Biometric information acquisition device, processing device, and computer program

Abstract

To restrict generation of a risk in which a user is forced a useless action by an alarm.SOLUTION: A receiving part 121 receives a waveform data W corresponding to a measured waveform of a biological parameter of a subject from a sensor 20. A notification part 11 outputs an alarm A showing that the biological parameter has not been normally acquired. A processing part 122 causes the notification part 11 to output the alarm A based on the waveform data W. An estimating part 124 estimates a percentage P of erroneously calculating a value of the biological parameter based on the waveform data W. The processing part 122 causes the notification part 11 to notify at least one of a quality of the waveform data W, the state of the sensor 20, and an action that a user should take based on the percentage P.SELECTED DRAWING: Figure 1

Description

[Technical field]

[0001] The present invention relates to a biometric information acquiring device that acquires biometric information of a subject through a sensor, and also to a processing device that processes the biometric information, and a computer program executable by a processing unit of the processing device. [Background technology]

[0002] The bioinformation acquiring device disclosed in Patent Document 1 receives waveform data corresponding to a measured waveform of a bioparameter of a subject. The bioinformation acquiring device outputs an alarm indicating an abnormality in at least one of the value of the bioparameter and the measured waveform based on the waveform data. The bioinformation acquiring device also outputs an alarm indicating an abnormality in the information acquisition state based on the waveform. [Prior art documents] [Patent documents]

[0003] [Patent Document 1] JP 2018-102671 A Summary of the Invention [Problem to be solved by the invention]

[0004] An object of the present invention is to prevent occurrence of a situation in which an alarm forces a user to take an unnecessary action. [Means for solving the problem]

[0005] A first aspect of the present invention to achieve the above object is a biometric information acquisition device, comprising: a reception unit that receives waveform data corresponding to a measured waveform of a biological parameter of the subject from the sensor; a notification unit that outputs an alarm indicating that the biological parameters have not been normally acquired; a processing unit that causes the notification unit to output the alarm based on the waveform data; an inference unit for inferring a probability that the value of the biological parameter is erroneously calculated based on the waveform data; It is equipped with The processing unit causes the notification unit to notify at least one of the quality of the waveform data, the state of the sensor, and an action to be taken by a user based on the probability.

[0006] A second aspect of the present invention to achieve the above object is a processing device for processing biological information of a subject, comprising: a receiving unit that receives waveform data corresponding to a measured waveform of a biological parameter of the subject from a sensor; a processing unit that causes an output device to output an alarm indicating that the biological parameter is not normally acquired based on the waveform data; an inference unit for inferring a probability that the value of the biological parameter is erroneously calculated based on the waveform data; It is equipped with The processing unit causes the output device to notify at least one of the quality of the waveform data, the state of the sensor, and an action to be taken by a user based on the probability.

[0007] A third aspect for achieving the above object is a computer program executable by a processing unit of a processing device for processing biological information of a subject, When executed, the processing device: receiving waveform data corresponding to a measured waveform of a biological parameter of the subject from a sensor; outputting an alarm indicating that the biological parameter is not normally acquired on the basis of the waveform data to an output device; inferring the probability that the value of the biological parameter is erroneously calculated based on the waveform data; Based on the probability, the output device is caused to provide at least one of an indication of the quality of the waveform data, a state of the sensor, and an action to be taken by a user.

[0008] The inventors of the present application have discovered that when an alarm is output indicating that bioinformation is not being acquired normally, there may be cases where the probability that the value of a bioparameter will be erroneously calculated based on the measured waveform of the bioinformation remains high.

[0009] According to the configurations of the first to third aspects, by referring to the probability inferred by the inference unit, it is possible to estimate the symptoms that may lead to a state in which an alarm is frequently generated, and to prompt the user to take action, thereby preventing the occurrence of a situation in which the user is forced to take an unnecessary action due to an alarm.

[0010] A fourth aspect of the present invention to achieve the above object is a biometric information acquisition device, comprising: A reception unit that receives waveform data corresponding to a measured waveform of a biological parameter of a subject; a notification unit that outputs an alarm indicating an abnormality in at least one of the value of the biological parameter and the measured waveform; a processing unit that causes the notification unit to output the alarm based on the waveform data; an inference unit for inferring a probability that the value of the biological parameter is erroneously calculated based on the waveform data; It is equipped with The processing unit controls output of the alarm when the probability exceeds a threshold.

[0011] A fifth aspect for achieving the above object is a processing device for processing biological information of a subject, comprising: a reception unit for receiving waveform data corresponding to a measured waveform of a biological parameter of the subject; a processing unit that causes an output device to output an alarm indicating an abnormality in at least one of the value of the biological parameter and the measured waveform based on the waveform data; an inference unit for inferring a probability that the value of the biological parameter is erroneously calculated based on the waveform data; It is equipped with The processing unit controls output of the alarm when the probability exceeds a threshold.

[0012] A sixth aspect for achieving the above object is a computer program executable by a processing unit of a processing device for processing biological information of a subject, When executed, the processing device: receiving waveform data corresponding to a measured waveform of a biological parameter of the subject; outputting an alarm indicating an abnormality in at least one of the value of the biological parameter and the measured waveform to an output device; inferring the probability that the value of the biological parameter is erroneously calculated based on the waveform data; If the probability exceeds a threshold, the output of the alarm is controlled.

[0013] Even if some abnormality is found in the waveform data received by the receiving unit, if there is a high probability that the biological parameters inferred by the inference unit based on the same waveform data are erroneously calculated, there is a high probability that the abnormality has been erroneously detected. According to the configurations according to the fourth to sixth aspects for controlling the alarm output in such a case, it is possible to make the user aware that there is a possibility that an abnormality has been erroneously detected. Therefore, it is possible to prevent the occurrence of a situation in which the user is forced to take unnecessary action due to an alarm. [Brief description of the drawings]

[0014] [Figure 1] 2 illustrates an example of a functional configuration of a biological information acquisition device according to an embodiment. [Diagram 2] 2 illustrates a method for generating the inference unit of FIG. 1; [Diagram 3] 4 illustrates an example of a relationship between electrocardiogram waveform data and first heart rate data. [Figure 4] 13 illustrates an example of a relationship between electrocardiogram waveform data and second heart rate data. [Diagram 5] 1 illustrates exemplary data obtained from a subject showing changes in heart rate over time. [Figure 6] 13 illustrates an example of data showing changes in heart rate over time that have been processed by an inference unit. [Figure 7]2 is a diagram for explaining an example of processing performed by the processing unit in FIG. 1; [Figure 8] 2 is a diagram for explaining an example of processing performed by the processing unit in FIG. 1; [Figure 9] 2 is a diagram for explaining an example of processing executed by the processing unit in FIG. 1; [Figure 10] 10 illustrates another example of processing executed by the processing unit in FIG. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0015] Exemplary embodiments are described in detail below with reference to the accompanying drawings.

[0016] 1 illustrates a functional configuration of a bioinformation acquisition device 10 according to an embodiment. The bioinformation acquisition device 10 is a device that acquires bioparameters of a subject 30 through a sensor 20. Examples of the sensor 20 include an electrode for acquiring a heart rate, an electrode for acquiring an impedance respiration rate, a catheter for acquiring an invasive blood pressure value, a probe for acquiring a percutaneous arterial oxygen saturation (SpO2) or a pulse rate, and a flow sensor for acquiring a respiratory carbon dioxide concentration.

[0017] The bioinformation acquisition device 10 includes an alarm unit 11. The alarm unit 11 is configured to be capable of outputting an alarm A. The alarm A may include an alarm (a so-called technical alarm) indicating that a bioparameter has not been acquired normally, and an alarm (a so-called biological alarm) indicating that there is an abnormality in at least one of the value of the bioparameter and the measured waveform of the bioparameter. The alarm A includes at least one of a visual alarm, an auditory alarm, and a tactile alarm. The alarm unit 11 is an example of an output device.

[0018] The biometric information acquiring device 10 includes a processing device 12. The processing device 12 includes a receiving unit 121, a processing unit 122, and an output unit 123.

[0019] The receiving unit 121 is configured to receive waveform data W corresponding to a measured waveform of a biological parameter of the subject 30 from the sensor 20. The waveform data W may be in the form of analog data or digital data. When the waveform data W is in the form of analog data, the receiving unit 121 is configured to include an appropriate conversion circuit including an A / D converter. The processing unit 122 processes the waveform data W in the form of digital data.

[0020] The processing unit 122 is configured to output, from the output unit 123, first control data C1 for causing the notification unit 11 to output an alarm A based on the waveform data W. For example, if the sensor 20 is displaced from a predetermined position or if some malfunction occurs in the operation of the sensor 20, the appropriate measurement waveform may be distorted or the signal level may decrease. If the processing unit 122 determines that the waveform data W corresponds to such a state, it transmits the first control data C1 from the output unit 123 to the notification unit 11.

[0021] The first control data C1 transmitted to the notification unit 11 may be in the form of analog data or digital data. When the first control data C1 is in the form of analog data, the output unit 123 is configured to include an appropriate conversion circuit including a D / A converter.

[0022] The processing unit 122 may be realized by a general-purpose microprocessor that operates in cooperation with a general-purpose memory. Examples of the general-purpose microprocessor include a CPU, an MPU, and a GPU. Examples of the general-purpose memory include a ROM and a RAM. In this case, a computer program that executes the process described below may be stored in the ROM. The ROM is an example of a non-transient computer-readable medium in which a computer program is stored. The processor specifies at least a part of the computer program stored in the ROM, deploys it on the RAM, and executes the above-mentioned process in cooperation with the RAM. The above computer program may be pre-installed in the general-purpose memory, or may be downloaded from an external server device via a communication network and installed in the general-purpose memory. In this case, the external server device is an example of a non-transient computer-readable medium in which a computer program is stored.

[0023] The processing unit 122 may be realized by a dedicated integrated circuit capable of executing the above computer program, such as a microcontroller, an ASIC, or an FPGA. In this case, the above computer program is pre-installed in a memory element included in the dedicated integrated circuit. The memory element is an example of a non-transitory computer-readable medium on which a computer program is stored. The processing unit 122 may also be realized by a combination of a general-purpose microprocessor and a dedicated integrated circuit.

[0024] The processing device 12 includes an inference unit 124. The inference unit 124 is configured to infer a probability P of an erroneous calculation of a value of a biological parameter based on the waveform data W. The inference unit 124 is realized as a trained model or a classifier generated by machine learning, which will be described later.

[0025] With reference to FIG. 2, a method for generating the inference unit 124 capable of inferring the probability that the heart rate will be erroneously calculated based on the measured waveform of an electrocardiogram will be described as an example.

[0026] First, a process is performed to obtain first heart rate data H1 from the electrocardiogram waveform data E (STEP 1). The first heart rate data H1 is data indicating a heart rate value correctly calculated from the electrocardiogram waveform data E. The electrocardiogram waveform data E is an example of waveform data W. The heart rate is an example of a biological parameter.

[0027] FIG. 3 illustrates the relationship between electrocardiogram waveform data E and first heart rate data H1. The electrocardiogram waveform data E shows changes over time in physiological electrical activity of the subject's heart. The electrocardiogram waveform data E changes with the beating of the heart. The first heart rate data H1 is calculated by counting the number of such changes per minute. The broken lines extending vertically in FIG. 3 indicate the portion of the electrocardiogram waveform data E that is used for counting the heart rate. In the figure, 12 heart beats are counted within the time period in which the waveform is displayed (e.g., 10 seconds).

[0028] The heart rate corresponding to the first heart rate data H1 may be counted by visually observing the electrocardiogram waveform data E, or by using appropriate counting software. The electrocardiogram waveform data E may be waveform data actually obtained from a subject using an electrocardiogram sensor or the like, or waveform data published as a database for performance evaluation of electrocardiographs may be used.

[0029] Subsequently or in parallel, a process (STEP 2 in FIG. 2) is performed to acquire second heart rate data H2 from the electrocardiogram waveform data E. The second heart rate data H2 is data indicating a heart rate value automatically calculated from the electrocardiogram waveform data E by a specific heart rate calculation algorithm. The heart rate calculation algorithm may be executed, for example, by the processing unit 122 of the processing device 12 mounted on the biometric information acquisition device 10. The heart rate calculation algorithm is configured to automatically calculate the heart rate based on the characteristics of the waveform shape in the input electrocardiogram waveform data E.

[0030] Fig. 4 illustrates the relationship between the electrocardiogram waveform data E and the second heart rate data H2. The electrocardiogram waveform data E illustrated in Fig. 4 is the same as the electrocardiogram waveform data E illustrated in Fig. 3. Based on the same electrocardiogram waveform data E, 19 heart beats are counted in the example illustrated in Fig. 4, instead of 12. This fact means that the heart rate calculation algorithm used to obtain the second heart rate data H2 may perform counting based on erroneously detected heart beats.

[0031] Fig. 5 shows an example of the change over time in heart rate that is automatically calculated by inputting electrocardiogram waveform data E actually obtained from a subject into the above-mentioned heart rate calculation algorithm. Each of the multiple points included in the graph corresponds to the heart rate detected within a predetermined time (e.g., one minute).

[0032] In FIG. 6, the erroneously calculated values ​​of the heart rate values ​​illustrated in FIG. 5 are represented by white circles. One method for excluding erroneously calculated heart rate data is to set a threshold range for the heart rate. For example, heart rate data whose heart rate is not between a lower threshold Th1 and an upper threshold Th2 can be excluded. However, as shown in FIG. 6, erroneously calculated heart rate data may exist even if it is within the threshold range, and conversely, correctly calculated heart rate data may exist even if it is outside the threshold range. The former is based on erroneous analysis of the heart rate calculation algorithm, and the latter corresponds to a case where the heart rate increases due to some physiological phenomenon in the subject. The erroneous analysis of the heart rate calculation algorithm may occur, for example, due to a phenomenon (so-called double counting) in which a high-amplitude T wave is erroneously recognized as a QRS wave and the heart rate is counted higher. The erroneous analysis of the heart rate calculation algorithm may also occur due to external noise such as the subject's body movement and deterioration or detachment of electrodes attached to the subject.

[0033] Next, a process is performed (STEP 3 in Figure 2) in which second heart rate data H2 corresponding to the heart rate automatically calculated by the heart rate calculation algorithm is compared with the first heart rate data H1 to generate teacher data T including a teacher label indicating whether the heart rate is correct or incorrect.

[0034] When the first heart rate data H1 illustrated in FIG. 3 and the second heart rate data H2 illustrated in FIG. 4 are subjected to matching, the teacher data T includes a teacher label indicating that the heart rate corresponding to the second heart rate data H2 is incorrect.

[0035] In order to include a teacher label indicating a correct answer, it is not necessary that the heart rate corresponding to the first heart rate data H1 and the heart rate corresponding to the second heart rate data H2 completely match. As long as the error between the heart rate corresponding to the first heart rate data H1 and the heart rate corresponding to the second heart rate data H2 is within an allowable range, the teacher data T may include teacher data indicating a correct answer.

[0036] That is, the heart rate calculated automatically by the heart rate calculation algorithm is subjected to comparison with the heart rate that is known to be correctly calculated. Therefore, the teacher label obtained as a result of the comparison can reflect the tendency or habit of the heart rate calculation algorithm to correctly or erroneously calculate the heart rate for the input electrocardiogram waveform data E.

[0037] Next, machine learning is performed by the inference unit 124 based on a combination of the electrocardiogram waveform data E and the teacher data T (STEP 4 in FIG. 2). For the machine learning, a well-known method related to supervised learning is appropriately used. That is, the inference unit 124 learns the feature amount of the electrocardiogram waveform data E from which the heart rate has been erroneously calculated, and corrects the calculation parameters so as to output inference data indicating that the probability P of the heart rate being erroneously calculated is high when the electrocardiogram waveform data E having such feature amount is input. The correction of the calculation parameters may be performed by the person who performs the learning, or may be performed by the inference unit 124 itself through learning. The inference data may indicate the calculated probability P itself, or may be associated with a score (e.g., any of the values ​​from 1 to 5) corresponding to the probability P, or the like.

[0038] 1, the processing unit 122 is configured to output second control data C2 from the output unit 123 based on the probability P calculated by the inference unit 124. The second control data C2 is configured to cause the notification unit 11 to output a preliminary notification N, which will be described later. The second control data C2 transmitted to the notification unit 11 may be in the form of analog data or digital data. When the second control data C2 is in the form of analog data, the output unit 123 is configured to include an appropriate conversion circuit including a D / A converter.

[0039] The preliminary notification N performed by the biometric information acquisition device 10 will be described in detail with reference to FIGS.

[0040] FIG. 7 is a bar graph illustrating the frequency of alarms output when a conventional biometric information acquisition device is used to acquire an electrocardiogram waveform of a subject. An alarm is output to indicate that acquisition of the heart rate is not being performed normally. An alarm is output due to the subject's body movement, deterioration, misalignment, or falling off of the electrodes attached to the subject, etc. A blank bar represents an alarm with an output time of less than 5 seconds. A bar with diagonal hatching represents an alarm with an output time of 5 seconds or more but less than 30 seconds. A bar with grid hatching represents an alarm with an output time of 30 seconds or more.

[0041] In the example shown in Figure 7, an alarm starts to be output at time t1. It can be seen that the aforementioned cause that prevents the normal acquisition of the heart rate occurs. As time passes, the frequency of alarm output increases, and the number of alarms with long output times increases, indicating that the situation is worsening. At time t2, the electrodes are replaced or rearranged, and the alarm output stops.

[0042] Fig. 8 is a graph plotting the probability P output from the inference unit 124 when the electrocardiogram waveform acquired in the example shown in Fig. 7 is input to the inference unit 124 according to this embodiment. Fig. 9 is a graph showing the change over time in the probability P smoothed by calculating the time moving average value of the plot group shown in Fig. 8, superimposed on the bar graph of Fig. 7. It can be seen that there is a correlation between the period during which an alarm is output and the period during which the probability P of the heart rate output by the inference unit 124 being erroneously calculated is high.

[0043] The inventors of the present application focused on the period prior to time t1. Rather than a sudden transition from a state in which the probability P remains low before time t0 prior to time t1 to a state in which the probability P remains high after time t1, the frequency with which the probability P assumes a relatively high value increases gradually during the period from time t0 to time t1. In other words, there is a period in which the state transitions from a state in which the heart rate is being normally acquired to a state in which the heart rate is not being normally acquired without the output of an alarm.

[0044] In view of the above, the processing unit 122 is configured to cause the alarm unit 11 to output a preliminary notification N based on the upward trend of the probability P calculated by the inference unit 124. For example, the processing unit 122 may cause the alarm unit 11 to output the preliminary notification N when the frequency (number of times per unit time) at which the probability P exceeds a predetermined probability threshold exceeds a predetermined frequency threshold. Alternatively, the processing unit 122 may cause the alarm unit 11 to output the preliminary notification N when the cumulative number of times that the probability P exceeds the predetermined probability threshold exceeds a predetermined number threshold. Alternatively, the processing unit 122 may cause the alarm unit 11 to output the preliminary notification N when a time moving average value of the probability P exemplified in FIG. 9 exceeds a predetermined probability threshold.

[0045] The preliminary notification N notifies at least one of the quality of the waveform data W, the state of the sensor 20, and the action that the user should take. The preliminary notification N includes at least one of a visual notification, an audio notification, and a tactile notification.

[0046] In the examples shown in Figures 8 and 9, the increase in the probability P of erroneously calculating the heart rate can be caused by deviation of the electrocardiogram waveform from the normal state due to body movement, deviation of the electrocardiogram waveform from the normal state due to deterioration or misalignment of the electrodes, etc. For example, the preliminary notification N can be a message such as "The waveform is distorted," "There is a lot of body movement," or "Please check the condition of the electrodes" provided to the user through display or voice.

[0047] According to this configuration, it is possible to estimate the symptoms that may lead to a state where an alarm is frequently issued and to prompt the user to take action, thereby preventing the occurrence of a situation where the user is forced to take an unnecessary action due to an alarm.

[0048] It is preferable that the conditions for outputting the preliminary notification N are set so that the preliminary notification N is issued before the state in which the alarm A is output by the notification unit 11. In the example shown in Fig. 8 and Fig. 9, the conditions related to the probability P are set so that the preliminary notification N is issued before time t1. By prompting the user to take an appropriate action by the preliminary notification N prior to the output of the alarm A, it is possible to increase the possibility of avoiding the output of the alarm A, which causes a certain degree of tension for the user.

[0049] The electrocardiogram waveform data E is a waveform obtained by synthesizing a plurality of lead waveforms, or a waveform obtained by simply acquiring a single lead (typically lead II). However, the inference unit 124 may be generated so as to estimate the probability that the heart rate will be erroneously calculated based on each of the plurality of lead waveforms. This configuration allows for detailed status management for each electrode from which each lead waveform is acquired, making it easy to prevent the occurrence of a situation in which an alarm forces the user to take unnecessary action.

[0050] Next, another configuration example of the bioinformation acquisition device 10 will be described with reference to Fig. 10. In this example, the processing unit 122 of the processing device 12 judges whether or not there is an abnormality in at least one of the value of the bioparameter and the measured waveform of the bioparameter based on the waveform data W received by the receiving unit 121 (STEP 11). This process is repeated until it is judged that there is an abnormality in at least one of the value of the bioparameter and the measured waveform of the bioparameter (NO in STEP 11).

[0051] When it is determined that there is an abnormality in at least one of the value of the biological parameter and the measured waveform of the biological parameter (YES in STEP 11), the processing unit 122 determines whether the probability P that the value of the biological parameter inferred by the inference unit 124 is erroneously calculated exceeds a probability threshold Pth (STEP 12).

[0052] If it is determined that the probability P does not exceed the probability threshold Pth (NO in STEP 12), the processing unit 122 outputs first control data C1 for causing the notification unit 11 to output a predefined normal alarm A from the output unit 123 (STEP 13). As described above, the alarm A can be output using at least one of a visual technique, an auditory technique, and a tactile technique.

[0053] When it is determined that the probability P exceeds the probability threshold Pth (YES in STEP 12), the processing unit 122 outputs a third control signal CS3 from the output unit 123 to cause the notification unit 11 to output a controlled alarm A' (STEP 14). As used in this specification, the expressions "controlled alarm" or "controlling the output of an alarm" refer to applying some change to a normal alarm A assigned to a certain situation. Examples of changes include changing the urgency of the alarm, stopping the output of the alarm, adding information to be provided, omitting information to be provided, etc. In other words, the expression "controlled output of an alarm" includes the meaning that a normal alarm is not output.

[0054] For example, alarm A may include levels of "emergency," "alert," and "caution" depending on the urgency. When an electrocardiogram waveform is to be acquired, an "emergency" level alarm may be assigned to the occurrence of ventricular fibrillation or ventricular tachycardia. An "alert" level alarm may be assigned to the occurrence of abnormal heart rate or premature ventricular contractions. An "alert" level alarm may be assigned to other abnormalities.

[0055] The notification mode differs between levels. For example, if alarm A includes a visual alarm, the displayed color and symbol will differ. If alarm A includes an auditory alarm, the output volume, tone, length, and voice will differ.

[0056] For example, if it is determined that ventricular fibrillation is occurring based on the electrocardiogram waveform data E and the probability P that the heart rate is incorrectly calculated does not exceed the probability threshold Pth (NO in STEP 12), the processing unit 122 causes the notification unit 11 to output an alarm A of the "emergency" level assigned to ventricular fibrillation (STEP 13).

[0057] On the other hand, if it is determined that ventricular fibrillation is occurring based on the electrocardiogram waveform data E and the probability P that the heart rate is incorrectly calculated exceeds the probability threshold Pth (YES in STEP 12), the processing unit 122 applies a change to alarm A of the "emergency" level assigned to ventricular fibrillation and outputs it as a controlled alarm A' (STEP 14).

[0058] Even if some abnormality is found in the electrocardiogram waveform received by the receiving unit 121, if the probability P that the heart rate inferred by the inference unit 124 based on the same electrocardiogram waveform is erroneously calculated is high, the probability that the abnormality has been erroneously detected is high. In such a case, by applying some change to the alarm A output by the notification unit 11, it is possible to make the user aware that there is a possibility that an abnormality has been erroneously detected. Therefore, it is possible to prevent the occurrence of a situation in which the user is forced to take unnecessary action due to an alarm.

[0059] The processing unit 122 may cause the notification unit 11 to output an indicator indicating the probability P estimated by the inference unit 124. The indicator may be provided using at least one of a visual indicator, an auditory indicator, and a tactile indicator.

[0060] According to this configuration, it is possible to more reliably make the user aware of the possibility that an abnormality has been erroneously detected.

[0061] When the alarm A includes a visual alarm and an audible alarm, the processing unit 122 may output an alarm A' that is controlled so as to reduce the urgency of at least one of the visual alarm and the audible alarm.

[0062] For example, if normal alarm A includes a visual alarm and an audible alarm assigned to an "emergency" level, at least one of the visual alarm and the audible alarm may be changed to one assigned to an "alert" level or a "caution" level. For example, the color display assigned to the "emergency level" is maintained, while the sound assigned to the "emergency level" is changed to a sound assigned to the "alert level."

[0063] With this configuration, the alarm output itself is maintained to alert the user, while the sense of tension associated with an output of a higher level of urgency can be alleviated. When the urgency of one of the audiovisual and auditory alarms is reduced, the user can be made aware of the possibility that an abnormality has been erroneously detected based on facts different from those of a normal alarm combination.

[0064] The processing unit 122 can cause the notification unit 11 to provide additional information corresponding to the fact that the probability P of the erroneous detection of the biological parameter inferred by the inference unit 124 exceeds the probability threshold Pth, while maintaining the urgency of the normal alarm A. The additional information can be provided using at least one of a visual method, an auditory method, and a tactile method.

[0065] With this configuration, the alarm output itself is maintained to attract the user's attention, but additional information is provided to make the user aware of the possibility that an abnormality has been mistakenly detected, thereby easing the sense of tension.

[0066] The above-described embodiment is merely an example for facilitating understanding of the present invention. The configurations according to the above-described embodiment may be appropriately modified or improved without departing from the spirit and scope of the present invention.

[0067] In the above embodiment, in order to generate the inference unit 124 for inferring the probability that the heart rate is erroneously calculated, the first heart rate data H1 and the second heart rate data H2 are acquired using the common electrocardiogram waveform data E. As described above, the first heart rate data H1 is data indicating a heart rate correctly calculated from the electrocardiogram waveform data E. The second heart rate data H2 is data indicating a heart rate calculated by a specific heart rate calculation algorithm, and may include an erroneously calculated heart rate.

[0068] As another example, the first respiratory rate data and the second respiratory rate data may be obtained using common capnogram data to generate the inference unit 124 for inferring the probability that the respiratory rate is erroneously calculated. The capnogram data is an example of waveform data. The respiratory rate is an example of a biological parameter. The first respiratory rate data is data indicating a respiratory rate correctly calculated from the capnogram data. The second respiratory rate data is data indicating a respiratory rate calculated by a specific respiratory rate calculation algorithm, and may include an erroneously calculated respiratory rate.

[0069] The biological parameter for which the inference unit 124 estimates the probability of miscalculation may be different from the biological parameter for which an alarm is output. For example, the inference unit 124 may be used to infer the probability that the heart rate is erroneously calculated in a configuration in which an alarm is output for abnormalities such as ST value, QT interval, and QTc in electrocardiogram waveform data. Alternatively, the inference unit 124 may be used to infer the probability that the respiratory rate is erroneously calculated in a configuration in which an alarm is output for abnormalities such as end-tidal carbon dioxide partial pressure (EtCO2) in capnogram data.

[0070] Different types of biological parameters may be used to generate the inference unit 124. As an example, heart rate data obtained from electrocardiogram waveform data and pulse rate data obtained from invasive arterial pressure waveform data may be used to generate the inference unit 124 for inferring the probability that the pulse rate is erroneously calculated. The electrocardiogram waveform data and the invasive arterial pressure waveform data must be obtained from the same subject. The heart rate data is data indicating a heart rate correctly calculated from the electrocardiogram waveform data. The pulse rate data is data indicating a pulse rate calculated by a specific pulse rate calculation algorithm, and may include an erroneously calculated pulse rate.

[0071] The inference unit 124 generated as described above is used to infer the probability that the invasive arterial pressure is erroneously calculated. The invasive arterial pressure waveform data is an example of waveform data. The invasive arterial pressure is an example of a biological parameter.

[0072] Pulse waveform data may be used to obtain pulse rate data. The pulse waveform data may be obtained by non-invasively measuring the change in blood absorbance over time using a pulse photometry probe or the like. Alternatively, the pulse waveform data may be obtained by measuring the change in intra-cuff pressure over time using a non-invasive arterial pressure meter or the like. In this case, the generated inference unit 124 is used to infer the probability that the non-invasive arterial pressure, pulse rate, and percutaneous arterial oxygen saturation (SpO2) are erroneously calculated. The pulse waveform data is an example of waveform data. The invasive arterial pressure, pulse rate, and SpO2 are examples of biological parameters.

[0073] As another example, first respiratory rate data obtained from impedance respiratory waveform data and second respiratory rate data obtained from capnogram data may be used to generate the inference unit 124 for inferring the probability that the respiratory rate is erroneously calculated. The impedance respiratory waveform data and the capnogram data must be obtained from the same subject. The first respiratory rate data is data indicating a respiratory rate correctly calculated from the impedance respiratory waveform data. The second respiratory rate data is data indicating a respiratory rate calculated by a specific respiratory rate calculation algorithm, and may include an erroneously calculated respiratory rate. The impedance respiratory waveform data is an example of waveform data. The respiratory rate is an example of a biological parameter.

[0074] In the above embodiment, the notification unit 11 that outputs the alarm A, the controlled alarm A', and the preliminary notification N is mounted on the biometric information acquisition device 10. However, the function of the notification unit 11 may be realized in an independent output device that is capable of data communication with the biometric information acquisition device 10 via a communication network. In this case, the processing unit 122 of the processing device 12 transmits, from the output unit 123, first control data C1 and second control data C2 that cause the output device to output the alarm A, the controlled alarm A', and the preliminary notification N.

[0075] In the above embodiment, both the notification unit 11 and the processing device 12 are mounted on the biometric information acquisition device 10. However, the processing device 12 may be mounted on an external server device capable of communicating with the biometric information acquisition device 10 via a communication network. In this case, the processing device 12 executes the above-mentioned process based on the waveform data W received from the sensor 20 or the biometric information acquisition device 10 via the communication network, transmits appropriate control data via the communication network, and causes the notification unit 11 to output the alarm A, the controlled alarm A', and the preliminary notification N. The device that executes the operation based on the control data transmitted from the external server device may be an output device independent of the biometric information acquisition device 10. [Explanation of symbols]

[0076] 10: monitor device, 11: notification unit, 12: control device, 121: reception unit, 122: processing unit, 124: inference unit, 20: sensor, 30: subject, A: alarm, A': controlled alarm, N: preliminary notification, P: probability, W: waveform data

Claims

1. a reception unit that receives waveform data corresponding to a measured waveform of a biological parameter of the subject from the sensor; A notification department; a processing unit that outputs an alarm to the notification unit when it is determined that the biological parameter is not normally acquired based on the waveform data; an inference unit which is a trained model or a classifier generated by learning a feature of waveform data corresponding to a measured waveform in which the biological parameter is not normally acquired using a machine learning model, and outputting inference data indicating that there is a high probability that the value of the biological parameter will be erroneously calculated when waveform data having the feature is input; Equipped with When a probability corresponding to the inference data output by inputting the waveform data to the inference unit satisfies a condition relating to an upward trend, the processing unit causes the notification unit to notify at least one of the quality of the waveform data, the state of the sensor, and an action to be taken by a user. Biometric information acquisition device.

2. The processing unit causes the notification unit to perform the notification based on a frequency at which the probability exceeds a threshold. The biological information acquiring device according to claim 1 .

3. The processing unit may cause the notification unit to perform the notification when the probability satisfies a condition related to the upward trend before it is determined that the biological parameter has not been normally acquired based on the waveform data. The biological information acquiring device according to claim 1 .

4. The biological parameter is heart rate. The biometric information acquiring device according to claim 1 .

5. the sensor includes a plurality of electrodes attached to the subject to measure an electrocardiogram waveform; the inference unit calculates the probability for the waveform data corresponding to each of a plurality of lead waveforms included in the electrocardiogram waveform. The biological information acquiring device according to claim 4 .

6. A processing device for processing biological information of a subject, a receiving unit that receives waveform data corresponding to a measured waveform of a biological parameter of the subject from a sensor; a processing unit that outputs an alarm to an output device when it is determined that the biological parameter is not normally acquired based on the waveform data; an inference unit which is a trained model or a classifier generated by learning a feature of waveform data corresponding to a measured waveform in which the biological parameter is not normally acquired using a machine learning model, and outputting inference data indicating that there is a high probability that the value of the biological parameter will be erroneously calculated when waveform data having the feature is input; It is equipped with When the probability corresponding to the inference data outputted by inputting the waveform data to the inference unit satisfies a condition relating to an upward trend, the processing unit causes the output device to notify at least one of the quality of the waveform data, the state of the sensor, and an action to be taken by a user. Processing unit.

7. A computer program executable by a processing unit of a processing device that processes biological information of a subject, When executed, the processing device: receiving waveform data corresponding to a measured waveform of a biological parameter of the subject from a sensor; When it is determined that the biological parameter is not normally acquired based on the waveform data, an alarm is outputted to an output device. inputting the waveform data into an inference unit which is a trained model or a classifier generated by learning a feature amount of waveform data corresponding to a measured waveform in which the biological parameter is not normally acquired using a machine learning model, and outputting inference data indicating that there is a high probability that the value of the biological parameter will be erroneously calculated when waveform data having the feature amount is input; When the probability corresponding to the inference data satisfies a condition relating to an upward trend, the output device is caused to notify at least one of the quality of the waveform data, the state of the sensor, and an action to be taken by a user. Computer program.

8. A reception unit that receives waveform data corresponding to a measured waveform of a biological parameter of a subject; A notification department; an inference unit which is a trained model or a classifier generated by learning a feature of waveform data corresponding to a measured waveform in which the biological parameter is not normally acquired using a machine learning model, and outputting inference data indicating that there is a high probability that the value of the biological parameter will be erroneously calculated when waveform data having the feature is input; a processing unit that outputs an alarm to the notification unit when it is determined that the biological parameter is not normally acquired based on the waveform data and the probability corresponding to the inference data output by inputting the waveform data to the inference unit does not exceed a threshold value; It is equipped with the processing unit, when it is determined that the biological parameter has not been normally acquired based on the waveform data and a probability corresponding to the inference data output by inputting the waveform data to the inference unit exceeds a threshold, modifies the alarm to be output; The change includes at least one of changing the urgency of the alarm, stopping the output of the alarm, adding information to be provided, and omitting information to be provided. Biometric information acquisition device.

9. The notification unit notifies an index corresponding to the probability. The biological information acquiring device according to claim 8 .

10. The alarm includes a visual alarm and an audible alarm; The processor changes at least one of the visual alarm and the audible alarm to a lower level of urgency. The biological information acquiring device according to claim 8 .

11. The processor provides an additional notification corresponding to the probability exceeding the threshold while maintaining the urgency of the alarm. The biological information acquiring device according to claim 8 .

12. A processing device for processing biological information of a subject, a reception unit for receiving waveform data corresponding to a measured waveform of a biological parameter of the subject; an inference unit which is a trained model or a classifier generated by learning a feature of waveform data corresponding to a measured waveform in which the biological parameter is not normally acquired using a machine learning model, and outputting inference data indicating that there is a high probability that the value of the biological parameter will be erroneously calculated when waveform data having the feature is input; a processing unit that causes an output device to output an alarm when it is determined that the biological parameter is not normally acquired based on the waveform data and a probability corresponding to the inference data output by inputting the waveform data to the inference unit does not exceed a threshold value; It is equipped with the processing unit, when it is determined that the biological parameter has not been normally acquired based on the waveform data and the probability corresponding to the inference data output by inputting the waveform data to the inference unit exceeds a threshold, modifies the alarm to be output; The change includes at least one of changing the urgency of the alarm, stopping the output of the alarm, adding information to be provided, and omitting information to be provided. Processing unit.

13. A computer program executable by a processing unit of a processing device that processes biological information of a subject, By executing the computer program, the processing device receiving waveform data corresponding to a measured waveform of a biological parameter of the subject; inputting the waveform data into an inference unit which is a trained model or a classifier generated by learning a feature amount of waveform data corresponding to a measured waveform in which the biological parameter is not normally acquired using a machine learning model, and outputting inference data indicating that there is a high probability that the value of the biological parameter will be erroneously calculated when waveform data having the feature amount is input; a processing unit that causes an output device to output an alarm when it is determined that the biological parameter is not normally acquired based on the waveform data and a probability corresponding to the inference data output by inputting the waveform data to the inference unit does not exceed a threshold value; When it is determined that the biological parameter has not been normally acquired based on the waveform data, and the probability that the waveform data corresponds to the inference data output by the inference unit exceeds a threshold value, modifying the alarm to be output; The change includes at least one of changing the urgency of the alarm, stopping the output of the alarm, adding information to be provided, and omitting information to be provided. Computer program.

Images