JP2022041969A - Computer-assisted method for noninvasive detection of carbohydrate metabolism disorders from heart rate variability and wearable autonomous device for its implementation - Google Patents

Abstract

To provide a computer-assisted method for noninvasively detecting carbohydrate metabolism disorders (CMD) from heart rate variability (HRV).SOLUTION: A computer-assisted method includes the steps of: selecting a cardio signal from a group consisting of an ECG and a PPG; determining a sequence of time stamps selected from a group consisting of R, P, and T-peaks on the ECG and systoles and diastoles on the PPG; determining time intervals between adjacent time stamps in the sequence to obtain a cardiointervalogram (CIG) characterizing a patient's HRV; carrying out transformation of the CIG by removing a linear trend to obtain a cardiointervalogram (TCIG); obtaining a set of statistical parameters of HRV for each of the TCIG; and carrying out computer processing of the statistical spectral parameters of HRV and making a decision on the presence of CMD of the patient according to comparison between healthy patients and patients with CMD and the degree of closeness of the patient's value to values of the patients with CMD.SELECTED DRAWING: Figure 1

Description

The present invention relates to medicine, and more particularly to a computer-aided method for detecting an inborn error of glucose metabolism from heart rate variability (HRV) and a wearable autonomous device for carrying out the method.

The method and device can be used to construct a large two-step screening of the population to identify patients with inborn errors of glucose metabolism (CMD).

In the first phase of screening, the proposed methods and devices are used to identify patients with suspected inborn errors of carbohydrate metabolism and send them to a medical facility for further testing.

In the second stage, a medical professional (a medical institution expert based on national recommendations) makes the final decision on the presence or absence of an inborn error of glucose metabolism in a patient who has passed the first stage of screening.

There are quite a few technical means, including scales, blood pressure, and blood glucose sensors to control the health of the human body.

Many technical solutions are known for assessing the human condition using technical means and various biosensors, followed by data analysis using computing devices.

Known from the prior art is a method for continuously monitoring functional status and providing functional diagnostics (see, for example, WO2017 / 160186A1 published September 21, 2017). The method consists of sending data from a biometrics monitor (wristwatch or bracelet) to a mobile application on an Android or iOS platform by wired or wireless means. The stress level is then calculated based on the histogram plotted using the measured pulse values, and then the histogram from the concentration of the maximum number of intervals between adjacent heart rates within a given range of X millisecond. The stress level is determined based on the data, and an index of the strength of the user's movement is recorded using an electric accelerometer built into the biometrics monitor. Then, based on the distribution of the obtained indicators over time, conclusions were drawn about the individual's daily motor activity and lifestyle, and based on the information gathered, what exactly caused the change in stress level. A conclusion is drawn about the stress and the user is advised to avoid the cause of the problem.

The technical problems solved by the present invention are the ability to take into account the situation of external factors of the environment in the process of human life, taking into account the physiological response to the action of these factors, the level of physical activity. And the ability to take into account sleep duration, simplify the user's perception of final information, and obtain the consequences of the impact of the environment on humans online.

Certain methods of continuously monitoring a person's level of stress state are based on the use of a biometrics detector, and the data from this detector is the value of the interval between adjacent heartbeats within a particular time frame. Is used to buffer and create a histogram of the distribution of these intervals to calculate stress levels based on heart rate variability.

The level of stress is associated with human activity and interaction with surrounding people and objects by forming a time delay between the event that occurs and the spike in the level of stress, according to which stress It is concluded what was the exact cause of the change in the level of, and the user is asked to rule out this cause.

The disadvantage of this method is that it compares the statistical spectral parameters of a patient's heart rate variability (HRV) with the reference statistical spectral parameters of HRV from healthy patients and patients with impaired carbohydrate metabolism. Includes the fact that the presence of impaired carbohydrate metabolism in patients cannot be assessed due to parameter deviations.

Known from prior art is a device for determining subject heart rate, respiratory rate, and / or blood pressure data based on pulse waveform analysis (eg, published January 19, 2017). See WO2017009465 (A1)).

The device provides an implementation of a method for determining health, for example, data related to a subject's heart rate, respiratory rate, and / or blood pressure are determined and processed based on analysis of pulse waveforms. do.

The device for determining the health condition of the subject comprises a control module and means for providing pulse wave data reflecting the subject's cardiac rhythm. Pulse wave data can be obtained. Some of the pulse wave data that characterizes multiple periods of the cardiac cycle are selected. Based on the pulse wave data of the pulse wave data, the pulse fluctuation, the respiratory rate fluctuation, and the heart rate fluctuation are determined. Correlation values are determined based on blood pressure variability, respiratory rate variability, heart rate variability, and corresponding reference values. The health status of the subject is determined based on the correlation value.

In another embodiment, the first and second indicators that characterize heart rate variability are determined based on some of the pulse wave data that characterize multiple periods of the cardiac cycle. The second indicator is different from the first indicator. Subject's health status is determined based on the first and second indicators. Determining the first index is to determine multiple breathing rate intervals based on some of the pulse wave data in the pulse wave data and to determine the first index based on multiple breathing time intervals. And include.

The disadvantage of this method is that it compares the statistical spectral parameters of a patient's heart rate variability (HRV) with the reference statistical spectral parameters of HRV from healthy patients and patients with impaired carbohydrate metabolism. Includes the fact that the presence of impaired carbohydrate metabolism in patients cannot be assessed due to parameter deviations.

For the closest analogs to the technical solution claimed, the applicant is exploring methods of heart rate variability that can be used for cardiovascular and nervous system functional diagnosis (July 20, 2002). See published RU2185090C1). Heart rate variability (HRV), that is, continuous variation in the duration of the electrocardiogram (ECG) RR interval with constant correction of cardiac activity according to the state of physical function, is an indicator of the effects of nerves on the heart. It is regarded.

This method measures height and weight, and pulse arterial blood pressure. An electrocardiogram is recorded. Fluctuation analysis is performed. The current heart rate per minute and the appropriate heart rate are determined. The stress level generated is calculated in conditional units. A qualitative estimation of the test results of the stress level that has occurred is performed. The type of cardiac rhythm is determined as normal, bradycardic, or tachyarrhythmia based on the difference between the current heart rate and the appropriate heart rate. The variation of the RR interval is determined from its coefficient of variation. One of nine possible variants of cardiac rhythm is determined (low, normal, or hypervariable bradycardic arrhythmia, normal arrhythmia, or tachyarrhythmia).

Mathematical Fourier transforms were used to identify three frequency ranges in which wavy changes in the heartbeat are usually observed. The first range (0.4-0.04 Hz, or high frequency) is clearly correlated with respiratory movement, during which there is an increase in parasympathetic activity (inhalation) or a decrease in the latter (expiration). The nature of the second range (0.04 to 0.015 Hz, or low frequency) remains unclear and is associated with sympathetic or parasympathetic effects. And finally, the third band (below 0.015 Hz or very low frequencies) is also thought to be associated with metabolic processes that do not correlate with any particular neural effect.

The disadvantage of this method is that it compares the statistical spectral parameters of a patient's heart rate variability (HRV) with the reference statistical spectral parameters of HRV from healthy patients and patients with impaired carbohydrate metabolism. Includes the fact that the presence of impaired carbohydrate metabolism in patients cannot be assessed due to parameter deviations.

WO2017 / 160186A1 WO2017009465 (A1) RU2185090C1

R.M. Baevsky, "ANALYSIS OF HEART RATE VARIABILITY WHEN USING DIFFERENT ELECTROCARDIOGRAPHIC SYSTEMS" Peter D. Welch, "The Use of Fast Fourier Transform for the Estimation of Power Spectra: A Method Based on Time Aver-aging Over Short, Modified Periodograms"

An object of the present invention is a computer-assisted method for non-invasive detection of impaired carbohydrate metabolism from heart rate variability (HRV), which provides statistical spectral parameters of HRV in patients and reference statistics of HRV from healthy patients. Assess the presence of impaired carbohydrate metabolism in patients by deviations in heart rate variability (HRV) parameters determined by comparison with specific spectral parameters and reference statistical spectral parameters of HRV in patients with impaired carbohydrate metabolism (CMD). To provide a method.

A second object of the invention is to provide a wearable autonomous device for non-invasive detection of impaired carbohydrate metabolism due to patient heart rate variability (HRV), which provides the patient's statistical spectral parameter HRV. To assess the presence of impaired carbohydrate metabolism by deviation of a patient's heart rate variability (HRV) parameters by comparing with the HRV reference statistical spectral parameters from healthy patients and the HRV reference statistical spectral parameters of ICD patients. It is configured in.

A third object of the present invention is to provide a computerized screening method for a population to identify individuals with exacerbated inborn errors of carbohydrate metabolism by heart rate variability (HRV).

The technical result of the present invention is to provide the possibility of detecting CMD from cardiac signals, as opposed to the standard for detecting inborn errors of carbohydrate metabolism (CMD) from blood tests, thereby providing a patient. The solution to the problem of large-scale screening of the population, which manifests itself in the effect of "selection", is simplified, i.e., providing computerized (non-invasive) separation of healthy patients from patients with CMD pathology. That is. This guarantees a 65-70% reduction in the number of blood tests.

This objective was achieved by providing a computer-aided method for the non-invasive detection of inborn errors of carbohydrate metabolism (CMD) from heart rate variability (HRV), which method is:

A step of acquiring at least one cardiac signal, wherein the signal is selected from the group consisting of an electrocardiogram (ECG) and a photoplethysmogram (PPG) of at least the first lead of the patient.

In at least one received cardiac signal, a sequence of timestamps of the appearance of R peaks in ECG, a sequence of timestamps of appearance of P peaks in ECG, a sequence of timestamps of appearance of T peaks in ECG, a sequence of systolic periods in PPG. A step to determine at least one sequence of time stamps selected from the group consisting of a sequence of time stamps of appearance and a sequence of time stamps of diastolic appearance in PPG.

With the step of determining the time interval between adjacent time stamps within at least one sequence of the resulting time stamps and obtaining at least one cardiac spacing diagram (CIG) that characterizes the patient's heart rate variability (HRV).

In order to obtain at least one transformed cardiac spacing diagram (TCIG) that characterizes the non-linear variation of the heartbeat, the steps to perform the transformation of at least one CIG obtained by removing the linear trend of the values,

To obtain at least one set of HRV statistical spectral parameters, the steps to calculate the HRV statistical spectral parameters for each of at least one TCIG,

It is a step to perform computer processing of at least one set of the obtained HRV statistical spectral parameters, during which the obtained HRV statistical spectral parameters are the HRV reference statistical spectral parameters of a healthy patient, and Steps and, which are compared with the reference statistical spectral parameters of the HRV of the CMD patient,

Inborn errors of glucose metabolism (CMD) involves making decisions about the presence of CMD in a patient according to the degree of proximity of the statistical spectral parameter in the patient's HRV to the reference value of the statistical spectral parameter in the patient's HRV.

Preferably, the method further comprises the step of removing the linear tendency of the CIG value by a zero-order least squares method (subtraction of arithmetic mean).

Preferably, the method further comprises removing the linear tendency of the CIG value by at least a first-order least squares method.

Preferably, the method further comprises the step of removing the linear trend of the CIG values by subtracting the smoothed KIG representing the moving average / median envelope.

Preferably, the method further comprises the step of smoothing the envelope at least once additionally by the moving average / median method.

Preferably, the method further comprises calculating the spectral parameters of HRV heart rate variability by a method selected from the group consisting of Fourier spectrum calculation, Welch's method, and Lomb-Scargle periodogram calculation.

Preferably, the method is a linear regression, logistic regression, discriminative method, neural network construction, decision tree trained in training samples of CMD and healthy patients in the feature space of HRV statistical spectral parameters. Further includes the step of making a decision on the existence of a CMD by applying at least one method from the group consisting of building a regression, performing clustering, and building a classification model.

According to a second aspect of the claimed invention, the object is achieved by providing a computer-aided method for non-invasive detection of an inborn error of glucose metabolism (CMD) from heart rate variability (HRV), the method. teeth,

Steps to obtain one cardiac spacing diagram (CIG) that reflects the presence of the patient's heart rate variability (HRV),

In order to obtain at least one transformed cardiac spacing diagram (TCIG) that characterizes the non-linear variation of the heartbeat, the steps to perform the transformation of at least one CIG obtained by removing the linear trend of the values,

To obtain at least one set of HRV statistical spectral parameters, the steps to calculate TCIG's HRV statistical spectral parameters, and

It is a step to perform computer processing of the obtained HRV statistical spectral parameters, during which the obtained HRV statistical spectral parameters are the reference statistical spectral parameters of the HRV of healthy patients, and the HRV of CMD patients. The steps and, which are compared with the reference statistical spectral parameters,

Inborn errors of glucose metabolism (CMD) involves making decisions about the presence of CMD in a patient according to the degree of proximity of the statistical spectral parameter in the patient's HRV to the reference value of the statistical spectral parameter in the patient's HRV.

According to a third aspect of the claimed invention, the objective is achieved by providing a wearable autonomous device for non-invasive detection of inborn errors of carbohydrate metabolism (CMD) from heart rate variability (HRV), the device. teeth,

A module for acquiring at least one cardiac signal, in which the signal is selected from the group consisting of an electrocardiogram (ECG) and a photoplethysmogram (PPG) of at least the first lead of the patient, as well as

It ’s a computing means,

In at least one received cardiac signal, a sequence of timestamps of the appearance of R peaks in ECG, a sequence of timestamps of appearance of P peaks in ECG, a sequence of timestamps of appearance of T peaks in ECG, a sequence of systolic periods in PPG. A module for determining at least one sequence of timestamps selected from the group consisting of sequences of occurrence timestamps and sequences of diastolic appearance timestamps in PPG.

With a module to determine the time interval between adjacent time stamps within at least one sequence of the resulting time stamps and to obtain at least one cardiac spacing diagram (CIG) that characterizes the patient's heart rate variability (HRV).

A module for performing at least one CIG transformation obtained by removing the linear trend of the values to obtain at least one transformed cardiac spacing diagram (TCIG) that characterizes the non-linear variation of the heartbeat,

A module for calculating HRV statistical spectral parameters for each of at least one TCIG to obtain at least one set of HRV statistical spectral parameters,

A module for performing computer processing on at least one set of the resulting HRV statistical spectral parameters, during which the resulting HRV statistical spectral parameters are the reference statistical spectral parameters of the HRV of a healthy patient. , And the module, which is compared with the reference statistical spectral parameters of the HRV of the CMD patient,

Compute with a module for making decisions about the presence of CMD in a patient according to the degree of closeness of the statistical spectral parameter of the patient's HRV to the reference value of the statistical spectral parameter of the patient's HRV in an inborn error of glucose metabolism (CMD). Equipped with ing means.

Preferably, in order to obtain at least one transformed core spacing diagram (TCIG), the module for performing the transformation of at least one CIG obtained is a linear trend of the values of the CIG by the zero-order least squares method. Is configured to remove.

Preferably, the module for performing the transformation of at least one obtained CIG is configured to eliminate the linear tendency of the CIG value by at least a first-order least squares method.

Preferably, the module for performing at least one CIG transformation obtained is such that the linear trend of the CIG values is removed by subtracting the smoothed KIG representing the moving average / median envelope. It is composed.

Preferably, the module for performing the conversion of at least one obtained CIG is further configured to smooth the envelope at least once by the moving average / median method.

Preferably, in order to obtain at least one set of HRV statistical spectral parameters, a module for calculating HRV statistical spectral parameters for each of at least one TCIG is a Fourier spectrum calculation, Welch's method, Lomb-. It is configured to calculate the spectral parameters of HRV heart rate variability by a method selected from the group consisting of the calculation of the Scargle periodogram.

Preferably, the module for making decisions about the presence of impaired carbohydrate metabolism (CMD) in the patient is a linear regression trained with training samples of CMD and healthy patients in the feature space of the statistical spectral parameters of HRV. By applying at least one method from the group consisting of logistic regression, discriminative methods, neural network construction, decision tree construction, clustering execution, and classification model construction, judgments regarding the presence of carbohydrate metabolism disorders Configured to do.

According to a fourth aspect of the claimed invention, the objective is achieved by providing a wearable autonomous device for non-invasive detection of inborn errors of carbohydrate metabolism (CMD) from heart rate variability (HRV), the device. teeth,

Sensors configured to measure the time interval between heartbeats and provide a cardiac interval diagram (CIG) that reflects the patient's heart rate variability (HRV), as well as

It ’s a computing means,

A module for performing at least one CIG transformation obtained by removing the linear trend of the values to obtain at least one transformed cardiac spacing diagram (TCIG) that characterizes the non-linear variation of the heartbeat,

A module for calculating HRV statistical spectral parameters for each of at least one TCIG to obtain at least one set of HRV statistical spectral parameters,

A module for performing computer processing on at least one set of the resulting HRV statistical spectral parameters, during which the resulting HRV statistical spectral parameters are the reference statistical spectral parameters of the HRV of a healthy patient. , And the module, which is compared with the reference statistical spectral parameters of the HRV of the CMD patient,

Compute with a module for making decisions about the presence of CMD in a patient according to the degree of closeness of the statistical spectral parameter of the patient's HRV to the reference value of the statistical spectral parameter of the patient's HRV in an inborn error of glucose metabolism (CMD). Equipped with ing means.

According to a fifth aspect of the claimed invention, the object is computerized to screen a population to identify individuals with signs of impaired carbohydrate metabolism based on the patient's heart rate variability (HRV). Achieved by providing a method, this method is

Steps to set screening sensitivity and specificity goals,

A step in performing screening preparation, for which ECG has a minimum required number of ECGs (ECGs) and photoplethysmograms (PPGs) of the patients under examination, and an MCR symbol for each observed patient in the group. And when the PPG threshold exceeds this, the presence of CMD for the patients in the group is recognized, step and

Including a step of performing a population screening to identify a person with a characteristic of an inborn error of carbohydrate metabolism, wherein the method according to any one of paragraphs 1 and 8 is carried out.

Improving the quality of care for patients with various illnesses at an acceptable cost to develop a health system is a very urgent task. The most important direction is the development of various methods of telemedicine for patient diagnosis, screening, monitoring, and follow-up, regardless of patient location. One of these diseases is diabetes.

The method and device can be effectively used to construct a two-step screening of the population to identify patients with inborn errors of carbohydrate metabolism (CMD). In the first stage of screening, patients suspected of having CMD are identified by using the method and device outside the medical facility. Patients identified in the first stage are then sent to a medical facility for further testing. In the second stage, the medical facility makes a final decision on the presence or absence of CMD according to endocrinological criteria for the patients identified in the first stage.

Hereinafter, the present invention will be described by reference to the accompanying drawings by description of preferred embodiments thereof.

FIG. 6 shows a sequence of steps of a method for non-invasive detection of an inborn error of glucose metabolism (CMD) from heart rate variability (HRV) according to the present invention, in which a cardiac signal, which is an electrocardiogram (ECG), is acquired. It is an electrocardiogram obtained from a patient and shows the sequence of time stamps of the appearance of the R peak in ECG. It is an electrocardiogram taken from a patient and shows the sequence of time stamps of the appearance of the P peak in ECG. It is an electrocardiogram taken from a patient and shows the sequence of time stamps of the appearance of the T peak in ECG. FIG. 5 shows a cardiac spacing diagram (CIG) that characterizes a patient's heart rate variability (HRV). FIG. 5 shows a transformed cardiac spacing map (PCIG) that characterizes a patient's heart rate variability (HRV). The figure showing the sequence of steps of the method for non-invasive detection of inborn errors of carbohydrate metabolism (CMD) from heart rate variability (HRV) according to the present invention, and the acquisition of a cardiac signal which is a photoplethysmogram (PPG). Is. FIG. 6 shows a photoplethysmogram obtained from a patient showing a sequence of time stamps of systolic appearance. FIG. 5 shows a photoplethysmogram taken from a patient showing a sequence of time stamps of the appearance of diastole. The present invention presents a sequence of steps in a method for non-invasive detection of inborn errors of carbohydrate metabolism (CMD) from heart rate variability (HRV), where cardiac signals are taken from the patient's electrocardiogram (ECG) and photoplethysmogram (ECG) and photoplethysmogram. It is a figure which is PPG). It is a figure which shows the sequence of the steps of the method for the non-invasive detection of the disorder of carbohydrate metabolism (CMD) from the heart rate variability (HRV) according to this invention, and is the figure which obtains the cardiac signal which is the cardiac spacing diagram (CIG). .. FIG. 6 is a block diagram of an autonomous device for non-invasive detection of an inborn error of glucose metabolism (CMD) due to a patient's heart rate variability (HRV), which is a first embodiment. FIG. 2 is a block diagram of an autonomous device for non-invasive detection of an inborn error of glucose metabolism (CMD) due to a patient's heart rate variability (HRV), which is a second embodiment.

In a healthy person, the time interval from the beginning of one heartbeat cycle to the beginning of another heartbeat is not the same and is constantly changing. This phenomenon is called heart rate variability (HRV). In addition, the invariance of the intervals between cardiac cycles is within a certain mean, which is optimal for certain considered functional states of the body.

Currently, the definition of HRV is recognized as the most beneficial non-invasive method for quantitatively assessing the autonomous (nutritional) regulation of heart rate. By analyzing HRV, it is possible not only to evaluate the functional state of the body, but also to monitor its dynamics.

According to the present invention, a computerized method for non-invasive detection of an inborn error of glucose metabolism due to a patient's heart rate variability (HRV) is performed as follows.

At least one cardiac signal selected from the group consisting of an electrocardiogram (ECG) and a photoplethysmogram (PPG) of at least the first lead of a patient is acquired.

Electrocardiograms and photoplethysmograms are obtained using individual ECGs, PPGs, or CIGs.

According to the first embodiment of the present invention, at least one cardiac signal is acquired (S1) (FIG. 1), i.e., at least the electrocardiogram (ECG) of the first lead.

The electrocardiogram is stored in the observed patient database (S2) and the patient's electrocardiogram is accumulated.

In at least one ECG obtained, at least one sequence of time stamps was determined (S3), which sequence was the sequence of time stamps A1, A2, A3, An, the appearance of R peaks in the ECG (FIG. 2a). Select from the group consisting of the time stamps B1, B2, B3, Bn of the appearance of the P peak in ECG (Fig. 2b) and the time stamps of C1, C2, C3, Cn of the appearance of the T peak in ECG (Fig. 2c). Will be done.

In at least one sequence of the resulting timestamps, the time interval between adjacent timestamps A1, A2, A3, An or B1, B2, B3, Bn or C1, C2, C3, Cn S4 is determined (S4). (Figure 1), at least one heart rate diagram (CIG) (Figure 3) that characterizes the patient's heart rate variability (HRV) is obtained.

The resulting transformation of at least one CIG removes the linear tendency of the values to obtain at least one transformed cardiac spacing diagram (PCIG) (FIG. 4) that characterizes the non-linear variation of the heartbeat (S5) ( It is executed by Figure 1). The linear trend of values is shown by the thick line in Figure 4.

Is the linear tendency of the CIG value eliminated by a zero-order least squares method (subtraction of the arithmetic mean) (see, for example, RM Baevsky's ANALYSIS OF HEART RATE VARIABILITY WHEN USING DIFFERENT ELECTROCARDIOGRAPHIC SYSTEMS, or The linear tendency of the CIG value is removed by at least the least squares method of the first order. It is also possible to remove the linear tendency of the CIG value by subtracting the smoothed KIG, which is a smoothed KIG. Represents a moving average / median envelope. The envelope is then smoothed at least once by the moving average / median method. The smoothing is well known in the art.

For each of at least one PKIG of statistical spectral HRV parameters, it is necessary to obtain at least one set of statistical spectral HRV parameters (Fig. 4).

HRV statistical spectral parameters are calculated for each of at least one TCIG to obtain at least one set of HRV statistical spectral parameters (S6) (FIG. 1).

HRV heart rate variability spectral parameters include Fourier spectrum calculations (see, for example, RM Baevsky above), Welch's Method (Peter D. Welch's "The Use of Fast Fourier Transform for the Optimization of Power Spectra: A Method Based on". Time Aver-aging Over Short, Modified Periodograms), calculated by a method selected from the group consisting of Lomb-Scargle periodogram calculations (see, for example, RM Baevsky).

Computer processing (S7) of the obtained HRV statistical spectral parameters is performed, during which the obtained HRV statistical spectral parameters are the reference statistical spectral parameters of the HRV of healthy patients and the HRV of CMD patients. Reference Statistical spectral parameters are compared.

At least a piecewise linear division of the HRV statistical spectral parameter space into at least two reference regions with and without CMD is performed.

The determination of the presence or absence of CMD features in a patient is based on the fact that the vector of statistical spectral parameters of the patient's HRV falls into the CMD / non-CMD reference region.

The reference statistical spectral parameters of the HRV from healthy patients and the reference statistical spectral parameters of the HRV of CMD patients are pre-populated in the computer database.

Then, the presence of a patient's carbohydrate metabolism disorder is determined according to the degree of proximity of the patient's HRV statistical spectral parameter to the reference value of the patient's HRV statistical spectral parameter.

The above judgment is based on linear regression, logistic regression, discriminative method, neural network construction, decision tree construction, trained in training samples of CMD and healthy patients in the feature space of HRV statistical spectral parameters. It is done by applying at least one method from the group consisting of performing clustering, and building a classification model.

According to the second embodiment of the present invention, at least one cardiac signal is acquired in order to carry out a computerized method for non-invasive detection of an inborn error of glucose metabolism due to a patient's heart rate variability (HRV). (S8) (Fig. 5), which is the patient's photoplethysmogram (PPG).

Photoplethysmograms are stored in the database of observed patients (S9) and photoplethysmograms are accumulated.

At least one sequence of systolic appearance time stamps D1, D2, D3, Dn (Fig. 6a) in PPG and diastolic appearance time stamps E1, E2, E3, in at least one received signal. The sequence of En is determined (S10) (Fig. 6b), where the Y-axis is the signal amplitude PPG with sensor data (the percentage of reflected light from the IR LED backlight).

In the resulting sequence of at least one time stamp, the time interval between adjacent time stamps is determined to give at least one cardiac spacing diagram (CIG) that characterizes the patient's heart rate variability (HRV) (S11) (Figure). 3).

The resulting transformation of at least one CIG removes the linear tendency of the values (FIG. 5) to obtain at least one transformed cardiac spacing diagram (TCIG) (FIG. 4) that characterizes the non-linear variation of the heartbeat (FIG. 5). It is executed by S12).

Is the linear tendency of the CIG value eliminated by a zero-order least squares method (subtraction of the arithmetic mean) (see, for example, RM Baevsky's ANALYSIS OF HEART RATE VARIABILITY WHEN USING DIFFERENT ELECTROCARDIOGRAPHIC SYSTEMS, or The linear tendency of the CIG value is removed by at least the least squares method of the first order. It is also possible to remove the linear tendency of the CIG value by subtracting the smoothed KIG, which is a smoothed KIG. Represents a moving mean / median envelope. The envelope is then smoothed at least once by the moving mean / median method.

For each of the at least one PKIG, the statistical spectral HRV parameters are calculated to obtain at least one set of statistical spectral HRV parameters (S13) (Fig. 5).

Spectral parameters of HRV heart rate variability are calculated by a method selected from the group consisting of Fourier spectrum calculations, Welch's method, and Lomb-Scargle periodogram calculations.

Computer processing (S14) of at least one set of the resulting HRV statistical spectral parameters is performed, where the resulting HRV statistical spectral parameters are the HRV reference statistical spectral parameters from healthy patients and Compared to the reference statistical spectral parameters of the HRV of the CMD patient. Specific reference statistical spectral parameters for HRV from healthy patients and HRV reference statistical spectral parameters for CMD patients are pre-populated in the computer database.

Judgments regarding the presence of a patient's CMD are made according to the degree of proximity of the patient's HRV statistical spectral parameters to the reference value of the patient's HRV statistical spectral parameters for inborn errors of glucose metabolism (CMD).

Linear regression, logistic regression, discriminative methods, neural network construction, decision tree construction, clustering execution, trained in CMD and healthy patient training samples in the HRV statistical spectral parameter feature space. And by applying at least one method from the group consisting of building classification models, decisions about the existence of ICE are made.

According to a third embodiment of the invention (FIG. 7), cardiac signals are acquired to implement a computerized method for non-invasive detection of inborn errors of glucose metabolism due to a patient's heart rate variability (HRV). (S15), which is the ECG and photoplethysmogram (PPG) of at least the first lead of the patient.

The electrocardiogram is stored in the database of the observed patients (S16) and the electrocardiogram is accumulated.

Photoplethysmograms are stored in the database of observed patients (S17) and photoplethysmograms are accumulated.

Similar to the first embodiment of the method, at least one sequence of time stamps (FIG. 2) is determined in the received ECG signal (S18) and the sequence is the time stamp of the appearance of the R peak in the ECG. Selected from the group consisting of the sequence of, the time stamp sequence of the appearance of the P peak in the ECG, and the time stamp sequence of the appearance of the T peak in the ECG (Fig. 3), which characterizes the patient's heart rate variability (HRV). Obtain at least one center spacing diagram (CIG) (Figure 3).

Similar to the second embodiment of the method, the resulting photoplethysmogram signal determines the sequence of time stamps of the appearance of systole in PPG and the sequence of time stamps of appearance of diastole in PPG. (S19).

In the resulting sequence of time stamps, the time interval between adjacent time stamps is determined (S20) to obtain a cardiac spacing diagram (CIG) that characterizes the patient's heart rate variability (HRV).

From what is obtained, for example, at least one option is selected from five options for calculating CIG. The selection of a sequence or a combination thereof is performed based on an experimental assessment of the sensitivity and specificity of one of the resulting variants.

The resulting transformation of at least one CIG is performed by removing the linear trend of the values (S21) to obtain at least one transformed cardiac spacing diagram (PCIG) that characterizes the non-linear variation of the heartbeat (S21). Figure 3).

Similar to the above, the linear tendency of CIG is removed by the zero-order least squares method (subtraction of arithmetic mean). Alternatively, the linear tendency of RIG is removed by at least a first-order least squares method. It is also possible to remove the linear RIG trend by subtracting the smoothed KIG, which represents the moving average / median envelope. The moving average / median method then further smoothes the envelope at least once.

Statistical spectral HRV parameters are calculated for each PKIG to obtain a set of statistical spectral HRV parameters (S22). Spectral parameters of HRV heart rate variability are calculated by a method selected from the group consisting of Fourier spectrum calculations, Welch's method, and Lomb-Scargle periodogram calculations.

Similar to the second embodiment, it is a step of performing computer processing of at least one set of the obtained statistical spectral parameters of the HRV, during which the obtained statistical spectral parameters of the HRV are of a healthy patient. It is compared with the reference statistical spectral parameters of the HRV and the reference statistical spectral parameters of the HRV of the CMD patient. The particular reference statistical spectral parameters of the HRV from a healthy patient and the reference statistical spectral parameters of the HRV of a CMD patient are pre-populated in a computer database.

Judgments regarding the presence of CMD in patients are made according to the degree of closeness of the statistical spectral parameters of the patient's HRV to the reference value of the statistical spectral parameters of the patient's HRV in an inborn error of glucose metabolism (CMD) (S23).

Linear regression, logistic regression, discriminative methods, neural network construction, decision tree construction, clustering execution, trained in CMD and healthy patient training samples in the HRV statistical spectral parameter feature space. And by applying at least one method from the group consisting of building a classification model, the above determination of the existence of CMD is made.

In each implementation, the judgment is made by the inspector for proven sensitivity and specificity.

According to a fourth embodiment of the present invention (FIG. 8), the presence of a patient's heart rate variability (HRV) is used to implement a method for non-invasive detection of carbohydrate metabolism disorders by the patient's heart rate variability (HRV). A single heart spacing diagram (CIG) that reflects is obtained (S24) (Figure 3). Removal is done in the cloud using a local device or from ECG or PPG.

The CIG storage and storage steps are performed (S25).

Similar to the method above, the resulting CIG transformation (S26) removes the linear tendency of the values to obtain a transformed cardiac spacing diagram (PCIG) (FIG. 4) that characterizes the non-linear variation of the heartbeat. Is executed by.

Statistical spectral HRV parameters are calculated for PKIG to obtain a set of statistical spectral HRV parameters (S27).

In addition, at least one set of obtained HRV statistical spectral parameters is subjected to computer processing (S28), during which the obtained HRV statistical spectral parameters are the reference statistical spectral parameters of the HRV of healthy patients. , And CMD patients' HRVs are compared with reference statistical spectral parameters. The particular reference statistical spectral parameters of the HRV from a healthy patient and the reference statistical spectral parameters of the HRV of a CMD patient are pre-populated in a computer database.

Judgments regarding the presence of a patient's CMD are made according to the degree of proximity of the patient's HRV statistical spectral parameters to the reference value of the patient's HRV statistical spectral parameters for inborn errors of glucose metabolism (CMD).

Similar to the above method, the linear tendency of CIG is removed by the least squares zero-order or first-order least-squares method. Alternatively, the linear trend RAG is removed by subtracting the smoothed KIG, which represents the moving average / median envelope. The moving average / median method further smoothes the envelope at least once.

Spectral parameters of HRV heart rate variability are calculated by a method selected from the group consisting of Fourier spectrum calculations, Welch's method, and Lomb-Scargle periodogram calculations.

Linear regression, logistic regression, discriminative methods, neural network construction, decision tree construction, clustering execution, trained in CMD and healthy patient training samples in the HRV statistical spectral parameter feature space. And by applying at least one method from the group consisting of building classification models, decisions about the existence of CMD are made.

According to the present invention, a wearable autonomous device for non-invasive detection of an inborn error of glucose metabolism due to a patient's heart rate variability (HRV) is also proposed (Fig. 9).

The device is a module 29 for acquiring at least one cardiac signal, the signal being selected from the group consisting of an electrocardiogram (ECG) and a photoplethysmogram (PPG) of at least the first lead of the patient. It has a module 29 and a computing means 30.

The computing means 30 are connected in series,

In at least one received cardiac signal, a sequence of timestamps of the appearance of R peaks in ECG, a sequence of timestamps of appearance of P peaks in ECG, a sequence of timestamps of appearance of T peaks in ECG, a sequence of systolic periods in PPG. Module 31 for determining at least one sequence of time stamps selected from the group consisting of sequences of occurrence time stamps and sequences of diastolic appearance time stamps in PPG.

Module 32 to determine the time interval between adjacent time stamps within at least one sequence of the resulting time stamps and to obtain at least one cardiac spacing diagram (CIG) that characterizes the patient's heart rate variability (HRV).

With Module 33 to perform at least one CIG transformation obtained by removing the linear trend of the values to obtain at least one transformed cardiac spacing diagram (TCIG) that characterizes the non-linear variation of the heartbeat. ,

Module 34 for calculating HRV statistical spectral parameters for each of at least one TCIG to obtain at least one set of HRV statistical spectral parameters, and

Modul 35 for performing computer processing on at least one set of the resulting HRV statistical spectrum parameters, during which the obtained HRV statistical spectrum parameters are the reference statistical spectrum of the HRV of a healthy patient. With module 35, which is compared with the parameters, and the reference statistical spectral parameters of the HRV of the CMD patient,

Modul 35 will also make a judgment regarding the presence of CMD in a patient according to the degree of proximity of the statistical spectral parameter of the patient's HRV to the reference value of the statistical spectral parameter of the HRV in the patient with an inborn error of glucose metabolism (CMD). It is composed.

To obtain at least one transformed core spacing diagram (TCIG), Module 33 for performing the resulting transformation of at least one CIG removes the linear tendency of the CIG values by the zero-order least squares method. It is configured to do.

Module 33 for performing at least one CIG transformation obtained is also configured to eliminate the linear tendency of CIG values by at least a first-order least squares method.

Module 33 for performing at least one CIG transformation obtained is configured to eliminate the linear trend of the CIG values by subtracting the smoothed KIG representing the moving average / median envelope. Ru. Module 33 for performing at least one CIG transformation obtained is further configured to smooth the envelope at least once by the moving average / median method.

To obtain at least one set of HRV statistical spectral parameters, Module 34 for calculating HRV statistical spectral parameters for each of at least one TCIG is a Fourier spectrum calculation, Welch's method, Lomb-Scargle period. It is configured to calculate the spectral parameters of HRV heart rate variability by a method selected from the group consisting of gram calculations.

Module 35 for making decisions about the presence of impaired carbohydrate metabolism (CMD) in patients is a linear regression, logistic trained with training samples of CMD and healthy patients in the feature space of statistical spectral parameters of HRV. Make decisions about the presence of carbohydrate metabolism disorders by applying at least one method from the group consisting of regression, discriminative methods, building neural networks, building decision trees, performing clustering, and building classification models. It is composed of.

According to a second embodiment, a wearable autonomous device for non-invasive detection of carbohydrate metabolism disorders (CMD) from heart rate variability (HRV) measures the time interval between heartbeats and the patient's heart rate variability ( It comprises a sensor 36 configured to provide a heart rate diagram (CIG) that reflects HRV). The heart rate monitor can be used as a particular sensor 36.

The devices are also connected in series,

With Module 38 to perform at least one CIG transformation obtained by removing the linear trend of the values to obtain at least one transformed cardiac spacing diagram (TCIG) that characterizes the non-linear variation of the heartbeat. ,

Module 39 for calculating HRV statistical spectral parameters for each of at least one TCIG to obtain at least one set of HRV statistical spectral parameters, and

Module 40 for performing computing processing on at least one set of the resulting HRV statistical spectral parameters, during which the obtained HRV statistical spectral parameters are the reference statistical spectrum of the HRV of a healthy patient. The parameters, and the computing means 37 with the module 40, which is compared with the reference statistical spectral parameters of the HRV of the CMD patient,

The module 40 will also make a judgment regarding the presence of CMD in the patient according to the degree of proximity of the statistical spectral parameter of the patient's HRV to the reference value of the statistical spectral parameter of HRV in the patient with carbohydrate metabolism disorder (CMD). It is composed.

A computerized method for screening a population to identify individuals with features of inborn errors of carbohydrate metabolism based on the patient's heart rate variability (HRV) is performed in the following manner.

Screening sensitivity and specificity goals are set.

Preparations for screening are performed, for which the minimum required number of ECGs and photoplethysmograms (PPGs) of the patients under test, and the ECG and PPG with the MCR symbol per observed patient in the group. When the threshold is exceeded, the presence of CMD for the patients in the group is recognized.

Population screening is performed to identify individuals with impaired carbohydrate metabolism, for which one of the above methods is performed.

In the first stage of screening, this method is used to identify patients with suspected inborn errors of carbohydrate metabolism and send them to a medical institution for further testing.

In the second phase, a nationally recommended medical institution makes the final decision on the presence or absence of inborn errors of glucose metabolism in patients who have undergone the first phase of screening.

Computerized method for detecting carbohydrate metabolism disorders by patient heart rate variability (HRV) to establish a large two-step screening of the population to identify patients with carbohydrate metabolism disorders (HRV) And wearable autonomous devices can be used. In the first stage of screening, the method is used to identify patients with suspected inborn errors of carbohydrate metabolism, who are sent to medical institutions for further testing. In the second stage, a medical professional (a medical institution expert based on national recommendations) makes the final decision on the presence or absence of an inborn error of glucose metabolism in patients who have undergone the first stage of screening.

[Example 1]
(Acquisition of ECG)
Patient A, 63 years old. Diagnosis: Type 2 diabetes for 25 years. No associated chronic illness. The general condition according to the test results is good.

A wearable autonomous device was attached to the patient's wrist for non-invasive detection of carbohydrate metabolism disorders.

At least the first ECG read cardiac signal (ECG) was obtained 5-20 times a week.

Accumulation of multiple electrocardiogram (ECG) sequences of at least the first read marked with the patient identifier in the database was performed.

At least one sequence of time stamps is determined in the received signal, the time stamp sequence of the appearance of the R peak in the ECG, the time stamp sequence of the appearance of the P peak in the ECG, and the time stamp of the appearance of the T peak in the ECG. It is selected from the group consisting of sequences.

In the resulting sequence of time stamps, the time interval between adjacent time stamps was determined to obtain at least one cardiac spacing diagram (CIG) that characterizes the patient's heart rate variability (HRV).

The conversion of at least one obtained CIG was performed by removing the linear tendency of the values to obtain at least one transformed cardiac spacing diagram (TCIG) that characterizes the non-linear variation in heart rate. Calculations were then performed for each of the TCIG statistical spectral parameters of the HRV to obtain at least one set of statistical spectral parameters for the HRV.

Computer processing of the set of obtained HRV statistical spectral parameters is then performed, while the resulting HRV statistical spectral parameters are HRV reference statistical spectral parameters from healthy patients and CMD patient HRV. It was compared with the reference statistical spectral parameters of.

As a result, all ECGs were subdivided into two classes, one with CMD features and one without CMD features.

Then, a judgment was made regarding the presence of a patient's carbohydrate metabolism disorder according to the degree of proximity of the patient's HRV statistical spectral parameter to the reference value of the patient's HRV statistical spectral parameter.

[Example 2]
(Acquisition of PPG)
Patient B, 63 years old. Diagnosis: Diabetes has not been diagnosed. No associated chronic illness. The general condition according to the test results is good.

A wearable autonomous device was attached to the patient's wrist for non-invasive detection of carbohydrate metabolism disorders.

At least one cardiac signal, which is the patient's photoplethysmogram (PPG), is acquired and selected from the group consisting of a sequence of time stamps of systolic appearance in PPG and a sequence of time stamps of diastolic appearance in PPG. A sequence of at least one time stamp was determined for at least one signal received.

In at least one sequence of the resulting time stamps, the time interval between adjacent time stamps was determined to obtain a cardiac spacing diagram (CIG) that characterizes the patient's heart rate variability (HRV). The resulting CIG was transformed by removing the linear tendency of the values to obtain a transformed cardiac spacing map (TCIG) that characterizes nonlinear heart rate variability.

To obtain a set of HRV statistical spectral parameters, a TCIG calculation of the HRV statistical spectral parameters was performed. Computer processing of the resulting set of HRV statistical spectral parameters was performed and the resulting HRV statistical spectral parameters were HRV reference statistical spectral parameters from healthy patients and HRV reference statistical parameters from CMD patients. Compared with spectral parameters.

Judgments were made regarding the presence of inborn errors of glucose metabolism in patients according to the degree of proximity of the patient's HRV statistical spectral parameters to the reference values of the ICR patient's HRV statistical spectral parameters.

[Example 3]
(Acquisition of CIG)
Patient B, 50 years old. Diagnosis: Diabetes has not been diagnosed. No associated chronic illness. The general condition according to the test results is good.

A wearable autonomous device was attached to the patient's wrist for non-invasive detection of carbohydrate metabolism disorders.

A heart rate monitor was used to obtain a single cardiac spacing map (CIG) that reflected the presence of the patient's heart rate variability (HRV). The resulting CIG was transformed by removing the linear tendency of the values to obtain a transformed cardiac spacing map (TCIG) that characterizes nonlinear heart rate variability. To obtain a set of HRV statistical spectral parameters, a TCIG calculation of the HRV statistical spectral parameters was performed.

The resulting set of HRV statistical spectral parameters is then processed and the resulting HRV statistical spectral parameters are the HRV reference statistical spectrum parameters from healthy patients and the CMD patient HRV reference statistical spectrum. Compared with the parameters.

Judgments were made regarding the presence of inborn errors of glucose metabolism in patients according to the degree of proximity of the patient's HRV statistical spectral parameters to the reference values of the patient's HRV statistical spectral parameters.

29 modules
30 Computing means
31 modules
32 modules
33 modules
34 modules
35 modules
36 sensor
37 Computing means
38 modules
39 modules
40 modules

Claims (29)

A computer-aided method for non-invasive detection of inborn errors of glucose metabolism (CMD) from heart rate variability (HRV).
A step of acquiring at least one cardiac signal, wherein the cardiac signal is selected from the group consisting of an electrocardiogram (ECG) and a photoplethysmogram (PPG) of at least the first lead of the patient.
In the received at least one cardiac signal, a sequence of time stamps of appearance of R peak in the ECG, a sequence of time stamps of appearance of P peak in the ECG, a sequence of time stamps of appearance of T peak in the ECG, said. A step of determining at least one sequence of time stamps selected from the group consisting of a sequence of time stamps of the appearance of systole in the PPG and a sequence of time stamps of the appearance of diastole in the PPG.
A step of determining the time interval between adjacent time stamps within at least one sequence of the resulting time stamps to obtain at least one cardiac spacing diagram (CIG) that characterizes the patient's heart rate variability (HRV).
In order to obtain at least one transformed cardiac spacing diagram (TCIG) that characterizes the non-linear variation of the heartbeat, the step of performing the transformation of at least one obtained CIG by removing the linear tendency of the values,
To obtain at least one set of HRV statistical spectral parameters, the step of calculating the HRV statistical spectral parameters for each of the at least one TCIG,
A step of performing computer processing of at least one set of the obtained HRV statistical spectral parameters, during which the obtained HRV statistical spectral parameters are the reference statistical spectral parameters of the HRV of a healthy patient. , And the steps and, which are compared with the reference statistical spectral parameters of the HRV of the CMD patient.
A step of making a determination regarding the presence of the CMD of the patient according to the degree of proximity of the statistical spectral parameter of the patient's HRV to the reference value of the statistical spectral parameter of the HRV of the patient with an inborn error of glucose metabolism (CMD). Method. The method of claim 1, wherein the linear tendency of the value of the CIG is removed by a zero-order least squares method (subtraction of the arithmetic mean). The method of claim 1, wherein the linear tendency of the value of the CIG is removed by at least a first-order least squares method. The method of claim 1, wherein the linear trend of the CIG value is removed by subtracting a smoothed KIG representing a moving average / median envelope. The method of claim 4, wherein the envelope is additionally smoothed at least once by the moving average / median method. The method of claim 1, wherein the spectral parameters of HRV heart rate variability are calculated by a method selected from the group consisting of a Fourier spectrum calculation, a Welch method, and a Lomb-Scargle periodogram calculation. Linear regression, logistic regression, discriminative methods, neural network construction, decision tree construction, clustering execution, trained in CMD and healthy patient training samples in the HRV statistical spectral parameter feature space. The method according to claim 1, wherein the determination regarding the existence of the CMD is made by applying at least one method from the group consisting of the construction of the classification model and the construction of the classification model. A computer-aided method for non-invasive detection of inborn errors of glucose metabolism (CMD) from heart rate variability (HRV).
Steps to obtain one cardiac spacing diagram (CIG) that reflects the presence of the patient's heart rate variability (HRV),
In order to obtain at least one transformed cardiac spacing diagram (TCIG) that characterizes the non-linear variation of the heartbeat, the step of performing the transformation of at least one CIG obtained by removing the linear tendency of the values,
To obtain at least one set of HRV statistical spectral parameters, the step of calculating the TCIG HRV statistical spectral parameters,
It is a step of performing computer processing of the obtained HRV statistical spectral parameters, during which the obtained HRV statistical spectral parameters are the HRV reference statistical spectral parameters of a healthy patient and the CMD patient. Steps and, compared to HRV reference statistical spectral parameters,
A step of making a determination regarding the presence of the CMD of the patient according to the degree of proximity of the statistical spectral parameter of the patient's HRV to the reference value of the statistical spectral parameter of the HRV of the patient with an inborn error of glucose metabolism (CMD). Method. The method of claim 8, wherein the linear tendency of the value of the CIG is removed by a zero-order least squares method (subtraction of the arithmetic mean). The method of claim 8, wherein the linear tendency of the value of the CIG is removed by at least a first-order least squares method. 8. The method of claim 8, wherein the linear trend of the CIG value is removed by subtracting a smoothed KIG representing a moving average / median envelope. 11. The method of claim 11, wherein the envelope is additionally smoothed at least once by the moving average / median method. The method of claim 8, wherein the spectral parameters of HRV heart rate variability are calculated by a method selected from the group consisting of calculation of Fourier spectrum, Welch's method, calculation of Lomb-Scargle periodogram. Linear regression, logistic regression, discriminative methods, neural network construction, decision tree construction, clustering execution, trained in CMD and healthy patient training samples in the HRV statistical spectral parameter feature space. The method of claim 8, wherein the determination regarding the existence of the CMD is made by applying at least one method from the group consisting of the construction of the classification model and the construction of the classification model. A wearable autonomous device for non-invasive detection of inborn errors of glucose metabolism (CMD) from heart rate variability (HRV).
A module for acquiring at least one cardiac signal, wherein the cardiac signal is selected from the group consisting of an electrocardiogram (ECG) and a photoplethysmogram (PPG) of at least the first lead of a patient. As well as a means of computing,
In the received at least one signal, the sequence of the time stamp of the appearance of the R peak in the ECG, the sequence of the time stamp of the appearance of the P peak in the ECG, the sequence of the time stamp of the appearance of the T peak in the ECG, the sequence of the PPG. A module for determining at least one sequence of time stamps selected from the group consisting of a sequence of time stamps of the appearance of systole in the PPG and a sequence of time stamps of the appearance of diastole in the PPG.
A module for determining the time interval between adjacent time stamps within at least one sequence of the resulting time stamps and obtaining at least one cardiac spacing diagram (CIG) that characterizes the patient's heart rate variability (HRV). When,
With a module to perform the transformation of at least one CIG obtained above by removing the linear tendency of the values to obtain at least one transformed cardiac spacing diagram (TCIG) that characterizes the non-linear variation of the heartbeat. ,
To obtain at least one set of HRV statistical spectral parameters, a module for calculating HRV statistical spectral parameters for each of the at least one TCIG, and
A module for performing computer processing of at least one set of the obtained HRV statistical spectral parameters, while the obtained HRV statistical spectral parameters are the reference statistical of the HRV of a healthy patient. Modules and modules that are compared with spectral parameters, and reference statistical spectral parameters of the HRV of CMD patients.
A module for making a judgment on the presence of the CMD of the patient according to the degree of proximity of the statistical spectral parameter of the patient's HRV to the reference value of the statistical spectral parameter of the HRV of the patient with an inborn error of glucose metabolism (CMD). Wearable autonomous device with computing means. In order to obtain at least one transformed core spacing diagram (TCIG), the module for performing the transformation of at least one obtained CIG is said to have the value of said CIG by a zero-order least squares method. The wearable autonomous device of claim 15, configured to eliminate linear trends. 15. Wearable autonomous device. The module for performing the transformation of at least one obtained CIG removes the linear tendency of the value of said CIG by subtracting the smoothed KIG representing the moving average / median envelope. 15. The wearable autonomous device of claim 15. 18. The module of claim 18 for performing the transformation of at least one of the resulting CIGs, further configured to smooth the envelope at least one additional time by the moving average / median method. Wearable autonomous device. To obtain at least one set of HRV statistical spectral parameters, the module for calculating HRV statistical spectral parameters for each of the at least one TCIG is the Fourier spectrum calculation, Welch's method, Lomb-Scargle. The wearable autonomous device of claim 15, configured to calculate spectral parameters of HRV heart rate variability by a method selected from the group consisting of periodogram calculations. A linear regression, in which the module for making a determination of the presence of a patient's impaired carbohydrate metabolism (CMD) was trained in a training sample of a CMD patient and a healthy patient in the feature space of the statistical spectral parameters of HRV. Judgment regarding the presence of impaired carbohydrate metabolism by applying at least one method from the group consisting of logistic regression, discriminative methods, neural network construction, decision tree construction, clustering execution, and classification model construction. The wearable autonomous device of claim 15, configured to do so. A wearable autonomous device for non-invasive detection of inborn errors of glucose metabolism (CMD) from heart rate variability (HRV).
A sensor, as well as a computing instrument, configured to measure the time interval between heartbeats and provide a cardiac interval diagram (CIG) that reflects the patient's heart rate variability (HRV).
A module for performing at least one CIG transformation obtained by removing the linear trend of the values to obtain at least one transformed cardiac spacing diagram (TCIG) that characterizes the non-linear variation of the heartbeat,
To obtain at least one set of HRV statistical spectral parameters, a module for calculating HRV statistical spectral parameters for each of the at least one TCIG, and
A module for performing computer processing of at least one set of the obtained HRV statistical spectral parameters, while the obtained HRV statistical spectral parameters are the reference statistical of the HRV of a healthy patient. Modules and modules that are compared with spectral parameters, and reference statistical spectral parameters of the HRV of CMD patients.
A module for making a judgment on the presence of the CMD of the patient according to the degree of proximity of the statistical spectral parameter of the patient's HRV to the reference value of the statistical spectral parameter of the HRV of the patient with an inborn error of glucose metabolism (CMD). Wearable autonomous device with computing means. In order to obtain at least one transformed core spacing diagram (TCIG), the module for performing the transformation of at least one obtained CIG is said to have the value of said CIG by a zero-order least squares method. 22. The wearable autonomous device of claim 22, configured to eliminate linear trends. 22. Wearable autonomous device. The module for performing the transformation of at least one obtained CIG removes the linear tendency of the value of said CIG by subtracting the smoothed KIG representing the moving average / median envelope. 22. The wearable autonomous device according to claim 22. 18. Claim 18, wherein the module for performing the transformation of at least one obtained CIG is further configured to smooth the envelope at least one additional time by the moving average / median method. Wearable autonomous device. To obtain at least one set of HRV statistical spectral parameters, the module for calculating HRV statistical spectral parameters for each of the at least one TCIG is the Fourier spectrum calculation, Welch's method, Lomb-Scargle. 22. The wearable autonomous device of claim 22, configured to calculate said spectral parameters of HRV heart rate variability by a method selected from the group consisting of periodogram calculations. A linear regression, in which the module for making a determination of the presence of a patient's impaired carbohydrate metabolism (CMD) was trained in a training sample of a CMD patient and a healthy patient in the feature space of the statistical spectral parameters of HRV. Judgment regarding the presence of impaired carbohydrate metabolism by applying at least one method from the group consisting of logistic regression, discriminative methods, neural network construction, decision tree construction, clustering execution, and classification model construction. 22. A wearable autonomous device configured to do so. A computerized method for screening the population to identify individuals with signs of inborn errors of glucose metabolism (CMD) based on the patient's heart rate variability (HRV).
At the stage of setting screening sensitivity and specificity goals,
The stage of performing screening preparations, for which the minimum required number of ECGs and photoplethysmograms (PPGs) of the patients under test, and the MCR symbol for each observed patient in the group are mentioned above. When the ECG and PPG thresholds are exceeded, the stage in which the presence of the CMD for the patient in the group is recognized, and
A step of performing the screening of the population to identify a person with a characteristic of an inborn error of carbohydrate metabolism, wherein the method according to any one of paragraphs 1 and 8 is performed. Including, method.

Images