Sphygmomanometer, blood pressure measurement method, blood pressure measurement program, learning model construction method, and learning model construction program
The blood pressure monitor employs a machine learning correction model using cuff pressure time-series and waveform data to improve measurement accuracy, addressing noise issues and data limitations, enabling precise blood pressure readings with reduced training data.
Patent Information
- Application Number
- PCT/JP2024/016716
- Authority / Receiving Office
- WO · WO
- Patent Type
- Applications
- Current Assignee / Owner
- Priority Date
- 2023-05-12
- Filing Date
- 2024-04-30
- Publication Date
- 2024-11-21
AI Technical Summary
The oscillometric method for blood pressure measurement is prone to accuracy reduction due to noise from body movements and cuff positioning, making it challenging to achieve precise measurements, especially with limited training data.
A blood pressure monitor that uses a machine learning-based correction model, inputting time-series changes in cuff pressure and OscilloMetric Waveform (OMW) and OscilloMetric Waveform Envelope (OMWE) information to correct blood pressure measurements, reducing the range of the objective variable and thereby requiring fewer training data points.
This approach enhances the accuracy of blood pressure measurement by correcting for noise and reducing the need for extensive training data, allowing for precise readings even with a small amount of data, and improves measurement accuracy compared to traditional methods.
Smart Images

Figure JP2024016716_21112024_PF_FP_ABST
Abstract
Description
Sphygmomanometer, blood pressure measurement method, blood pressure measurement program, learning model construction method, and learning model construction program
[0001] The present invention relates to a sphygmomanometer, a blood pressure measurement method, a blood pressure measurement program, a learning model construction method, and a learning model construction program.
[0002] Health conditions are known to be assessed by measuring blood pressure. Patent Document 1 describes a technique for storing a correlation between a subject's blood pressure measured using a cuff and that measured by auscultation, and correcting the subject's blood pressure using the stored correlation. Patent Document 2 describes a technique for measuring blood pressure using a photoplethysmogram.
[0003] JP-A-5-309073 Publication Special Publication No. 2022-516820
[0004] In blood pressure measurement using the oscillometric method, a cuff wrapped around the subject's upper arm or wrist is inflated and then gradually depressurized while detecting pressure oscillations superimposed on the cuff's internal pressure, and blood pressure is measured based on the detected pressure oscillations. However, noise may be mixed into the detected pressure oscillations due to the subject's body movements, the way the cuff is wrapped, the position of the cuff, and other factors. Such noise contamination may reduce the accuracy of blood pressure measurement using the oscillometric method. While machine learning could be used to mitigate the reduction in blood pressure measurement accuracy, it is not easy to collect a large amount of training data that can mitigate the effects of noise.
[0005] One aspect of the disclosed technology aims to provide a sphygmomanometer, a blood pressure measurement method, a blood pressure measurement program, a learning model construction method, and a learning model construction program that can achieve highly accurate blood pressure measurement even with a small amount of training data.
[0006] One aspect of the disclosed technology is exemplified by the following blood pressure monitor: The blood pressure monitor includes a cuff to be wrapped around the arm of a person to be measured, a control unit that controls the pressure applied to the arm by the cuff, a detection unit that detects the cuff pressure received by the cuff from the arm, a measurement unit that measures the blood pressure of the person to be measured by an oscillometric method using a time-series change in the cuff pressure, a correction model constructed for the subject by machine learning using waveform information regarding the time-series change in the cuff pressure as an explanatory variable and an error in the blood pressure measured by the measurement unit relative to a blood pressure adopted as a true value as a target variable, a correction unit that corrects the blood pressure measured by the measurement unit based on the error output by the correction model by inputting the waveform information regarding the time-series change in the cuff pressure detected by the detection unit into the correction model, and an output unit that outputs the blood pressure corrected by the correction unit.
[0007] According to the above-described sphygmomanometer, the blood pressure measured by the oscillometric method is corrected using the error, thereby improving the accuracy of blood pressure measurement. Furthermore, the above-described sphygmomanometer uses the error of the blood pressure measured by the oscillometric method relative to the blood pressure adopted as the true value as the objective variable. The range of the error is considered to be approximately ±10 mmHg, and the range of human systolic blood pressure is considered to be approximately 80 mmHg to 200 mmHg. In machine learning, the amount of training data tends to increase in proportion to the range of the objective variable. In other words, by narrowing the range of the objective variable, the amount of training data used for learning can be reduced. Since the above-described sphygmomanometer uses the error, which has a narrower range than systolic blood pressure, as the objective variable, the amount of training data used for learning can be reduced. In other words, this sphygmomanometer can achieve high-accuracy blood pressure measurement even with a small amount of training data. The blood pressure adopted as the true value may be blood pressure actually measured by auscultation.
[0008] Here, the information on the waveform may include, for example, information on the cuff pressure. The measurement unit may acquire an Oscillometric Waveform (OMW) based on a time-series change in the cuff pressure, and the information on the waveform may include OMW information indicating the waveform of the OMW.
[0009] The measurement unit may also acquire an Oscillometric Waveform Envelope (OMWE) based on the OMW, and the information on the waveform may include OMWE information indicating the waveform of the OMWE.
[0010] Here, the blood pressure monitor may correct the blood pressure measured by the measurement unit based on an error output by the correction model by inputting both the time-series change in the cuff pressure and the OMW information into the correction model.Furthermore, the blood pressure monitor may correct the blood pressure measured by the measurement unit based on an error output by the correction model by inputting the time-series change in the cuff pressure and the OMWE information into the correction model.
[0011] The disclosed technology can also be understood from the aspects of a blood pressure measurement method, a blood pressure measurement program, a learning model construction method, and a learning model construction program.
[0012] According to the disclosed technology, highly accurate blood pressure measurement can be achieved even with a small amount of training data.
[0013] FIG. 1 is a diagram illustrating an example of a sphygmomanometer according to the present embodiment. FIG. 2 is a diagram schematically illustrating blood pressure correction by a correction unit according to the embodiment. FIG. 3 is a diagram schematically illustrating construction of a correction model according to the embodiment. FIG. 4 is a diagram illustrating an example of a processing flow of the sphygmomanometer according to the present embodiment. FIG. 5 is a diagram illustrating verification results of blood pressure measurement accuracy. FIG. 6 is a diagram illustrating an example of a sphygmomanometer according to a first modified example.
[0014] <Application Example> An application example of the present invention will be described. A blood pressure monitor 100 according to the application example is a device that measures the blood pressure of a subject by wrapping a cuff 120 around the subject's arm. The cuff 120 is wrapped around, for example, the upper arm or wrist of the subject's arm. That is, the blood pressure monitor 100 may be a wrist-type blood pressure monitor. In the blood pressure monitor 100, the measurement unit 114 measures the blood pressure of the subject using an oscillometric method. The measurement unit 114 inputs the measured blood pressure of the subject to the correction unit 115.
[0015] The correction unit 115 corrects the blood pressure measured by the measurement unit 114 using the correction model 116. The correction model 116 is a learning model constructed by machine learning using waveform information regarding the time-series change in cuff pressure by the cuff 120 as an explanatory variable and the error of the blood pressure measured by the measurement unit 114 relative to the blood pressure adopted as the true value as a target variable. Examples of waveform information regarding the time-series change in cuff pressure include waveform information from an Oscillometric Waveform (OMW) and waveform information from an Oscillometric Waveform Envelope (OMWE). The measurement unit 114 measures blood pressure using, for example, an oscillometric method. Note that the blood pressure adopted as the true value may be blood pressure actually measured using, for example, auscultation. Furthermore, machine learning includes deep learning.
[0016] When waveform information relating to time-series changes in cuff pressure is input, the correction model 116 outputs the error of the blood pressure measured by the oscillometric method relative to the blood pressure adopted as the true value. The correction unit 115 corrects the blood pressure measured by the measurement unit 114 using the error output from the correction model 116. In the sphygmomanometer 100, such correction allows the blood pressure measured by the measurement unit 114 to approach the blood pressure actually measured by the auscultation method as closely as possible.
[0017] As described above, the correction model 116 in this application example uses the error of the blood pressure measured by the oscillometric method relative to the blood pressure adopted as the true value as the objective variable. The error is considered to be within the range of approximately ±10 mmHg. On the other hand, the systolic blood pressure of a person is within the range of approximately 80 mmHg to 200 mmHg.
[0018] That is, when blood pressure is used as the objective variable of the correction model 116, the range of variation of the objective variable is about 10 times larger than when the error is used as the objective variable. In machine learning, the amount of training data tends to increase in proportion to the range of variation of the objective variable. Therefore, in this application example, in which the error is used as the objective variable, the amount of training data used to construct the correction model 116 can be reduced compared to when blood pressure is used as the objective variable. That is, according to this application example, high-precision blood pressure measurement can be achieved even with a small amount of training data.
[0019] Next, an embodiment will be described. Fig. 1 is a diagram showing an example of a blood pressure monitor 100 according to this embodiment. The blood pressure monitor 100 includes a main body 110 and a cuff 120. The blood pressure monitor 100 measures the blood pressure of a subject by wrapping the cuff 120 around the subject's upper arm or wrist (hereinafter also referred to as the arm).
[0020] The cuff 120 is formed in a band shape and has a bag inside which air is supplied. When air is supplied to the bag inside the cuff 120 while the cuff 120 is wrapped around the arm of the subject, the pressure exerted by the cuff on the arm increases. When air is inhaled from the bag inside the cuff 120 while the cuff 120 is wrapped around the arm of the subject, the pressure exerted by the cuff on the arm decreases.
[0021] The main body 110 includes, for example, a processor and a storage unit, and the processor executes a program stored in the storage unit to implement various processing units, such as a start switch 111, a control unit 112, a pressure sensor 113, a measurement unit 114, a correction unit 115, a correction model 116, and an output unit 117. In other words, the sphygmomanometer 100 can be regarded as a computer having a cuff 120. Note that the various processing units, such as the start switch 111, the control unit 112, the pressure sensor 113, the measurement unit 114, the correction unit 115, the correction model 116, and the output unit 117, may be dedicated hardware circuits. When the start switch 111 is turned on, the sphygmomanometer 100 is started up and blood pressure measurement is initiated.
[0022] The control unit 112 controls the pressure applied to the arm of the subject by the cuff 120 by supplying air to the cuff 120 and inhaling air from the cuff 120. When the activation switch 111 is turned on, the control unit 112 starts controlling the pressure of the cuff 120.
[0023] The pressure sensor 113 is a sensor that detects the pressure (cuff pressure) caused by the cuff 120. The pressure sensor 113 outputs the detected cuff pressure to the measurement unit 114. The cuff pressure includes, for example, the pressure of the cuff 120 pressing against the arm and pressure vibrations caused by the pulsation of blood flowing through the artery in the arm.
[0024] The measurement unit 114 measures blood pressure using the oscillometric method based on the pressure detected by the pressure sensor 113. When the pressure from the cuff 120 becomes higher than the systolic blood pressure, the artery is blocked, and pressure vibrations due to blood flow pulsation no longer occur. On the other hand, when the pressure from the cuff 120 becomes lower than the systolic blood pressure, the artery begins to open, and pressure vibrations due to blood flow pulsation become superimposed on the pressure from the cuff 120. When the pressure from the cuff 120 further decreases and becomes lower than the diastolic blood pressure, the artery inside the arm moves from its open state to slightly expand, and pressure vibrations no longer occur.
[0025] The measurement unit 114 acquires, from the pressure sensor 113, a time-series change in the cuff pressure of the cuff 120, which is increased to a pressure higher than the systolic blood pressure and then gradually decreased to a pressure lower than the diastolic blood pressure. The measurement unit 114 extracts a time-series change in pressure vibration caused by the pulsation of blood flow in an artery from the time-series change in the cuff pressure acquired from the pressure sensor 113. The extracted time-series change in pressure vibration is the OMW. The measurement unit 114 further extracts the OMWE from the OMW. The measurement unit 114 measures the blood pressure of the subject using a feature extracted from the OMW (hereinafter also referred to as the OMW feature) and a feature extracted from the OMWE (hereinafter also referred to as the OMWE feature). Note that the measurement unit 114 may measure blood pressure using an oscillometric method using either the OMW feature or the OMWE feature. Furthermore, the measurement unit 114 outputs waveform information relating to time-series changes in the cuff pressure to the correction unit 115. Examples of the waveform information relating to time-series changes in the cuff pressure include OMW information indicating an OMW waveform and OMWE information indicating an OMWE waveform. In the following, an example will be described in which at least one of the OMW information and the OMWE information is used as the waveform information relating to time-series changes in the cuff pressure.
[0026] The OMW information and OMWE information output by the measurement unit 114 can include, for example, the following information (1) to (28). Each piece of information (1) to (14) can be considered information about the waveform shape of the OMW or OMWE. Furthermore, each piece of information (15) to (20) can be considered information about the cuff pressure. (1) Minimum value, average value, minimum value. (2) Value at X% point on the time axis (X is a number between 0 and 100). (3) Time at which the value is maximum. (4) Area of the region enclosed by the OMW or OMWE waveform and the time axis. (5) Kurtosis and skewness of the OMW or OMWE waveform. (6) Number of peaks in the OMW or OMWE waveform. (7) Complexity of the OMW or OMWE waveform (for example, the sum of squares of the differences before and after the time series data). (8) The number of points in the OMW or OMWE waveform that exceed the average value up to a predetermined time point (e.g., 1 / 3 of the way through from the start). (9) The gradient in each of the divided regions, when the OMW or OMWE waveform is divided into a predetermined number of regions (e.g., 4). (10) The value obtained by dividing the time before the peak of the OMW or OMWE waveform by the time after the peak. (11) The rate of change in the height of the OMW or OWME waveform. (12) The ratio of the area of the waveform to the circumscribing rectangle, when the OMW or OMWE waveform is divided into a predetermined number of regions (e.g., 4). (13) The value obtained by subtracting the peak time of the OMW lower envelope from the peak time of the OMW upper envelope. (14) The area ratio between the area of the OMW upper envelope and the area of the OMW lower envelope. (15) The cuff pressure at the OMWE peak. (16) The cuff pressure at the point where the first derivative of OMWE is maximum and minimum. (17) The cuff pressure at the point where the first derivative of OMWE is N times (N is an integer greater than 0 and less than 1) lower than the peak in the first half of the OMWE peak. (18) The cuff pressure at the point where the first derivative of OMWE is N times (N is an integer greater than 0 and less than 1) lower than the peak in the second half of the OMWE peak. (19) The time from the maximum value of the first derivative of OMWE to the peak. (20) The time from the peak of OMWE to the minimum value of the first derivative. (21) The time length of OMWE. (22) The value obtained by dividing the time to the peak of OMWE by the time length of OMWE. (23) The position of the mean "-σ" when OMWE is Gaussian fitted.(24) The position of the mean "+σ" when OMWE is Gaussian-fitted. (25) The value of OMWE at the position of the mean "-σ" when OMWE is Gaussian-fitted. (26) The value of OMWE at the position of the mean "+σ" when OMWE is Gaussian-fitted. (27) The maximum value of the gradient of OMWE. (28) The minimum value of the gradient of OMWE.
[0027] The correction unit 115 corrects the blood pressure measured by the measurement unit 114. The correction unit 115 inputs the OMW information and OMWE information extracted by the measurement unit 114 to the correction model 116. The correction model 116 is a learning model that, when the OMW information and OMWE information are input, outputs a correction value representing the error between the blood pressure measured by the measurement unit 114 and the blood pressure actually measured by auscultation. The correction model 116 is stored, for example, in a storage unit housed in the main body 110. The correction unit 115 corrects the blood pressure measured by the measurement unit 114 using the correction value acquired from the correction model 116.
[0028] FIG. 2 is a diagram schematically illustrating blood pressure correction by the correction unit 115 according to an embodiment. The correction of blood pressure by the correction unit 115 will be described with reference to FIG. 2. The measurement unit 114 extracts the OMW of the subject C1 (step M1). The measurement unit 114 extracts OMW information indicating the OMW waveform from the extracted OMW (step M2). The measurement unit 114 also extracts an OMWE from the OMW (step M3) and measures the blood pressure of the subject C1 based on the extracted OMWE using an oscillometric method (step M4). The measurement unit 114 then extracts OMWE information indicating the OMWE waveform from the OMWE (step M5).
[0029] The correction unit 115 inputs the OMW information extracted in step M2 and the OMWE information extracted in step M5 into the correction model 116, obtains the error between the blood pressure measured by the oscillometric method and the blood pressure actually measured by the auscultation method, and corrects the blood pressure measured in step M4 based on the obtained error (step M6). The output unit 117 outputs the blood pressure corrected in step M6 to a display or the like (step M7).
[0030] Here, we will briefly explain how to construct the correction model 116 used for correction by the correction unit 115. The correction model 116 is constructed by machine learning using OMW information and OMWE information as explanatory variables and the error between blood pressure measured by the measurement unit 114 and blood pressure actually measured by auscultation as a response variable.
[0031] FIG. 3 is a diagram schematically illustrating construction of a correction model 116 according to an embodiment. Construction of the correction model 116 will be described with reference to FIG. 3. To construct the correction model 116, multiple subjects (E1, E2, ..., EN) are prepared. Note that, although multiple subjects are prepared in this embodiment, the number of subjects may be one. The measurement unit 114 extracts the OMW of subject E1 (step P1). The measurement unit 114 extracts OMW information indicating the waveform of the OMW from the extracted OMW (step P2). The measurement unit 114 also extracts an OMWE from the OMW (step P3) and measures the blood pressure of subject E1 using an oscillometric method based on the extracted OMWE (step P4). The measurement unit 114 then extracts OMWE information indicating the waveform of the OMWE from the OMWE (step P5).
[0032] Doctor D1 also measures the blood pressure of subject E1 by auscultation (step P6). The error between the blood pressure measured in step P4 and the blood pressure measured in step P6 is calculated (step P7). Machine learning is performed using the OMW information extracted in step P2 and the OMWE information extracted in step P5 as explanatory variables and the error calculated in step P7 as a target variable (step P8). A correction model 116 is constructed by the machine learning in step P8.
[0033] The processes from steps P1 to P8 are also performed for each of subjects E2 to EN. In other words, the correction model 116 is not a learning model that improves the accuracy of blood pressure measurement for a single subject, but rather a model that improves the accuracy of blood pressure measurement by the oscillometric method by using measurement results for multiple subjects.
[0034] 1 , the output unit 117 outputs the blood pressure corrected by the correction model 116. The output unit 117 is, for example, a liquid crystal display (LCD), a plasma display panel (PDP), an inorganic electroluminescence (EL) panel, or an organic EL panel. The output unit 117 may also be a printer or a speaker.
[0035] <Processing Flow> Fig. 4 is a diagram showing an example of a processing flow of the sphygmomanometer 100 according to this embodiment. Hereinafter, an example of the processing flow of the sphygmomanometer 100 will be described with reference to Fig. 4 .
[0036] In step S1, activation switch 111 is turned on to activate sphygmomanometer 100. In step S2, controller 112 supplies air to cuff 120 to pressurize the subject's arm. Controller 112 inflates cuff 120 to a pressure higher than the systolic blood pressure, and then gradually reduces the pressure in cuff 120.
[0037] In step S3, the measurement unit 114 extracts an OMW based on the cuff pressure detected by the pressure sensor 113, and extracts OMW information from the extracted OMW. The measurement unit 114 also extracts an OMWE from the OMW, and extracts OMWE information. The measurement unit 114 measures the blood pressure of the subject by an oscillometric method using the OMWE.
[0038] In step S4, the measurement unit 114 determines whether or not the blood pressure measurement of the subject has been completed. The measurement unit 114 may determine that the blood pressure measurement is completed when, for example, both the systolic blood pressure and the diastolic blood pressure of the subject have been measured. If the blood pressure measurement is completed (YES in step S4), the process proceeds to step S5. If the blood pressure measurement is not completed (NO in step S4), the process proceeds to step S2.
[0039] In step S5, the control unit 112 deflates the cuff 120. In step S6, the correction unit 115 inputs the OMW information and OMWE information extracted in step S3 into the correction model 116 to obtain the error between the blood pressure measured by the oscillometric method and the blood pressure actually measured by the auscultation method. The correction unit 115 corrects the blood pressure measured in step S3 using the error obtained from the correction model 116.
[0040] In step S7, the output unit 117 outputs the blood pressure corrected in step S6.
[0041] <Verification> The blood pressure measurement accuracy of the sphygmomanometer 100 according to the embodiment described above was verified, and will now be described. FIG. 5 is a diagram illustrating verification results of blood pressure measurement accuracy. For comparison with the embodiment, FIG. 5 shows a comparison example in which blood pressures (diastolic blood pressure: DBP, systolic blood pressure: SBP) measured by the oscillometric method are not corrected by a correction model (no correction), and a comparison example in which a correction model is constructed using blood pressure as the objective variable is prepared. For each of the comparison example and the embodiment, a comparison example using OMWE information as input data to the correction model (Comparative Example 1, Embodiment 1) and a comparison example using both OMWE information and OME information (Comparative Example 2, Embodiment 2) were prepared.
[0042] In this verification, blood pressure was measured multiple times for each of multiple subjects to obtain measurement results. In this verification, the obtained measurement results were evaluated based on Standard Deviation of Error (SDE).
[0043] 5, "Embodiment 1" and "Embodiment 2" exhibited higher accuracy than "No correction," "Comparative Example 1," and "Comparative Example 2." Also, referring to FIG. 5, whether OMWE information or both OMW information and OMWE information were used as input data to the correction model, "Embodiment 1" and "Embodiment 2" exhibited higher accuracy than "No correction," "Comparative Example 1," and "Comparative Example 2."
[0044] <Effects of the embodiment> According to this embodiment, the blood pressure measured by the measurement unit 114 is corrected by the correction unit 115, thereby increasing the accuracy of blood pressure measurement as much as possible. Here, the correction model 116 uses the error between the blood pressure measured by the oscillometric method and the blood pressure actually measured by the auscultation method as the dependent variable. As explained in the application example, the range of the error is approximately 1 / 10 of the range of the blood pressure. Since the amount of training data tends to increase in proportion to the range of the dependent variable, this embodiment, in which the error is used as the dependent variable, can reduce the amount of training data used to construct the correction model 116 compared to when blood pressure is used as the dependent variable. In other words, this embodiment makes it possible to achieve more accurate blood pressure measurement with less training data.
[0045] Furthermore, in this embodiment, noise due to the subject's body movement, the way the cuff is wrapped, the position at which the cuff is wrapped, and so on is also included in the OMW information and OMWE information. Here, the correction model 116 is constructed using the OMW information and OMWE information as explanatory variables. Therefore, the correction model 116 can output the error between the blood pressure measured by the oscillometric method and the blood pressure actually measured by the auscultation method, taking such noise into account. In other words, according to this embodiment, blood pressure measurement by the oscillometric method can be performed with higher accuracy.
[0046] However, since blood pressure measurement using a cuff applies pressure to the human body, it is difficult to collect a large amount of blood pressure data measured by the oscillometric method, which is the basis for calculating the error. In this embodiment, by using the error as the objective variable, the range of variation of the objective variable can be narrowed, and the influence of bias in the objective variable can be suppressed. Therefore, even if bias occurs in the measured blood pressure data due to a small amount of blood pressure data measured by the oscillometric method, blood pressure measurement can be performed with higher accuracy.
[0047] <First Modification> In the embodiment described above, the correction model 116 is stored in a storage unit within the main body 110. However, the correction model 116 may be stored in a location other than a storage unit within the main body 110. FIG. 6 is a diagram showing an example of a sphygmomanometer 100A according to a first modification. The sphygmomanometer 100A according to the first modification differs from the sphygmomanometer 100 according to the embodiment in that the correction model 116 is stored not in a storage unit within the main body 110 but in a server 200 connected to the main body 110 via a computer network N1, and in that the sphygmomanometer 100A includes a correction unit 115A instead of the correction unit 115.
[0048] The computer network N1 is a network that interconnects information processing devices. The computer network N1 may be wired or wireless. The computer network N1 is, for example, the Internet.
[0049] The correction unit 115A accesses the correction model 116 stored in the server 200 via the computer network N1. That is, the correction unit 115A transmits the extracted OMW information and OMWE information to the server 200 via the computer network N1.
[0050] The server 200 inputs the OMW information and OMWE information received from the sphygmomanometer 100A into the correction model 116 to obtain an error between the blood pressure measured by the oscillometric method and the blood pressure measured by the auscultatory method. The server 200 transmits the error to the main body 110 via the computer network N1. The correction unit 115A corrects the blood pressure measured by the measurement unit 114 using the error received from the server 200.
[0051] According to the first modification, the server 200 can be caused to execute the calculation process for calculating the error using the correction model 116. Therefore, the calculation load in the main body 110 can be reduced.
[0052] <Other Modifications> In the embodiment described above, pressure-type blood pressure measurement has been described with reference to FIG. 4 , but the sphygmomanometer 100 according to this embodiment may also perform pressure-reduction type blood pressure measurement.
[0053] The embodiments and modifications disclosed above can be combined with each other.
[0054] <Computer-readable recording medium> An information processing program that causes a computer or other machine or device (hereinafter referred to as a computer, etc.) to realize any of the above functions can be recorded on a computer-readable recording medium. Then, the computer, etc. can provide the function by reading and executing the program from this recording medium.
[0055] Here, a computer-readable recording medium refers to a recording medium that stores information such as data and programs electrically, magnetically, optically, mechanically, or chemically and that can be read by a computer, etc. Among such recording media, those that are removable from a computer, etc. include, for example, flexible disks, magneto-optical disks, Compact Disc Read Only Memory (CD-ROM), Compact Disc-Recordable (CD-R), Compact Disc-Rewritable (CD-RW), Digital Versatile Disc (DVD), Blu-ray Disc (BD), Digital Audio Tape (DAT), 8mm tape, flash memory, external hard disk drives, and Solid State Drives (SSD). Furthermore, examples of recording media fixed to a computer or the like include built-in hard disk drives, SSDs, ROMs, and the like.
[0056] <Supplementary Note 1> A cuff (120) to be wrapped around the arm of a person to be measured; a control unit (112) to control the pressure applied to the arm by the cuff (120); a detection unit (113) to detect the cuff pressure applied to the arm by the cuff (120); a measurement unit (114) to measure the blood pressure of the person to be measured by an oscillometric method using a time-series change in the cuff pressure; a correction model (116) constructed by machine learning for a subject using waveform information relating to the time-series change in the cuff pressure as an explanatory variable and an error in the blood pressure measured by the measurement unit from a blood pressure adopted as a true value as a target variable; and a correction unit (115) to correct the blood pressure measured by the measurement unit (114) based on the error output by the correction model (116) by inputting the waveform information relating to the time-series change in the cuff pressure detected by the detection unit (113) into the correction model (116). and an output unit (117) that outputs the blood pressure corrected by the correction unit (115). <Supplementary Note 2> The blood pressure monitor (100) according to Supplementary Note 1, wherein the blood pressure adopted as the true value is a blood pressure actually measured by auscultation. <Supplementary Note 3> The blood pressure monitor (100) according to Supplementary Note 1, wherein the information on the waveform includes information on the cuff pressure. <Supplementary Note 4> The blood pressure monitor (100) according to any one of Supplements 1 to 3, wherein the measurement unit (114) acquires an Oscillometric Waveform (OMW) based on a time-series change in the cuff pressure, and the information on the waveform includes OMW information indicating a waveform of the OMW. <Supplementary Note 5> The sphygmomanometer (100) according to any one of Supplementary Notes 1 to 4, wherein the measurement unit (114) acquires an Oscillometric Waveform Envelope (OMWE) based on the OMW, and the information on the waveform includes OMWE information indicating a waveform of the OMWE. <Supplementary Note 6> The sphygmomanometer (100) according to Supplementary Note 4, wherein the correction unit (115) corrects the blood pressure measured by the measurement unit (114) based on an error output by the correction model (116) by inputting the time-series change in the cuff pressure and the OMW information into the correction model (116).<Supplementary Note 7> The sphygmomanometer (100) according to Supplementary Note 5, wherein the correction unit (115) corrects the blood pressure measured by the measurement unit (114) based on an error output by the correction model (116) by inputting the time-series change in the cuff pressure and the OMWE information into the correction model (116). <Supplementary Note 8> A blood pressure measurement method, the method comprising: a control step of controlling pressure on the arm of a person to be measured by a cuff (120) wrapped around the arm; a detection step of detecting cuff pressure received from the arm by the cuff (120); a measurement step of measuring the blood pressure of the person to be measured by an oscillometric method using time-series changes in the cuff pressure; a correction step of correcting the blood pressure measured in the measurement step based on an error output by a correction model (116) constructed by machine learning using, for the subject, information on a waveform related to the time-series changes in the cuff pressure detected in the detection step as an explanatory variable and an error of the blood pressure measured in the measurement step from a blood pressure adopted as a true value as a target variable; and an output step of outputting the blood pressure corrected in the correction step. <Supplementary Note 9> A blood pressure measurement program that causes a computer to execute the following steps: a control step of controlling pressure on the arm of a person to be measured by a cuff (120) wrapped around the arm; a detection step of detecting cuff pressure received from the arm by the cuff (120); a measurement step of measuring the blood pressure of the person to be measured by an oscillometric method using time-series changes in the cuff pressure; a correction step of correcting the blood pressure measured in the measurement step based on an error output by a correction model (116) constructed by machine learning using information on a waveform related to the time-series changes in the cuff pressure detected in the detection step as an explanatory variable and an error of the blood pressure measured in the measurement relative to a blood pressure adopted as a true value as a target variable; and an output step of outputting the blood pressure corrected in the correction step.<Supplementary Note 10> A method for generating a learning model, in which a computer executes the following steps: a receiving step for receiving an input of an error in blood pressure measured by the oscillometric method relative to a blood pressure to be adopted as a true value; and a generating step for generating a learning model by machine learning using, as an explanatory variable, information on a waveform related to time-series changes in cuff pressure caused by a cuff wrapped around the arm of a subject, and the error as a dependent variable. <Supplementary Note 11> A program for generating a learning model, in which a computer executes the following steps: a receiving step for receiving an input of an error in blood pressure measured by the oscillometric method relative to a blood pressure to be adopted as a true value; and a generating step for generating a learning model by machine learning using, as an explanatory variable, information on a waveform related to time-series changes in cuff pressure caused by a cuff wrapped around the arm of a subject, and the error as a dependent variable.
[0057] 100: Blood pressure monitor 100A: Blood pressure monitor 110: Main body 111: Start switch 112: Control unit 113: Pressure sensor 114: Measurement unit 115: Correction unit 115A: Correction unit 116: Correction model 117: Output unit 120: Cuff 200: Server C1: Person to be measured D1: Doctor E1: Subject N1: Computer network
Claims
1. A blood pressure monitor comprising: a cuff to be wrapped around the arm of a person to be measured; a control unit that controls the pressure applied to the arm by the cuff; a detection unit that detects the cuff pressure received by the cuff from the arm; a measurement unit that measures the blood pressure of the person to be measured by an oscillometric method using a time-series change in the cuff pressure; a correction model constructed for the subject by machine learning using waveform information regarding the time-series change in the cuff pressure as an explanatory variable and an error in the blood pressure measured by the measurement unit relative to a blood pressure adopted as a true value as an objective variable; a correction unit that corrects the blood pressure measured by the measurement unit based on the error output by the correction model by inputting the waveform information regarding the time-series change in the cuff pressure detected by the detection unit into the correction model; and an output unit that outputs the blood pressure corrected by the correction unit.
2. The blood pressure monitor according to claim 1, wherein the blood pressure adopted as the true value is a blood pressure actually measured by auscultation.
3. The blood pressure monitor of claim 1, wherein the information about the waveform includes information about the cuff pressure.
4. The blood pressure monitor according to any one of claims 1 to 3, wherein the measurement unit acquires an Oscillometric Waveform (OMW) based on a time series change in the cuff pressure, and the information on the waveform includes OMW information indicating a waveform of the OMW.
5. The blood pressure monitor according to claim 4, wherein the measurement unit acquires an Oscillometric Waveform Envelope (OMWE) based on the OMW, and the information on the waveform includes OMWE information indicating a waveform of the OMWE.
6. The blood pressure monitor according to claim 4, wherein the correction unit corrects the blood pressure measured by the measurement unit based on an error output by the correction model by inputting the time-series change in the cuff pressure and the OMW information into the correction model.
7. The blood pressure monitor according to claim 5, wherein the correction unit corrects the blood pressure measured by the measurement unit based on an error output by the correction model by inputting the time-series change in the cuff pressure and the OMWE information into the correction model.
8. A blood pressure measurement method, in which a computer executes the following steps: a control step of controlling pressure applied to a subject's arm by a cuff wrapped around the arm; a detection step of detecting cuff pressure received by the cuff from the arm; a measurement step of measuring the blood pressure of the subject by an oscillometric method using the time-series change in the cuff pressure; a correction step of correcting the blood pressure measured in the measurement step based on the error output by the correction model by inputting information on the waveform related to the time-series change in the cuff pressure detected in the detection step into a correction model constructed by machine learning using, for the subject, information on the waveform related to the time-series change in the cuff pressure as an explanatory variable and an error of the blood pressure measured in the measurement step relative to the blood pressure adopted as a true value as a target variable; and an output step of outputting the blood pressure corrected in the correction step.
9. A blood pressure measurement program that causes a computer to execute the following steps: a control step of controlling pressure applied to a subject's arm by a cuff wrapped around the arm; a detection step of detecting cuff pressure received by the cuff from the arm; a measurement step of measuring the blood pressure of the subject by an oscillometric method using the time-series change in the cuff pressure; a correction step of correcting the blood pressure measured in the measurement step based on the error output by the correction model by inputting the information on the waveform related to the time-series change in the cuff pressure detected in the detection step into a correction model constructed by machine learning using, for the subject, information on the waveform related to the time-series change in the cuff pressure as an explanatory variable and an error in the blood pressure measured in the measurement relative to a blood pressure adopted as a true value as an objective variable; and an output step of outputting the blood pressure corrected in the correction step.
10. A method for constructing a learning model, the method comprising: a receiving step for receiving an input of an error in blood pressure measured by an oscillometric method relative to a blood pressure to be adopted as a true value; and a construction step for constructing a learning model by machine learning using waveform information relating to time-series changes in cuff pressure caused by a cuff wrapped around the arm of a person being measured as an explanatory variable and the error as a target variable, the method comprising: a receiving step for receiving an input of an error in blood pressure measured by an oscillometric method relative to a blood pressure to be adopted as a true value; 11. A learning model construction program that causes a computer to execute the following steps: an acceptance step for accepting input of an error in blood pressure measured by an oscillometric method relative to a blood pressure to be adopted as a true value; and a construction step for constructing a learning model by machine learning using waveform information regarding time-series changes in cuff pressure caused by a cuff wrapped around the arm of a person being measured as an explanatory variable and the error as a target variable.
Citation Information
Patent Citations
Blood pressure estimation method
JP2022516820A
Pressure value determination system, method and device and readable storage medium
CN114692668A
Electronic sphygmomanometer
JP1993309073A
Vehicle body gravity center slip angle measuring apparatus
JP1994273187A
Electronic sphygmomanometer measuring method, system, electronic sphygmomanometer, and storage medium
US20210000354A1
