapsed time from a predetermined action of the subject to the blood drawing; and an estimation formula creation step of creating an estimation formula that estimates the index from the pulse wave feature and the elapsed time by treating the index, the pulse wave feature, and the elapsed time as one set of data and using machine learning to learn multiple sets of data.

19. An index presentation system comprising: an estimation unit that estimates an index related to blood components of a subject using pulse wave features that indicate the characteristics of the subject's pulse wave and the elapsed time since the subject performed a specified action; and a presentation unit that presents the estimation result by the estimation unit to the subject.

20. The index presentation system according to claim 19, wherein the presentation unit presents the elapsed time together with the estimation result.

21. An index estimation system as described in any one of claims 1 to 9, wherein the estimation formula used by the estimation unit includes a first estimation formula and a second estimation formula different from the first estimation formula, and the estimation unit performs the estimation using the first estimation formula when the elapsed time is within a first hour, and performs the estimation using the second estimation formula when the elapsed time exceeds the first hour.

22. An index estimation system as described in claim 21, wherein the first estimation equation includes a plurality of parameters including a first parameter indicating the elasticity of a blood vessel, and the weighting of the first parameter is higher than the weighting of other parameters.

23. An index estimation system as described in claim 21 or 22, wherein the second estimation formula includes a plurality of parameters including a first parameter indicating the elasticity of the blood vessel, and the weighting of the first parameter is lower than the weighting of the other parameters.

24. An index estimation system according to any one of claims 21 to 23, wherein the first time is the time when a person's blood glucose level peaks after a meal.

25. The index estimation system of claim 24, wherein the first time is one of the following: (1) 60 minutes from the end of a meal when the specified behavior is a meal; (2) a time set according to the content of the meal when the specified behavior is a meal; (3) a time set by the subject as the time when the blood glucose level peaks; or (4) a time set according to the attributes of the subject.

26. An index estimation system according to any one of claims 21 to 25, comprising an evaluation unit that evaluates the state of the autonomic nervous system of the subject from the pulse wave feature amount, wherein the first estimation formula includes a 1-1 estimation formula and a 1-2 estimation formula, and the second estimation formula includes a 2-1 estimation formula and a 2-2 estimation formula, and wherein the estimation unit, when the evaluation unit evaluates that the sympathetic nervous system is dominant, performs the estimation using the 1-1 estimation formula or the 2-1 estimation formula, and when the evaluation unit evaluates that the parasympathetic nervous system is dominant, performs the estimation using the 1-2 estimation formula or the 2-2 estimation formula.

27. An index estimation system according to claim 26, wherein the first-1 estimation formula and the second-1 estimation formula are obtained by correcting the pulse wave feature value to a lower value.

28. An index estimation system according to claim 26 or 27, wherein the first-second estimation formula and the second-second estimation formula are obtained by correcting the pulse wave feature value to a high value.

29. An estimation formula creation system as claimed in any one of claims 10 to 13, wherein the estimation formula creation unit creates a first estimation formula and a second estimation formula different from the first estimation formula as the estimation formula, the first estimation formula being created using the group data whose elapsed time is within a first hour, and the second estimation formula being created using the group data whose elapsed time exceeds the first hour.

30. The estimation formula creation system described in claim 29, wherein the acquisition unit acquires an evaluation of the autonomic nervous system of the subject, the estimation formula creation unit creates a 1-1 estimation formula and a 1-2 estimation formula as the first estimation formula, and creates a 2-1 estimation formula and a 2-2 estimation formula as the second estimation formula, the 1-1 estimation formula and the 2-1 estimation formula being created when the evaluation of the autonomic nervous system of the subject is sympathetic dominant, and the 1-2 estimation formula and the 2-2 estimation formula being created when the evaluation of the autonomic nervous system of the subject is parasympathetic dominant.

31. An index estimation method as described in claim 17, wherein the estimation formula used in the estimation step includes a first estimation formula and a second estimation formula different from the first estimation formula, and in the estimation step, if the elapsed time is within a first time, the estimation is performed using the first estimation formula, and if the elapsed time exceeds the first time, the estimation is performed using the second estimation formula.

32. The estimation formula creation method of claim 18, wherein in the estimation formula creation step, a first estimation formula and a second estimation formula different from the first estimation formula are created as the estimation formula, the first estimation formula being created using the group data whose elapsed time is within a first hour, and the second estimation formula being created using the group data whose elapsed time exceeds the first hour.

Images