Blood pressure measurement device, pressurization control method for blood pressure measurement device, and program

Abstract

A blood pressure measurement device for estimating a blood pressure value using pulse wave data, having: a cuff for compressing a part under measurement; a sensor that detects the pressure of the cuff; a pulse wave acquisition unit that acquires pulse wave data from an output signal from the sensor; an index estimation unit that, at a prescribed periodicity, performs index estimation processing to estimate an index on the basis of the pulse wave data acquired up to the present time while the cuff is gradually pressurized, said index expressing the position of the present time with respect to a predicted envelope that is predicted to be obtained if the pressurization of the cuff is continued; and a pressurization control unit that stops the pressurization of the cuff when the value of the index estimated by the index estimation processing reaches a predetermined threshold.

Description

Blood pressure measurement device, pressure control method for blood pressure measurement device, and program

[0001] The present invention relates to a blood pressure measurement device, and more particularly to cuff inflation control.

[0002] In recent years, health management technologies have become widespread, using measuring devices to measure information about an individual's body, such as blood pressure, and recording and analyzing the measurement results. One example of such measuring devices is a sphygmomanometer, which is configured to attach a cuff to a measurement site, such as the upper arm or wrist, measure a pressure pulse wave while inflating the attached cuff, and measure blood pressure, including systolic blood pressure (SBP), based on the measured pressure pulse wave. For example, Patent Document 1 proposes a method for completing blood pressure measurement before the cuff completely occludes the blood vessel by determining a pulse score from a profile of pressure pulse waves measured recently for multiple heartbeats, generating an envelope of the pressure pulse wave, and estimating blood pressure when the pulse score exceeds a threshold.

[0003] U.S. Pat. No. 9,750,419

[0004] The pulse score proposed in Patent Document 1 is designed to numerically represent the likelihood that the current cuff pressure exceeds the cuff pressure corresponding to the peak of the envelope. Specifically, a pulse score of 3 indicates a 100% likelihood that the current cuff pressure exceeds the cuff pressure corresponding to the peak of the envelope, and a pulse score of 2 indicates a 90% likelihood. The method of Patent Document 1 can be said to be an approach that sets "collection of pressure pulse waves until the peak of the envelope is exceeded" as a necessary condition for accurate blood pressure estimation and attempts to determine whether this necessary condition is met using the pulse score.

[0005] However, while studying blood pressure estimation algorithms, the inventors discovered that, in order to improve estimation accuracy, it is more important to clearly identify the position on the envelope of the collected pressure pulse wave than to collect the pressure pulse wave up to the peak of the envelope. When attempting to estimate blood pressure from an incomplete envelope, if prior knowledge is available as to whether the given envelope represents a stage on the way to the peak, a stage near the peak, or a stage beyond the peak, estimation accuracy can be improved by, for example, switching the blood pressure estimation algorithm or tuning parameters according to the stage of the envelope. In other words, even if only the pressure pulse wave up to the envelope peak has been collected, accurate blood pressure estimation can be expected as long as it is guaranteed that the data represents the position just before the envelope peak.

[0006] From this perspective, it can be said that the pulse score of Patent Document 1 has the following problems.

[0007] First, using the pulse score does not clearly indicate the extent to which the collected pressure pulse wave is measured on the envelope. Even if the pulse score threshold is set to 3, this merely guarantees that the pressure pulse wave will be collected up to the peak of the envelope. For example, one subject may receive a pulse score of 3 at a cuff pressure slightly above the peak, while another subject may receive a pulse score of 3 at a pressure significantly above the peak and close to the systolic pressure. The resulting envelope may vary greatly from subject to subject. Furthermore, setting the pulse score threshold to 2 guarantees that the envelope peak will be exceeded with a 90% probability, but conversely, there is a 10% probability (i.e., once in 10 times) that the pressure pulse wave will only be measured up to a position just before the peak. Thus, the pulse score-based determination limits the accuracy and reliability of subsequent blood pressure estimation due to the indeterminate section of the collected pressure pulse wave (especially the extent to which the pressure pulse wave is measured on the envelope).

[0008] Second, when using the pulse score, it is not possible to control the measurement or pressure application to end before the peak of the envelope. This is because the measurement of the pressure pulse wave continues until the peak of the envelope is exceeded. (As mentioned above, there is a probability that the measurement may end before the peak, but this is an exception and it cannot be actively controlled.) Therefore, there are limitations to speeding up the measurement or reducing the pressure of the cuff when using the pulse score.

[0009] An object of the present invention is to provide a technique that enables measurement of a pressure pulse wave and cuff inflation to be stopped at a desired position on the envelope.

[0010] The present disclosure provides a blood pressure measurement device that estimates a blood pressure value using pulse wave data, the blood pressure measurement device having: a cuff for compressing a measurement target; a sensor that detects the pressure of the cuff; a pulse wave acquisition unit that acquires pulse wave data from an output signal of the sensor; an index estimation unit that executes, at a predetermined cycle, an index estimation process that estimates an index that represents a current position relative to a predicted envelope that is expected to be obtained if the cuff is continued to be inflated, based on pulse wave data acquired up to the present time while gradually inflating the cuff; and an inflation control unit that stops inflating the cuff when the value of the index estimated by the index estimation process reaches a predetermined threshold.

[0011] The blood pressure measurement device may further include a blood pressure estimation unit that estimates a blood pressure value using an envelope obtained from pulse wave data acquired until inflation of the cuff is stopped.

[0012] The index may be designed to take a first value at a position where the amplitude of the predicted envelope is at a maximum value, and to take a second value at a position where the amplitude of the predicted envelope is a predetermined ratio of the maximum value.

[0013] The predetermined ratio may be half of the maximum value.

[0014] The index may be designed to take a first value at a position on the predicted envelope corresponding to the lowest blood pressure and a second value at a position on the predicted envelope corresponding to the highest blood pressure.

[0015] The index estimation unit may acquire an envelope up to the present time and a pulse wave for the most recent heartbeat from the pulse wave data acquired up to the present time, and estimate the value of the index based on a feature of the envelope up to the present time and a feature of the pulse wave for the most recent heartbeat.

[0016] The index estimation unit may perform the index estimation process using a trained model that has been machine-learned to output the value of the index when the feature of the envelope up to the current time point and the feature of the pulse wave for the most recent heartbeat are given as input.

[0017] The index estimation unit may acquire an envelope up to the current time and pulse waves for the most recent multiple heartbeats from the pulse wave data acquired up to the current time, and estimate the value of the index based on features of the envelope up to the current time and features of the pulse waves for the most recent multiple heartbeats.

[0018] The index estimation unit may perform the index estimation process using a trained model that has been machine-learned to output the value of the index when the envelope feature up to the current time point and the pulse wave feature for the most recent multiple heartbeats are given as input.

[0019] The present disclosure provides an inflation control method for a blood pressure measurement device that estimates a blood pressure value using pulse wave data, the inflation control method including: executing, at a predetermined cycle, an index estimation process that estimates an index representing the current position relative to a predicted envelope that is expected to be obtained if the cuff is continued to be inflated, based on pulse wave data acquired up to the present time while gradually inflating the cuff; and stopping inflation of the cuff when the value of the index estimated by the index estimation process reaches a predetermined threshold value.

[0020] The present disclosure includes a program for causing a processor of a blood pressure measurement device that estimates a blood pressure value using pulse wave data to execute inflation control, including: executing, at a predetermined cycle, an index estimation process that estimates an index representing the current position relative to a predicted envelope that is expected to be obtained if the cuff is continued to be inflated, based on pulse wave data acquired up to the present time while gradually inflating the cuff; and stopping inflation of the cuff when the value of the index estimated by the index estimation process reaches a predetermined threshold.

[0021] The present invention may be understood as a blood pressure measurement device, a blood pressure monitor, a biological information acquisition device, etc., having at least a part of the above configuration. The present invention may also be understood as a pressurization control method or a blood pressure measurement method that includes at least a part of the above processing, or a program for realizing such a method or a recording medium on which the program is recorded. The above means and processing can be combined with each other as much as possible to constitute the present invention.

[0022] According to the present invention, it is possible to provide a technique that enables measurement of a pressure pulse wave and cuff inflation to be stopped at a desired position on the envelope.

[0023] FIG. 1 is a diagram for explaining the pressure level index P. FIG. 2 is a diagram conceptually explaining the technical meaning and effect of the pressure level index P. FIG. 3 is a diagram schematically showing an example of the hardware configuration of a blood pressure measurement device according to an embodiment. FIG. 4 is a block diagram showing an example of the functional configuration of the blood pressure measurement device 1 according to an embodiment. FIG. 5 is a flowchart showing the procedure of the overall processing of a blood pressure measurement method. FIG. 6 is a diagram schematically showing the flow of index estimation processing by an index estimation unit. FIG. 7A is a diagram for explaining design example 1 of the pressure level index P. FIG. 7B is a diagram for explaining design example 1 of the pressure level index P. FIG. 8 is a diagram for explaining design example 2 of the pressure level index P. FIG. 9 is a schematic diagram for explaining the learning processing of a machine learning model used in the index estimation processing.

[0024] In a blood pressure measurement device that estimates blood pressure values ​​using pressure pulse wave data, in order to achieve the goal of making it possible to stop pressure pulse wave measurement and cuff inflation at a desired position on the envelope, the proposed method introduces an "inflation level index P," which is an index that represents the position relative to the envelope.

[0025] Figure 1 illustrates the inflation level index P in the proposed method. A cuff is attached to the subject's part, and the cuff pressure is monitored by a sensor while gradually inflating the cuff. The upper part of Figure 1 shows the sensor output signal. The sampling rate of the sensor output signal is, for example, approximately 250 Hz. A pressure pulse wave signal at the part being measured is superimposed on the sensor output signal. The middle part of Figure 1 shows the pressure pulse wave signal extracted (separated) from the sensor output signal. A high-pass filter can be used, for example, to extract the pressure pulse wave signal. The envelope is the curve formed by connecting the peak values ​​of the pulse waveform (called "pulse") of each heartbeat in the pressure pulse wave signal. The lower part of Figure 1 shows the envelope obtained from the pressure pulse wave signal. Note that while the horizontal axis in Figure 1 represents time, because cuff pressure and time are correlated, the envelope may also be expressed in a Cartesian coordinate plane with cuff pressure as the horizontal axis and pulse amplitude as the vertical axis.

[0026] The proposed blood pressure measurement device estimates the value of an inflation level index P based on pulse wave data (encircled by a dashed line) acquired up to the current time T while gradually inflating the cuff. This index P is a dimensionless index designed to represent the position of time T relative to the predicted envelope curve expected to be obtained if the cuff continues to be inflated after time T. The index P represents, for example, a relative position based on the peak position (the position where the amplitude is maximum) of the predicted envelope curve. In the example of FIG. 1 , the index P is 0 at the peak position of the predicted envelope curve, and the value of the index P at each position is set according to the ratio of the maximum amplitude value to the peak position. The index P is negative before the peak and positive after the peak. For example, if the value of the index P is estimated to be −0.5 based on pulse wave data acquired up to the current time T, the sign (positive or negative) of the index P indicates that the current time T is just before the peak of the envelope curve, and the absolute value of the index P indicates the relative position (distance) to the peak position of the envelope curve.

[0027] FIG. 2 conceptually explains the technical meaning and effect of the pressure level index P. As shown on the left side of FIG. 2, the envelope curve (the horizontal axis represents cuff pressure or time) obtained by measurement varies in peak position and envelope width (spread) depending on the subject. Therefore, it is not easy to determine the position on the envelope curve from the current pulse amplitude value, etc. Introducing the pressure level index P has the effect of canceling out such variations in the envelope curve depending on the subject. In other words, as shown on the right side of FIG. 2, by scaling the envelope curve using the pressure level index P on the horizontal axis, the envelope curve is normalized in the cuff pressure direction (time direction), making it possible to indicate the same position on the envelope curve using the value of index P regardless of the subject.

[0028] By using this inflation level index P, it is possible to stop pressure pulse wave measurement and cuff inflation at a desired position on the envelope. For example, the process of estimating the inflation level index P may be performed at a predetermined interval while gradually inflating the cuff, and cuff inflation may be stopped when the value of index P reaches a predetermined threshold Pth. For example, in FIG. 1 , if the threshold Pth is set to −0.5, cuff inflation is stopped at time T, and blood pressure estimation is performed based on the pressure pulse wave signal data acquired up to that time T. In this case, the value of index P may be provided to the blood pressure estimation process as prior knowledge, and the blood pressure estimation algorithm or parameters used in the algorithm may be changed depending on the value of index P. For example, three types of algorithms or parameters may be prepared: when P<−0.3, when −0.3≦P≦0.3, and when 0.3<P, and the algorithm or parameters used may be changed depending on the value of index P at the time inflation was stopped. In this way, high-accuracy and reliable blood pressure estimation can be expected regardless of whether the envelope before the peak is reached, the envelope including the peak, or the envelope beyond the peak is used.

[0029] 3 schematically shows an example of the hardware configuration of the blood pressure measurement device 1 according to this embodiment. The blood pressure measurement device 1 according to this embodiment includes a control unit 10, a storage unit 11, an oscillator circuit 121, a pump drive circuit 122, a valve drive circuit 123, a cuff 130, a pressure sensor 131, a pressure pump 132, an exhaust valve 133, an air tube 134, an operation switch 14, a display unit 15, and a power supply 16. The control unit 10 and the storage unit 11 are an example of a computer part of the blood pressure measurement device 1.

[0030] The cuff 130 includes an air bag 1301 containing air. The cuff 130 is provided with a pressure sensor 131, a pressure pump 132, and an exhaust valve 133 via an air tube 134. The pressure sensor 131 is configured to detect the pressure (cuff pressure) inside the air bag 1301 of the cuff 130. The pressure pump 132 is configured to supply air into the air bag 1301. The exhaust valve 133 is provided as a boundary between the internal space and the external space of the air bag 1301 and is configured to be openable and closable. By closing the exhaust valve 133, air is trapped inside the air bag 1301, and the pressure inside the air bag 1301 can be maintained. On the other hand, by opening the exhaust valve 133, the air inside the air bag 1301 can be discharged, thereby reducing the pressure.

[0031] The control unit 10 includes a hardware processor such as a CPU, and is configured to execute information processing based on programs and various data. The control unit 10 (CPU) is an example of a processor resource of the estimation device. The oscillation circuit 121 outputs a signal having an oscillation frequency corresponding to the output value of the pressure sensor 131 to the control unit 10. In this embodiment, the control unit 10 processes the signal from the oscillation circuit 121 to obtain cuff pressure and pulse wave data. The pump drive circuit 122 is configured to control the drive of the pressure pump 132 based on a control signal output from the control unit 10. The valve drive circuit 123 is configured to control the opening and closing of the exhaust valve 133 based on a control signal output from the control unit 10.

[0032] The storage unit 11 may be configured, for example, as a semiconductor memory. The storage unit 11 is an example of a memory resource of the estimation device. In this embodiment, the storage unit 11 stores various information such as a program 110 and model data 111. The program 110 includes a program for causing the CPU of the blood pressure measurement device 1 to execute information processing (described later), such as cuff inflation control processing, pressure pulse wave acquisition processing, inflation level index P estimation processing, blood pressure value estimation processing, and measurement result display processing. The program 110 includes a series of instructions for the information processing. Note that each processing may be configured with a different program module, or all processing may be executed by a single program. The model data 111 is a trained model. As long as the trained model can be reproduced when estimating the index P or blood pressure, the configuration of the model data 111 is not particularly limited and may be determined appropriately depending on the embodiment. The model data 111 may be training result data. The model data 111 may be incorporated into the program 110. The memory unit 11 may also store other information such as blood pressure measurement results (cuff pressure, pulse wave data, value of index P, estimated blood pressure, etc.) as appropriate.

[0033] The operation switch 14 is used to perform operations such as starting blood pressure measurement. The operation switch 14 may be configured as at least one of a physical switch and a virtual switch. The display unit 15 is configured to display various information such as blood pressure measurement results. The operation switch 14 and the display unit 15 may be integrated into a touch panel display. The power supply 16 is configured to supply power to each unit such as the control unit 10.

[0034] Note that, with regard to the specific hardware configuration of the blood pressure measurement device 1, components can be omitted, replaced, or added as appropriate depending on the embodiment. For example, the control unit 10 may include multiple hardware processors. The hardware processor may be configured with a microprocessor, FPGA, DSP, ASIC, etc. The blood pressure measurement device 1 may be equipped with a communication interface. For example, the blood pressure measurement device 1 may be configured with a near-field communication module to enable data communication with a user terminal such as a smartphone. As a result, the blood pressure measurement device 1 may acquire arbitrary data via the user terminal. At least one of the program 110 and the model data 111 may be stored in a storage medium of an external computer, such as an NAS. The blood pressure measurement device 1 may acquire at least one of the program 110 and the model data 111 from an external computer. The blood pressure measurement device 1 may acquire data directly from an external computer or indirectly from an external computer via a user terminal. The blood pressure measurement device 1 may be configured as a general blood pressure monitor or as a wearable device such as a wristwatch.

[0035] FIG. 4 is a block diagram showing an example of the functional configuration of the blood pressure measurement device 1 according to this embodiment.

[0036] The blood pressure measurement device 1 has, as its main functions related to blood pressure measurement, a pressure detection unit 40, an index estimation unit 43, a pressurization control unit 44, and a blood pressure estimation unit 45. These functions are realized by the CPU of the control unit 10 reading and executing a program stored in the storage unit 11.

[0037] A signal from the oscillator circuit 121 is input to the pressure detection unit 40. The pressure detection unit 40 detects the oscillation frequency of the signal input from the oscillator circuit 121 and converts the detected oscillation frequency into a pressure value signal. The pressure detection unit 40 includes an HPF (High Pass Filter) unit 41 that extracts and outputs a pressure pulse wave signal by HPF (High Pass Filter) processing the pressure value signal, and an LPF unit 42 that extracts and outputs a cuff pressure signal by LPF (Low Pass Filter) processing the pressure value signal. The time series data of the pressure pulse wave output from the HPF unit 41 and the time series data of the cuff pressure output from the LPF unit 42 are stored in the memory of the control unit 10. That is, the HPF unit 41 functions as a pulse wave acquisition unit that acquires pulse wave data from the output signal of the pressure sensor 131, and the LPF unit 42 functions as a cuff pressure acquisition unit that acquires cuff pressure data from the output signal of the pressure sensor 131. The sampling rate of the pulse wave and the cuff pressure is, for example, approximately 250 Hz.

[0038] The index estimation unit 43 has a function of executing an index estimation process to estimate an inflation level index P based on time-series data of pressure pulse waves acquired up to the present time while gradually inflating the cuff 130. The index estimation unit 43 repeatedly executes the index estimation process at a predetermined cycle once measurement (inflation of the cuff 130) begins. The execution cycle of the index estimation process may be set arbitrarily, but since it is preferable to estimate and evaluate the index each time a new pulse is obtained, in this embodiment the execution cycle is set to be approximately the same as or slightly faster than the average cardiac cycle (for example, approximately 1 Hz to several Hz).

[0039] The inflation control unit 44 has a function of inflating and deflating the cuff 130 by sending control signals to the pump drive circuit 122 and the valve drive circuit 123 and controlling the operation of the pressure pump 132 and the exhaust valve 133 .

[0040] The blood pressure estimation unit 45 has a function of reading time-series data of the pulse wave from the memory and estimating the blood pressure value using the data. The estimated blood pressure value may include the systolic blood pressure (SBP) and the diastolic blood pressure (DBP).

[0041] In this embodiment, the functions shown in Fig. 4 are realized by a software program, but all or part of these functions may be replaced by circuits such as ASICs, FPGAs, etc. Alternatively, the functions shown in Fig. 4 may be realized in cooperation with other computers (such as a cloud server, a user's PC or smartphone) connected via a network.

[0042] <Operation of Blood Pressure Measurement Device> Figure 5 is a flowchart showing the overall processing procedure of the blood pressure measurement method by the blood pressure measurement device 1. The processing shown in Figure 5 is realized by the CPU of the control unit 10 reading and executing a program 110 from the storage unit 11. Note that all or part of the program 110 may be read into the blood pressure measurement device 1 from a computer-readable external storage medium (such as a flash memory) or may be obtained via a network.

[0043] When measuring blood pressure, the subject wraps the cuff 130 around the part to be measured in advance. In the following description, an example is given in which the part to be measured is the upper arm, but the part to be measured is not limited to this and may be the wrist, etc. The description also assumes that the subject performs predetermined settings using the operation switch 14 and issues a command to start blood pressure measurement. Upon receiving the command to start blood pressure measurement, the blood pressure measurement device 1 performs predetermined initialization, such as opening the exhaust valve 133 and setting the cuff pressure to atmospheric pressure (initial pressure).

[0044] When blood pressure measurement is started, the inflation control unit 44 starts inflation control to inflate the cuff 130 (step S500). In the inflation control, the cuff pressure is increased at a constant rate.

[0045] While gradually increasing the cuff pressure, the pressure detection unit 40 acquires the pressure pulse wave and the cuff pressure (step S501). The pressure detection unit 40 monitors the cuff pressure and waits until the cuff pressure reaches a predetermined lower limit (step S502). When the cuff pressure reaches the lower limit and the cuff 130 begins to compress the measurement area with a certain strength, a reliable pulse wave can be measured, and the process proceeds to step S503.

[0046] Steps S503 to S505 are index estimation processing by the index estimation unit 43. An example of the index estimation processing will be described with reference to Fig. 6. Fig. 6 is a diagram schematically showing the flow of the index estimation processing by the index estimation unit 43.

[0047] In step S503, the index estimation unit 43 acquires an envelope of the pulse wave up to the present time based on the pulse wave data for multiple heartbeats acquired up to the present time. A known method may be used to generate the envelope. For example, the envelope may be a simple connection of the pulse peak positions of each heartbeat, or the envelope may be obtained by fitting a curve to a sequence of the pulse peak positions of each heartbeat. Furthermore, although not shown, the index estimation unit 43 may perform preprocessing, such as noise removal, on the pulse wave data before acquiring the envelope.

[0048] In step S504, the index estimation unit 43 extracts "envelope feature quantities" using the envelope acquired in step S503, and also extracts "pulse feature quantities" using the pulse (pulse wave) for the most recent heartbeat. Examples of envelope feature quantities include the maximum amplitude of the envelope, the full width at half maximum of the envelope, the peak cuff pressure of the envelope (the cuff pressure at which the amplitude of the envelope is maximized), the area of ​​the envelope, the standard deviation of the amplitude of the envelope, the skewness of the envelope, the kurtosis of the envelope, and parameters obtained when a predetermined model (curve) is fitted to the envelope (e.g., parameters of a Gaussian function fitted to the envelope). Examples of pulse feature quantities include the maximum amplitude of the pulse, the full width at half maximum of the pulse, the peak cuff pressure of the pulse, the area of ​​the pulse, the standard deviation of the amplitude of the pulse, the skewness of the pulse, and the kurtosis of the pulse. Note that the feature quantities listed here are merely examples, and other feature quantities may also be used.

[0049] In step S505, the index estimation unit 43 estimates the value of the pressure level index P based on the envelope feature values ​​up to the current time point and the pulse feature values ​​for the most recent heartbeat. In this embodiment, as shown in Fig. 6 , a trained model that has been machine-learned to output the value of the index P when the envelope feature values ​​and the pulse feature values ​​for one heartbeat are given as input is used in the index estimation process. As the machine learning algorithm, any method may be used, such as multiple regression, random forest, neural network, support vector regression, Lasso regression, Ridge regression, or Naive Bayes (Gaussian distribution).

[0050] When the value of the index P (i.e., the current position on the envelope) is estimated by the index estimation process of steps S503 to S505, the inflation control unit 44 compares the value of the index P with a predetermined threshold value Pth (step S506). The threshold value Pth may be preset in the storage unit 11 of the blood pressure measurement device 1, or may be changeable by the user.

[0051] If the value of index P has reached the threshold value Pth (Y in step S506), the inflation control unit 44 stops inflation of the cuff 130 at that point (step S507). Note that even if the value of index P has not reached the threshold value Pth (N in step S506), if the cuff pressure has reached a predetermined upper limit value (Y in step S508), it is assumed that some abnormality may have occurred, and inflation of the cuff 130 is stopped. As long as the cuff pressure is below the upper limit value (N in step S508), the index estimation process is repeatedly executed at a predetermined cycle until the value of index P reaches the threshold value Pth.

[0052] In step S509, the blood pressure estimation unit 45 estimates blood pressure values ​​(SBP, DBP) using information such as an envelope obtained from pulse wave data acquired up until the end of inflation of the cuff 130. The measurement results (blood pressure value, heart rate, etc.) obtained by the above measurement process are displayed on the display unit 15 (step S510).

[0053] 7A and 7B show design example 1 of the pressure level index P. In design example 1, the envelope is modeled using a Gaussian function, and the parameters of the Gaussian function are used to define the relational expression between the index P and the cuff pressure.

[0054] First, pressure pulse wave data is prepared. At this time, as shown in FIG. 7A , pressure pulse wave data from a cuff pressure lower than the diastolic blood pressure (DBP) to a cuff pressure higher than the systolic blood pressure (SBP) is used. Next, an envelope is obtained from the pressure pulse wave. For example, the envelope may be obtained by connecting the peak points of the pulses of each heartbeat. A Gaussian function expressed by the following equation (1) is then fitted to the envelope. The variable x in equation (1) is the cuff pressure, and A, B, and C are parameters that define the shape of the Gaussian function. The inflation level index P is defined as shown in equation (2) below, using parameters B and C of the Gaussian function obtained by fitting. This equation (2) defines the relationship (conversion) between the index P and the cuff pressure x.

[0055]

[0056] 7B shows the relationship between the value of index P and the amplitude of the envelope curve according to equation (2). The index P is 0 (first value) at the peak position where the amplitude of the envelope curve is at its maximum, and the absolute value of index P is (2 log 2) at the position where the amplitude of the envelope curve is half of the maximum value. １／２ (second value), where log is the natural logarithm.

[0057] Statistically, it is known that the position where the amplitude of the envelope is half of the maximum value roughly corresponds to DBP and SBP. Therefore, when the index P of design example 1 is used, P = -(2 log 2) １／２ indicates the approximate location of the DBP, and P = (2 log 2) １／２ The position of the pressure stop threshold Pth can be determined based on this knowledge.

[0058] Next, as another example, a description will be given of design example 2 of the inflation level index P. In design example 2, the relational expression between the index P and the cuff pressure is defined by normalizing the envelope in the section between the minimum blood pressure and the maximum blood pressure.

[0059] Specifically, the true value of the systolic blood pressure (SBP) ｇｔ and the true value of the lowest blood pressure (DBP) ｇｔ Using the above, the relationship (conversion) between the index P and the cuff pressure x can be defined as in the following formula (3): In the case of design example 2, since there is no need to calculate an envelope or fit a formula to the envelope as in design example 1, the relational formula for the index P can be obtained by simple processing.

[0060]

[0061] 8 shows the relationship between the index P and the cuff pressure x according to equation (3). That is, the index P is −1 (first value) at a position on the envelope corresponding to the diastolic blood pressure (DBP), and the index P is +1 (second value) at a position on the envelope corresponding to the systolic blood pressure (SBP). Furthermore, by setting the envelope to peak midway between the diastolic and systolic blood pressures, the index P becomes 0 at the peak of the envelope. Based on this knowledge, a desired inflation stop threshold value Pth can be determined.

[0062] <Learning of pressurization level index P> Figure 9 is a schematic diagram for explaining the learning process of the machine learning model used in the index estimation process. This learning process can be executed by the control unit 10 of the blood pressure measurement device 1, but is typically executed by a learning device prepared separately from the blood pressure measurement device 1. The learning device is configured by, for example, a general-purpose computer equipped with a CPU, GPU, memory, storage, etc., and the learning process described below is realized by expanding a learning program stored in the storage into the memory and executing the program by the CPU and GPU.

[0063] First, the learning device acquires pressure-pulse wave data for learning as training data (step S90). At this time, it is preferable to prepare a large number of pressure-pulse wave data obtained from a large number of subjects.

[0064] The learning device first selects one piece of pressure pulse wave data as a target and determines the relational expression between the index P and the cuff pressure for that target pressure pulse wave data (step S91). When the index P of design example 1 is adopted, relational expression (2) is obtained, and when the index P of design example 2 is adopted, relational expression (3) is obtained.

[0065] The learning device extracts the target pressure pulse wave data at an arbitrary cuff pressure (step S92). This cuff pressure is hereinafter referred to as the "cut cuff pressure." This extraction process simulates a state in which measurements have been taken up to the cut cuff pressure.

[0066] The learning device extracts envelope features and pulse features from the pressure pulse wave data extracted in step S92 (step S93), and inputs these features into the machine learning model being trained to obtain an estimate of the index P (step S94).

[0067] On the other hand, the learning device calculates the true value of the index P by substituting the cut cuff pressure into the relational expression obtained in step S91 (step S95). The learning device calculates the difference between the estimated value of the index P obtained in step S94 and the true value (step S96), and corrects the machine learning model using the difference (step S97).

[0068] By repeating the processing of steps S92 to S97 while changing the cut cuff pressure, it is possible to learn the correlation between the inflation level index P and the position on the envelope in the target pressure pulse wave data. By repeating this process for multiple given pressure pulse wave data, it is possible to obtain a trained model that can be used for general purposes.

[0069] <Others> The above-described embodiments merely exemplify exemplary configurations of the present invention. The present invention is not limited to the specific embodiments described above, and various modifications are possible within the scope of the technical concept thereof. In the above-described embodiments, the envelope feature and the pulse feature for one heartbeat are used in the index inference process. However, other feature may also be used in the index inference process. For example, the envelope feature and the pulse feature for the most recent several heartbeats may be used in the index inference process. Alternatively, only the envelope feature may be used in the index inference process, or the pulse feature for the most recent several heartbeats may be used in the index inference process. Furthermore, the above-described design examples 1 and 2 are merely examples, and any definition may be used as long as the relationship (conversion) between the inflation level index P and the cuff pressure can be specified.

[0070] 1: Blood pressure measuring device

Claims

1. A blood pressure measurement device that estimates a blood pressure value using pulse wave data, comprising: a cuff for compressing a part to be measured; a sensor that detects the pressure of the cuff; a pulse wave acquisition unit that acquires pulse wave data from an output signal of the sensor; an index estimation unit that executes, at a predetermined cycle, an index estimation process that estimates an index that represents a current position relative to a predicted envelope that is expected to be obtained if the cuff is continued to be inflated, based on pulse wave data acquired up to the present time while gradually inflating the cuff; and an inflation control unit that stops inflating the cuff when the value of the index estimated by the index estimation process reaches a predetermined threshold.

2. The blood pressure measurement device according to claim 1, further comprising a blood pressure estimation unit that estimates a blood pressure value using an envelope obtained from pulse wave data acquired up until inflation of the cuff is stopped.

3. A blood pressure measurement device as described in claim 1 or 2, wherein the index is designed to take a first value at a position where the amplitude of the predicted envelope is at its maximum value, and to take a second value at a position where the amplitude of the predicted envelope is at a predetermined ratio to the maximum value.

4. The blood pressure measuring device according to claim 3, wherein the predetermined ratio is 1 / 2 of the maximum value.

5. A blood pressure measurement device as described in claim 1 or 2, wherein the index is designed to take a first value at a position corresponding to the lowest blood pressure in the predicted envelope and to take a second value at a position corresponding to the highest blood pressure in the predicted envelope.

6. A blood pressure measurement device according to any one of claims 1 to 5, wherein the index estimation unit acquires an envelope up to the present time and a pulse wave for the most recent heartbeat from the pulse wave data acquired up to the present time, and estimates the value of the index based on the feature of the envelope up to the present time and the feature of the pulse wave for the most recent heartbeat.

7. The blood pressure measurement device according to claim 6, wherein the index estimation unit executes the index estimation process using a trained model that has been machine-learned to output the value of the index when the envelope feature up to the current time point and the pulse wave feature for the most recent heartbeat are given as input.

8. A blood pressure measurement device according to any one of claims 1 to 5, wherein the index estimation unit acquires an envelope up to the present time and pulse waves for the most recent multiple heartbeats from the pulse wave data acquired up to the present time, and estimates the value of the index based on features of the envelope up to the present time and features of the pulse waves for the most recent multiple heartbeats.

9. The blood pressure measurement device according to claim 8, wherein the index estimation unit executes the index estimation process using a trained model that has been machine-learned to output the value of the index when the envelope feature up to the current time point and the pulse wave feature for the most recent multiple heartbeats are given as input.

10. A method for controlling inflation of a blood pressure measuring device that estimates a blood pressure value using pulse wave data, comprising: executing, at a predetermined cycle, an index estimation process that estimates an index representing the current position relative to a predicted envelope that is expected to be obtained if the cuff is continued to be inflated, based on pulse wave data acquired up to the present time while gradually inflating the cuff; and stopping inflation of the cuff when the value of the index estimated by the index estimation process reaches a predetermined threshold.

11. A program for causing a processor of a blood pressure measuring device that estimates a blood pressure value using pulse wave data to execute inflation control, including: executing, at a predetermined cycle, an index estimation process that estimates an index representing the current position relative to a predicted envelope that is expected to be obtained if the cuff is continued to be inflated, based on pulse wave data acquired up to the present time while gradually inflating the cuff; and stopping cuff inflation when the value of the index estimated by the index estimation process reaches a predetermined threshold.

Images