RU2832523C1 - Wearable device, method and system for determining level of glycated haemoglobin - Google Patents

Abstract

FIELD: medicine.

SUBSTANCE: group of inventions relates to a wearable device and a method for determining the level of glycated haemoglobin. Biological tissues of the user are exposed to a light source emitting waves of at least two different lengths. At least one photoplethysmographic (PPG) sensor detects PPG signals from user tissues at said at least two different wavelengths in reflection mode. Detected PPG signals are converted from the time domain to the frequency domain, features of the PPG signals in the time and frequency domains are extracted and an event occurring in the daily activities of the user is determined, in which blood component values change. Identifying fast and slowly changing signs, respectively, characterizing fast and slowly changing blood components, forming a set of quickly and slowly changing signs of PPG-signals and determining the level of glycated haemoglobin in blood of the user.

EFFECT: permanent control of glycated haemoglobin level and possibility of prevention, detection and control of course of malfunctions in organism, in particular microvascular complications associated with diabetes.

49 cl, 16 dwg, 3 tbl

Description

Field of technology

[0001] The present invention relates to a wearable device and a system for non-invasive continuous, personal and/or on-demand monitoring of blood parameters, in particular glycated hemoglobin. In addition, the invention relates to a method for determining the level of glycated hemoglobin, implemented in wearable devices, such as modern smart watches and fitness bracelets, and in the said system.

Background of the invention

[0002] The average life expectancy in the world is gradually increasing due to the development of medicine. The increase in average life expectancy is influenced not only by the development of medicine, but also by the growing interest of the population in checking their health and managing it.

[0003] Many personal health devices have been developed that allow people to check their health without visiting a hospital. Adult diseases that can be monitored include cerebrovascular disease, hypertension, diabetes, liver disease, etc. For example, automatic sphygmomanometers are available in many places such as public buildings. As another example, to check for liver disease, a blood sample is taken to measure the liver function index. As another example, many diabetic patients frequently check their blood sugar levels with a compact tester without visiting a hospital. Diabetes is a group of metabolic diseases in which a person has high blood glucose (blood sugar) levels either because they do not produce enough insulin, because the body's cells do not respond properly to insulin, or both. Diabetic patients may check their blood sugar levels several times a day to ensure that they have sufficient insulin levels.

[0004] Blood parameters are markers of a person's health status and their timely monitoring makes it possible to prevent, timely detect and control the progression of a disease, as well as other disorders in the body. Therefore, constant monitoring of blood parameters is necessary, in particular the level of glycated hemoglobin.

[0005] Hemoglobin is the main respiratory pigment and the main component of red blood cells, performing important functions in the human body: transporting inhaled oxygen from the lungs to biological tissues and organs, and carbon dioxide from tissues and organs to the lungs, where it is exhaled. Hemoglobin is a complex protein belonging to the group of hemoproteins, the protein component of which is represented by globin, the non-protein component is a prosthetic group. The prosthetic group in the hemoglobin molecule is represented by four identical iron-porphyrin compounds called hemes. The heme molecule consists of porphyrin IX, bound to iron by two nitrogen atoms by covalent and two other nitrogen atoms by coordination bonds. The iron (II) atom is located in the center of the heme and gives the blood a characteristic red color, the degree of its oxidation does not change regardless of the addition or release of oxygen

[0006] Glycated (glycosylated) hemoglobin or glycohemoglobin (symbol: hemoglobin A1c, HbA1c) is a biochemical blood indicator that reflects the average blood sugar level over a long period (two to three months), in contrast to blood glucose measurement, which gives an idea of the blood glucose level only at the time of the test.

[0007] Glycated hemoglobin reflects the percentage of blood hemoglobin irreversibly bound to glucose molecules. Glycated hemoglobin is formed as a result of the Maillard reaction between hemoglobin and glucose in the blood, this glycosylation reaction is irreversible. An increase in blood glucose levels in diabetes mellitus significantly accelerates this reaction, which leads to an increase in the level of glycated hemoglobin in the blood. In other words, glycated hemoglobin is an integral indicator of glycemia for approximately two to three months. The higher the level of glycated hemoglobin, the more frequent the glycemia, that is, the elevated glucose level over the past three months and, accordingly, the greater the risk of developing complications of diabetes mellitus. Thus, the level of glycated hemoglobin today is a generally recognized indicator of the severity and degree of compensation of carbohydrate metabolism disorders.

[0008] The study of the level of glycated hemoglobin is used in endocrinology for long-term monitoring of the course of diabetes mellitus, control of the adequacy and effectiveness of the therapy. With a high level of glycated hemoglobin, it is necessary to correct the treatment (prescribe insulin therapy or oral hypoglycemic drugs) and diet therapy.

[0009] Hemoglobin is a tetrameric protein, the molecule of which is formed by different types of polypeptide chains, designated as α, β, γ, ε, ξ. The molecule consists of 2 polypeptide chains of two different types, each of which binds one heme, therefore, there are four oxygen binding sites in the hemoglobin molecule. In the blood of an adult, the main hemoglobin is hemoglobin A (HbA), it accounts for 96-98% of the total amount of hemoglobin in the body, and its structure is described by the formula α2β2. Minor components are hemoglobin A2 (HbA2), which accounts for approximately 2.5% of the total amount of hemoglobin and the structure of which is described by the formula α2δ2, and hemoglobin A3 (HbA3), the content of which is less than 1% of the total amount of hemoglobin. In addition, in the blood of an adult there is fetal hemoglobin F (HbF), which has the α2γ2 structure, but normally it makes up less than 1-1.5% of the total amount of hemoglobin in the blood of an adult. However, in the fetus this form of hemoglobin is dominant, the main one. Hemoglobin F has an increased affinity for oxygen and allows the relatively small volume of fetal blood to more effectively perform the functions of oxygen supply. However, HbF has less resistance to destruction and less stability in a physiologically wide range of pH and temperatures. During the last trimester of pregnancy and soon after birth, HbF is gradually, during the first few weeks or months of life, in parallel with the increase in blood volume, replaced by "adult" HbA, a less active oxygen transporter, but more resistant to destruction and more stable at various values of blood pH and body temperature. This substitution occurs due to a gradual decrease in the production of γ-chains of globin and a gradual increase in the synthesis of β-chains by maturing erythrocytes, the α-chain, common to all types of hemoglobin, contains 141 amino acid residues, β-, γ-, δ-chains contain 146 residues with a very similar amino acid sequence. Although the amino acid sequences of the α- and β-chains are different, they have almost identical tertiary spatial structures. The hemoglobin molecule has an almost regular shape of a sphere with a diameter of 55 Å, with four chains arranged in a tetrahedron.

[0010] In arterial blood, almost all (93-95%) hemoglobin HbA is bound to oxygen, i.e. is in the oxygenated form, and serves to transport oxygen. The remaining part of HbA (5-7%) is glycated hemoglobin and there are at least three of its variants: HbA1a, HbA1b, HbA1c, but only the HbA1c variant, which is hemoglobin A modified by the covalent attachment of glucose to it, quantitatively predominates (4-6% of total hemoglobin) and gives a closer correlation with the degree of diabetes mellitus.

[0011] Each red blood cell contains about 270 million hemoglobin molecules, which are bound to glucose in the blood plasma by a slow, non-enzymatic glycation reaction. Glycated hemoglobin is hemoglobin in which the carbohydrate residue is condensed with the N-terminal valine in the β-chain of the HbA molecule. Glycation is a non-enzymatic process and occurs slowly throughout the life of the red blood cell. The glycation process is irreversible and its rate is proportional to the level of glycemia. It has been established that it occurs through the formation of aldimine, a relatively unstable, "reversible" compound, then aldimine is converted by chemical transformation into a relatively stable, "irreversible" compound, ketoamine. The resulting ketoamine remains attached to the protein for its entire life. As noted, glycated hemoglobin is divided into subtypes depending on the attached carbohydrate (HbA1a, HbA1b, HbA1c). HbA1c contains one glucose molecule, it is the main subtype of glycated hemoglobin and makes up 70-90% of the glycated fraction, the rest is HbA1a (has form 1 - with fructose-1,6-biphosphate and form 2 - with glucose-6-phosphate) and HbA1b (carbohydrate residue, not precisely identified). HbA1c predominates in venous blood and is associated with the level of hyperglycemia. The formation and disappearance of hemoglobin A1c is a continuous process, since new young erythrocytes daily replace old dead ones in the bloodstream. The degree of saturation of HbA with glucose and the rate of their connection with each other are directly dependent on the average content of it in the bloodstream over the past 90-120 days, and this period is determined by the lifespan of hemoglobin-containing erythrocytes (red blood cells), which is approximately 120-125 days. Thus, the result of determining the level of glycated hemoglobin reflects the level of glycemia for a period of 60 to 120 days.

[0012] Normal HbA1c values are considered to be from 4% to 5.7%, values from 5.7% to 6.4% indicate a predisposition to diabetes. In diabetes, the HbA1c level is 6.5% and higher, which indicates a higher risk of developing retinopathy, nephropathy and other complications. The International Diabetes Federation recommends keeping the HbA1c level below 6.5%, and an HbA1c value exceeding 8% means that diabetes is poorly controlled and therapy should be changed.

[0013] To summarize the above, the HbA1c level is proportional to the average blood glucose level over the past 6-12 weeks, and constant monitoring of this level is necessary to prevent complications and to predict the development and phase of microvascular complications associated with diabetes. It has been proven that hyperglycemia contributes to the development and progression of micro-, macroangiopathy, polyneuropathy with damage, dysfunction and subsequent failure of vital organs and systems of the body. Damage to the vessels of the retina with rapid loss of vision, glomerulosclerosis with the development of chronic renal failure, diabetic foot, arterial hypertension, acute / chronic ischemic heart disease, brain and their consequences (myocardial infarction, heart failure, stroke, cerebral cortex atrophy with a decrease in cognitive functions and the development of dementia) - this is not a complete list of diseases and syndromes associated with diabetes mellitus.

Thus, the following target groups can be identified for which monitoring of glycated hemoglobin levels is necessary:

- healthy people for the purpose of preventive control;

- children with diabetes;

- patients with diabetes with abnormal renal glucose threshold;

- patients with type I diabetes mellitus, insulin-dependent;

- pregnant women with type II diabetes;

- other people when changing their diet or other habits.

Therefore, the level of glycated hemoglobin is an important and significant indicator of human health and requires continuous monitoring.

[0014] Invasive and non-invasive methods are used to determine the level of glycated hemoglobin. Invasive methods are laboratory methods, which include low-pressure ion exchange chromatography, capillary electrophoresis, immunoturbidimetric method, high-performance liquid chromatography, and colorimetric method. No special preparation is required to test blood for glycated hemoglobin. The results of the study are not affected by food intake, smoking, alcohol consumption, intensity of exercise, or the composition of foods consumed the day before. Blood for glycated hemoglobin can be donated in the procedure room of a clinic, or at a biological material collection point in laboratories. Biomaterial samples are taken from a peripheral vein or finger and delivered to the laboratory.

[0015] The colorimetric method is the most common, fairly inexpensive and accurate. Its essence lies in the fact that the stable form of HbA1c contains 1-deoxy-1-fructose. When phosphoric acid is added to the blood sample, it is dehydrated to 5-hydroxymethyl-2-furaldehyde. This compound, interacting with 2-thiobarbituric acid, forms a substance with maximum absorption. Using a spectrophotometer at a wavelength of 443 nm, the optical density of the resulting solution is determined and calculations are made.

[0016] However, regarding direct blood testing, it should be noted that regular blood sampling may be painful and create psychological stress. In addition, since frequent blood sampling may create a potential risk of infection for a person, frequent blood sugar testing is not advisable. Laboratory methods for performing blood tests require qualified personnel and expensive laboratory equipment, and waste disposal is necessary. Moreover, laboratory methods are not suitable for continuous monitoring of glycated hemoglobin levels, and patients cannot do them themselves.

[0017] It should also be noted that the HbA1c content is 4-6% of the total amount of hemoglobin in the blood, which means that a very high sensitivity of the device used for the HbA1c measurement method is required. In addition, in optical measurement methods in the visible and near infrared range, it is necessary to take into account that the absorption/reflectance spectra for oxygenated, glycated and total hemoglobin overlap each other, which means that they are difficult to separate, which will be discussed below for the specified range of the spectrum.

[0018] Currently, high-speed, ultra-compact portable analyzers for studying glycated hemoglobin have begun to gain popularity, but due to the high cost, this method has not yet found sufficiently wide application.

[0019] To date, a number of devices have been developed for non-invasive measurement of glycated hemoglobin, reducing discomfort for patients. However, they are not always applicable for continuous monitoring. Moreover, there are no wearable devices on the market, such as smart watches or fitness bracelets, designed to determine glycated hemoglobin non-invasively, which make it easier for the user to monitor themselves.

[0020] Patent application US 20160061810 Al (Samsung Electronics Co.) discloses a non-invasive device and method for diagnosing glycated hemoglobin. The device includes a first light measuring device configured to emit a first light, which is visible light or is in the near IR region of the spectrum, onto an object and detect first information about the first light reflected from the object, a second light measuring device configured to emit a second light, which is laser radiation, onto the object and detect second information about the second light reflected from the object, a data extracting device configured to extract first data related to hemoglobin from the first information and second data related to glucose from the second information, and a data processor configured to determine information related to glycated hemoglobin from the first and second data. The photodetector of the first light-measuring device is an absorption spectrophotometer, and the photodetector of the second light-measuring device is a Raman scattering spectrophotometer.

[0021] Patent US 9885698 B2 (Omni Medsci) discloses devices and methods using near infrared lasers for non-invasive monitoring of glucose, ketones, HbA1c and other blood components. A wearable device for use with a smartphone or tablet PC, which is auxiliary, includes a measuring device having a light source with a plurality of light-emitting diodes, intended for measuring physiological parameters and configured to create an optical beam with wavelengths including a wavelength of near IR light from 700 to 2500 nm. The measuring device includes lenses configured to deliver the optical beam to a sample of skin or biological tissue, which reflects the optical beam to a receiver located at a first distance from one of the light-emitting diodes and at a different distance from the other of the light-emitting diodes, and is also configured to generate an output signal representing a non-invasive measurement on blood contained in the sample. The wearable device is designed with the possibility of communicating with a smartphone or tablet PC, which receives, processes, stores and displays the output signal, wherein the processed output signal is designed with the possibility of transmission via a wireless communication line.

[0022] Patent US 11363973 B2 (Leman Micro Devices) discloses a personal handheld monitor (PHM) for non-invasive blood analysis, namely for measuring the concentration of an analyte in the blood. The PHM comprises a signal receiving device for receiving signals that can be used to extract the results of measuring a parameter related to the health of the user, wherein the signal receiving device comprises a blood photosensor having one or more photoemitters for transmitting light to a part of the user's body, one or more photodetectors for detecting light transmitted through the part of the body or scattered by it, and two or more optical cells, at least one of which contains the analyte to be detected or which simulates the absorption spectrum of the analyte to be detected, through which light that has been or will be transmitted through or reflected by the part of the body passes before it reaches said photodetector or each of the photodetectors, wherein the PHM processor is designed to process the signals.

[0023] The disadvantages of the presented solutions are the lack of separation of the spectra of oxygenated, glycated hemoglobin and other blood components, the lack of correction of the measurement of glycated hemoglobin due to the influence of other blood components.

[0024] International publication WO 2023038254 A1 (Korea I.T.S. Co.) discloses a method and device for non-invasively assessing glycated hemoglobin or glucose in blood using machine learning. For machine learning, features of biosignals obtained by a device with a PPG sensor and other measuring sensors, as well as user data, are used. For example, biosignals may include body temperature, pulse rate, blood pressure, respiratory rate, blood sugar level, brain waves, electrocardiogram and oxygen saturation. The feature extraction tool can extract various features by applying frequency analysis, statistical analysis, waveform analysis, etc. based on the obtained PPG signals.

[0025] The disadvantage of this solution is the lack of consideration of the influence of other blood components on the level of glycated hemoglobin, and, in addition, the said publication does not disclose the possibility of using a wearable device for non-invasive assessment of glycated hemoglobin.

[0026] Thus, the main problem of this technical field is the lack of a method and a convenient-to-use wearable device for determining and continuously monitoring such a blood parameter as glycated hemoglobin, taking into account the influence of other blood components. At the same time, the advantages of measuring glycated hemoglobin using a wearable device are instantaneous or continuous measurement, ease of use, and the possibility of long-term monitoring.

[0027] Accordingly, there is a need for a wearable device and a method for non-invasively determining the level of glycated hemoglobin, providing an accurate measurement and suitable for non-professional use. In other words, an easy-to-use wearable device is needed for non-invasive, unobtrusive, continuous monitoring of the level of glycated hemoglobin. However, the principles of the present disclosure can be used not only in a wearable device. Possible products that use the method of the invention are wearable electronic devices, such as smart watches or smart bracelets, stationary diagnostic devices, home appliances and gadgets for personal medical monitoring.

The essence of the invention

[0028] In view of the above technical problems, the present invention will be described hereinafter with reference to the description and the accompanying drawings by way of example and not limitation.

[0029] This summary of the invention precedes the detailed description of specific exemplary embodiments in order to provide a general idea of aspects of the claimed invention that will be further explained below, and is in no way intended to define or limit the scope of the present invention.

[0030] The objective of the present invention is to create a wearable device with the function of determining the user's blood parameters, in particular, the level of glycated hemoglobin, based on data collected from the sensors of the wearable device, taking into account the influence of other blood components. In particular, the present invention mainly uses PPG sensors, and other auxiliary sensors can also be used, for example, a gyroscopic sensor, an accelerometer, which help to recognize periods of events occurring in the user's daily activities (life), for example, periods of motor activity, rest or sleep. The principles of measuring PPG signals in the time and frequency domains are described in detail in patent RU 2805810 C1 dated 24.10.2023 issued in the name of the same applicant, which discloses a wearable device and a method for determining the hemoglobin level.

[0031] Within the framework of the present invention, an extensive study of the correlation between the features of PPG signals and blood parameters has been conducted, in particular, the possibility of accurately assessing the level of glycated hemoglobin using a wearable device has been confirmed. The main inventive concept is to separate the features (characteristics) of PPG signals into rapidly and slowly changing features and to compare them accordingly with rapidly changing parameters of blood components, such as, for example, total hemoglobin, oxygenated hemoglobin, blood glucose, and with slowly changing parameters of blood components, which include, for example, glycated hemoglobin. It is known that, for example, the level of oxygenated hemoglobin and glucose in the blood changes rapidly during physical activity, moving from a room to fresh air, holding your breath or other types of daily activities, such as sleeping, eating, etc., while the level of glycated hemoglobin remains virtually unchanged. Therefore, the algorithm for estimating the glycated hemoglobin level using the wearable device according to the present invention can be described as follows:

- continuous irradiation of the biological tissues of the user with radiation of at least two different wavelengths using PPG sensors of the wearable device during the user's daily activities;

- detecting backscattered PPG signals at said at least two different wavelengths;

- extraction of features (components) of detected PPG signals in time and frequency domains;

- determination of the moment of occurrence of a trigger event, at which a change in blood components occurs;

- identification of rapidly and slowly changing features characterizing rapidly and slowly changing parameters of blood components according to detected PPG signals,

- extraction from PPG signals detected during a trigger event of rapidly and slowly changing features characterizing, respectively, rapidly and slowly changing parameters of blood components;

- formation of a set of rapidly and slowly changing features characterizing, respectively, rapidly and slowly changing parameters of blood components, and

- determination of the level of glycated hemoglobin in the user's blood based on a comparison of the obtained set of features and information from a pre-compiled database using a predictive model.

Signal processing, including conversion, is performed using a microprocessor built into the wearable device, which is also used at all other stages of PPG signal processing. In addition, signal processing can also be performed on a remote server or device that communicates with the wearable device and can exchange data with it.

In the context of the present description, a rapidly changing component parameter is a component parameter whose value changes by more than 5% within 30-250 seconds after the trigger event (e.g., within 30 seconds during a breath hold and 250 seconds during an occlusion test), and a slowly changing component parameter is a component parameter whose value does not change by more than 3% within 60-120 days. As for the signal features, the calculation of the features is essentially the same for rapidly and slowly changing signal characteristics (time and frequency characteristics, shown, for example, in Table 1 below). The difference is that for the calculation of rapidly changing characteristics, signal sections during and after the trigger event are taken, with a duration of no more than 250 seconds, and for slowly changing characteristics, any signal sections within 60-120 days. Rapidly changing signal characteristics are characteristics that change within no more than 250 s during and after the trigger event, while slowly changing signal characteristics do not change during this time period.

[0032] Thus, one of the advantageous effects of the present invention is to take into account the influence of rapidly changing parameters of blood components on the assessment of the level of glycated hemoglobin.

[0033] It should be noted that the detection (registration) of signals from the PPG sensors of the wearable device can occur in a continuous mode during the user's daily activities and after the occurrence of a trigger event or during specified periods of time according to the user's wishes. Continuous registration is necessary to ensure the possibility of detecting the moment of occurrence of the trigger event. Registration of signals after the occurrence of the trigger event is necessary to ensure the possibility of identifying rapidly and slowly changing features of the PPG signals. For example, after the trigger event occurs, the wearable device continues to continuously detect PPG signals, however, to extract the signal features that are then identified as fast and slowly changing features, signals are used that are detected starting, for example, from 5 seconds, 10 seconds, 20 seconds, 30 seconds, 40 seconds, 50 seconds, 1 minute and more after the trigger event occurs, but are not limited to this, during several pulse wave cycles, for example, for 5 seconds, 10 seconds, 20 seconds, 30 seconds, 40 seconds, 50 seconds, 1 minute and more, but are not limited to this. In this case, determining the time of the trigger event occurrence can be carried out in the simplest case by changing the user's pulse rate, since the PPG signal is a pulse wave and an increase in pulse rate, for example, during physical activity, during sleep, will directly affect the repetition rate of the pulse signal and can be easily determined. In addition, as an example, determining the time of occurrence of a trigger event can be done by monitoring the user's breathing, which, like the pulse, changes with different types of activity, as disclosed in more detail in the patent application of the same applicant RU 2023125468 A1 dated 04.10.2023.

[0034] The solution to the problem of assessing the level of glycated hemoglobin taking into account the influence of other blood components is based on identifying empirical patterns in training data placed in a common database using a machine method. The training data includes sets of features measured using the disclosed wearable device and compared with the parameters of blood components measured using laboratory methods. The database can be constantly updated, and, in addition, the predictive model (prediction algorithm) is refined using machine learning. In the present invention, the predictive model is created for a wrist-worn device, since the training of the predictive model was based on data recorded by the sensors of the wrist-worn device.

[0035] PPG signals obtained by the wearable device characterize the biological tissue of the user through which the radiation passes. Optical characteristics of the biological tissue are obtained using PPG signals at least at two wavelengths in the radiation range from visible light to the near IR region of the spectrum, i.e. in the range of 400-1100 nm. The selection of specific radiation wavelengths is not particularly limited and can be carried out depending on various factors, such as, for example, the user's skin color and pigmentation. The PPG sensor detects light at the at least two wavelengths to obtain a set of detected PPG signals in the time domain. Then, the detected PPG signals are converted from the time domain to the frequency domain to obtain a set of detected PPG signals in the frequency domain.

[0036] More specifically, the measurement technique used in the present disclosure is based on the use of photoplethysmography (PPG) principles. When the biological tissue of the wearable device user is irradiated, absorption, reflection and scattering of radiation occurs in biological tissues (epithelial, connective, nervous, muscular), bones, venous blood and arterial blood, which, when studying blood flow, is determined by the size of the vessels or the volume of blood passing through the examined tissue area. The narrowing and expansion of the vessel under the action of arterial blood flow pulsation, which is caused mainly by a change in blood flow in the arteries and arterioles, causes a corresponding change in the amplitude of the signal obtained at the output of the photodetector. Hemoglobin serves as a kind of filter for the light flux, and its "color" and "thickness" can change. The "color" of the filter depends mainly on the percentage of oxyhemoglobin. As an example, we note that the ability of, for example, a pulse oximeter to determine the degree of blood oxygenation is based on this. The change in the "thickness" of the filter is affected by the growth of the blood volume in the arteries and arterioles with each pulse wave. After measuring the pulse rate and the amplitude of the pulse wave, the ratio of the absorption of radiation of different wavelengths, for example, the PC and red regions of the spectrum, is analyzed using a microprocessor, then the saturation of the pulsating flow of arterial blood with oxygen is calculated. In this case, the simultaneous use of the PC and red regions of the spectrum is associated with the difference in the absorption of radiation by oxyhemoglobin, which absorbs less radiation with a wavelength of 660 nm, and deoxyhemoglobin, which absorbs less radiation with a wavelength of 940 nm.

[0037] Traditionally, devices with PPG sensors operate in the mode of measuring a biological tissue sample in transmission, when the device (or part of it) is placed on a finger, earlobe, i.e. where radiation can be transmitted through the tissue sample. However, in recent years, devices have been developed that operate in the reflection mode, when radiation backscattered by biological tissue is measured. The reflection mode is also used in the wearable device according to the present disclosure.

[0038] The processes of blood supply to biological tissue are associated with a pulse wave, i.e. a change in the amount of blood in the measured volume of tissue. Accordingly, the signal from the sensor output, proportional to the absorption of light passing through the tissue, includes two components: a weak variable component (AC), caused by a change in the volume of arterial blood with each heartbeat, against the background of a strong constant component (DC) of the signal, determined by the optical properties of the skin, bones, venous blood and other tissues of the area under study, the total absorption of which does not change during the propagation of the pulse wave.

[0039] The traditional approach to PPG signal analysis is based on separating the constant and variable components of the time series (DC and AC components in the time domain or DC and AC components) of the PPG signal. In this case, in the case of operation in the transmission mode, the integral area of the signal curve is traditionally measured, and in the case of operation in the reflection mode, the signal shape is analyzed. At the same time, the DC component is used to normalize the signal, since the transmission of the signal depends not only on the absorption/reflection/scattering coefficients of the skin, bones, blood and other tissues that make up the "filter" for the signal, but also on the thickness of the "filter" itself.

[0040] Extraction of PPG signal features in the time domain consists of analyzing the shape of the signal curve, i.e. their AC components. For correct measurement of the AC component of the PPG signal, accumulation of several pulse wave cycles is required, and the measurement process or data collection usually takes several seconds or tens of seconds. At the same time, analysis of the AC component of the PPG signal can be difficult, since the AC component often has a complex shape, which depends on the wavelength and carries information about the spectral properties of the tissue.

[0041] The set of features of PPG signals in the time domain is a set of spectral-temporal features (v _λ1 ^tl , v _λ2 ^tl , …, v _λn ^tl , … v _λ1 ^tk , … v _λn ^tk ) of signals characterizing the molecular structure of blood, i.e. the spectral and dynamic properties of blood components, where

λ ₁ , λ ₂ , …, λ _n - index indicating the wavelength,

t _l ,…t _k - index indicating the temporal feature, where the spectral-temporal feature v ^tk _λn corresponds to the temporal feature t _k , measured at the wavelength λ _n .

[0042] Additionally, the PPG signals are transformed from the time domain to the frequency domain, for example, by means of the Fourier or Hilbert-Huang transform, and the features of the PPG signals in the frequency domain are extracted. First, in the time domain, among all available signals at different wavelengths, a reference PPG signal is selected at a certain wavelength based on the highest signal-to-noise ratio or the highest amplitude (usually this is a signal at the wavelength of green light - 495-570 nm).

[0043] In the frequency domain, the position of the first harmonic (fundamental component) f ₀ is determined based on the reference PPG signal and is used to analyze signals at other wavelengths. Based on the position of the first harmonic f ₀ , the amplitude-phase characteristics of the harmonics of all the obtained optical signals are determined. Thus, for all combinations of the wavelengths of the PPG signal and harmonics, a set of spectral-frequency features is obtained (ƒ ₀ , a _λ1 ^ƒ0 , a _λ2 ^ƒ0 , a _λn ^ƒ0 , … a _λn ^ƒk ), which uniquely describes the shape of the pulse wave at each of the used wavelengths λ. It should be noted here that when converting a PPG signal from the time to the frequency domain, the periodicity of the pulse signal is transferred to the position of the first harmonic f ₀ , and its shape, i.e. peaks, troughs, notches, are transferred to harmonics of higher orders.

[0044] The PPG signals are recorded continuously during the user's daily activities and during the trigger event. After recording the PPG signals, the signal features are extracted in the time and frequency domains. In this case, determining the moment of occurrence of the trigger event is necessary, according to the present disclosure, for identifying the fast and slowly changing features of the PPG signals by comparing the features of the signals during the user's daily activities and the trigger event. Accordingly, a set of spectral-wave features and a set of spectral-frequency features obtained during the trigger event are used to form a set of fast and slowly changing features characterizing the fast and slowly changing parameters of the blood components.

[0045] Thus, as a result of processing the signals recorded by the PPG sensors, a set of rapidly and slowly changing features is formed, characterizing the molecular structure of the blood, which is used to predict the level of glycated hemoglobin, taking into account the influence of other blood components. In this case, since the level of glycated hemoglobin is characterized only by slowly changing features, and other blood components are characterized by rapidly changing features, it is possible to take into account the influence of other blood components on the accuracy of measuring the level of glycated hemoglobin.

[0046] As discussed above, said set of fast and slow changing features consists of a set of spectral-temporal features (v _λ1 ^tl , v _λ2 ^tl , …, v _λn ^tl , … v _λ1 ^tk , … v _λn ^tk ) and a set of spectral-frequency features (ƒ ₀ , a _λ1 ^ƒ0 , a _λ2 ^ƒ0 , a _λn ^ƒ0 , … a _λn ^ƒk ), which include, for example, at least the amplitudes of the diastolic and systolic peaks, the magnitude of the dicrotic notch (incisure) of the pulse wave recorded from the user, the magnitude of the first harmonic (ƒ ₀ ), the magnitudes of the amplitude-phase characteristics of the harmonics (at least the amplitude and the total power (area under the curve) of at least two harmonics within the spectral a band specified relative to a central frequency, the position of which is determined by a reference frequency corresponding to the heart rate (HR), of at least two PPG signals with different wavelengths.

[0047] Molecular structure refers to the presence and concentration of dye molecules (hemoglobin or other component) in the blood. The optical signal of the PPG sensor is predominantly sensitive to the molecular structure of the blood, i.e. the concentration of the dye. At the same time, the formation of a set of rapidly and slowly changing features of the PPG signals and their correlation with rapidly and slowly changing parameters of the blood components, respectively, makes it possible to determine the level of a slowly changing parameter of a blood component, such as, for example, glycated hemoglobin, taking into account other blood components.

[0048] The proposed wearable device, method and system provide the possibility of personal, non-professional use and the possibility of instantaneous or continuous determination of blood parameters with high accuracy, for example, the level of glycated hemoglobin (HbA1c).

[0049] At the same time, the declared device:

- does not require the use of additional separate measuring instruments, can be used in the form of existing wearable devices (smart watches, fitness bracelets, etc.),

- has high accuracy in measuring the level of glycated hemoglobin,

- has low energy consumption,

- capable of continuous monitoring without requiring user intervention.

Brief description of the drawings

[0050] The above and other features and advantages of the present disclosure will become clearer in the following description with reference to the accompanying drawings. The specific exemplary embodiments of the claimed invention, when considered in conjunction with the drawings, are not intended to limit the scope of the invention. Additional embodiments, modifications, or equivalents of the present disclosure will become apparent to those skilled in the art from a study of the present description, and all such embodiments, modifications, and equivalents are intended to be included within the present disclosure.

[0051] The drawings are provided solely to aid in understanding the description and should not be construed as limiting the scope of the invention in any way. The drawings illustrate the following:

[0052] Fig. 1A illustrates a block diagram of a wearable device with a glycated hemoglobin determination function.

Fig. 1B illustrates a schematic view of the wearable device.

Fig. 2A illustrates the absorption spectrum of oxygenated hemoglobin and glycated hemoglobin in the visible-near infrared (VIS-NIR) range.

Fig. 2B illustrates theoretically calculated absorption spectra of glycated hemoglobin in the VIS-NIR range at different HbA1c levels in the solvent.

Fig. 3A schematically illustrates the composition of blood and the main components of blood. Fig. 3B schematically illustrates graphs of changes in the concentration of rapidly changing parameters of blood components and slowly changing parameters of blood components depending on time.

Fig. 4 schematically illustrates the setup of a breath-holding experiment in which changes in various physiological signals were investigated as a function of time.

Fig. 5 illustrates filtered high-quality physiological signals obtained during breath holding and normal breathing.

Fig. 6A, 6B illustrate graphs confirming the feasibility of practical implementation of measuring separate levels of HbO ₂ and HbA1c using the wearable device of the present invention.

Fig. 7 illustrates the first key feature (KO1) of the present invention.

Fig. 8 illustrates the second and third key features (KO2 and KO3) of the present invention.

Fig. 9, 10A, 10B, 10C illustrate in more detail the KO1 of the present invention, which consists in detecting the parameters of PPG signals.

Fig. 11 illustrates a graph schematically showing the dependence of the amplitude on time of the levels of slowly changing features (characteristics) (f ^{^} (t)) of the PPG signal and rapidly changing features (f*(t)) of the PPG signal during continuous monitoring of the PPG signals.

Fig. 12 illustrates in more detail the KO2 of the present invention, which consists in compiling a reference database.

Fig. 13 illustrates in more detail the KO3 of the present invention, which characterizes the final stage of assessing the level of glycated hemoglobin, taking into account the influence of other blood components.

Fig. 14 illustrates the features of determining events occurring in the daily activities of the user of the wearable device according to the present invention.

Fig. 15A schematically illustrates variants of an exemplary display of the results of measuring the HbA1c level during the day and the deviation of the results of measuring the HbA1c over several months on the screen of the wearable device according to the invention in the S-Health application.

Fig. 15B schematically illustrates a graph displayed on the screen of a wearable device in the S-Health application, which allows monitoring the deviation of the glycated hemoglobin level from the target value.

Fig. 16A illustrates a schematic diagram of a breath-holding experiment in which changes in PPG signals were monitored for different users in different modes and the level of glycated hemoglobin was determined using an experimental sample of the claimed wearable device.

Fig. 16B illustrates the results of determining the level of glycated hemoglobin in the blood of a wearable device user using a reference database.

Detailed description of embodiments of the invention

[0053] The proposed invention is a wearable device with a function of determining the level of glycated hemoglobin in the blood of a user, comprising at least one photoplethysmographic (PPG) sensor configured to irradiate the biological tissue of the user with radiation of at least two different wavelengths and to detect backscattered PPG signals at said at least two different wavelengths, wherein the wearable device is configured to, during the user's daily activities, detect PPG signals in reflection mode, extract features of the detected PPG signals, determine an event occurring in the user's daily activities during which a change in the values of the parameters of blood components occurs, identify, from the detected PPG signals, rapidly and slowly changing features characterizing the rapidly and slowly changing parameters of the blood components, respectively, extract from the PPG signals detected during the occurring event rapidly and slowly changing features characterizing the rapidly and slowly changing parameters of the blood components, respectively, and form from them a set of rapidly and slowly changing features of the PPG signals, and determining the level of glycated hemoglobin in the user's blood based on a comparison of a generated set of features and information from a pre-compiled database using a predictive model.

[0054] According to an embodiment, the wearable device further comprises a housing, a processor, a battery and a memory device located in the housing, wherein the memory device is configured to store a database.

[0055] According to an embodiment, the wearable device further comprises an input and output device configured to input user profile data and display the determined glycated hemoglobin level, and a communication module configured to exchange information with a remote server and/or cloud storage.

[0056] According to an embodiment, said at least one PPG sensor of the wearable device provides radiation of different wavelengths in the visible (VIS) - near infrared (NIR) (approximately 400-1100 nm) spectral region. In an embodiment, one of said different wavelengths is preferably in the range of 490-570 nm, and more preferably is 530 nm.

[0057] According to an embodiment, said at least one PPG sensor of the wearable device comprises at least one radiation source comprising two or more light-emitting diodes.

[0058] According to an embodiment, said at least one PPG sensor of the wearable device comprises at least one radiation receiver, preferably a broadband photodetector.

[0059] According to an embodiment, the database information comprises a feature matrix including sets of features characterizing fast and slowly changing features of PPG signals, compared with reference values of fast and slowly changing parameters of blood components, determined by laboratory methods for a plurality of users, and profile data of a plurality of users, wherein said sets of features are generated by the wearable device during events occurring in the daily activities of the users and trigger events associated with a change in the values of the fast changing parameters of blood components.

[0060] According to an embodiment, the wearable device is configured to identify fast and slowly changing features of PPG signals by comparing the features of the signals during the user's daily activities and an ongoing event in the daily activities.

[0061] According to an embodiment, the blood components with rapidly changing parameters include total hemoglobin, oxygenated hemoglobin, glucose level, and the blood components with slowly changing parameters include glycated hemoglobin.

[0062] According to an embodiment, the wearable device is further configured to convert said at least two PPG signals from the time domain to the frequency domain and describe the converted PPG signals with respect to harmonics, wherein the set of features characterizing rapidly and slowly changing parameters of blood components includes features of PPG signals determined in the time and frequency domains.

[0063] According to an embodiment, the transformation of PPG signals from the time domain to the frequency domain comprises a Fourier transform.

[0064] According to an embodiment, the wearable device is configured to, when describing PPG signals in the frequency domain, select from all PPG signals the most significant signal characterized by the highest signal-to-noise ratio or the highest amplitude and use it as a reference signal, extract harmonics from the reference signal, determine the coordinate of the fundamental harmonic (f ₀ ) and calculate the coordinates of other harmonics (f _l ... f _n ), decompose all PPG signals by said coordinates, measure for each harmonic the amplitude, the area of the curve, the width at half the peak height and/or the signal-to-noise ratio to obtain a set of amplitude-phase characteristics (t _λ ⁰ ) of the harmonics, which are features determined in the frequency domain.

[0065] According to an embodiment, the features determined in the time domain include critical points of the PPG signal waveform, such as the amplitude of the systolic peak, the amplitude of the diastolic peak, the magnitude of the dicrotic notch, and the pulse rate.

[0066] According to an embodiment, the features determined in the time domain additionally include starting points of the sliding (adaptive) time window of the PPG signals, such as the window step length, the window length, the segment length.

[0067] According to an embodiment, the comparison in determining the level of glycated hemoglobin is based on the use of a numerical method or a prediction algorithm, and the prediction model is pre-trained on a compiled database using a machine learning algorithm or a neural network.

[0068] According to an embodiment, the wearable device is configured to be placed on the wrist.

[0069] According to an embodiment, the wearable device is a smart device, preferably a smartwatch or a fitness bracelet.

[0070] The proposed invention also relates to a method for determining the level of glycated hemoglobin in the blood of a user, comprising the steps of irradiating the biological tissue of the user with radiation of at least two different wavelengths during the daily activities of the user of the wearable device, detecting backscattered PPG signals at said at least two different wavelengths in reflection mode, extracting features of the detected PPG signals, determining an event occurring in the daily activities of the user during which a change in the values of the parameters of the blood components occurs, identifying rapidly and slowly changing features characterizing the rapidly and slowly changing parameters of the blood components, respectively, from the detected PPG signals, extracting rapidly and slowly changing features characterizing the rapidly and slowly changing parameters of the blood components, respectively, from the PPG signals detected during the occurring event, and forming a set of rapidly and slowly changing features of the PPG signals from them, and determining the level of glycated hemoglobin in the blood of the user based on a comparison of the formed set of features and information from a pre-compiled database using predictive model.

[0071] According to an embodiment of the method, the biological tissue of the user is irradiated with radiation of various wavelengths in the visible (VIS) - near infrared (NIR) range of the spectrum.

[0072] According to an embodiment of the method, the biological tissue of the user is irradiated using at least one radiation source containing two or more light-emitting diodes.

[0073] According to an embodiment of the method, the backscattered PPG signals are detected using at least one radiation detector, preferably a broadband photodetector.

[0074] According to an embodiment of the method, the database information contains a matrix of features including sets of features characterizing rapidly and slowly changing features of PPG signals, compared with reference values of rapidly and slowly changing parameters of blood components, determined by laboratory methods for a plurality of users, and profile data of a plurality of users, wherein the sets of features are generated by the wearable device during events occurring in the daily activities of users and trigger events associated with a change in the values of rapidly changing parameters of blood components.

[0075] According to an embodiment of the method, fast and slow changing features of PPG signals are identified by comparing the features of the signals during the user's daily activities and an ongoing event in the daily activities.

[0076] According to an embodiment of the method, the blood components with rapidly changing parameters include total hemoglobin, oxygenated hemoglobin, and glucose level, and the blood components with slowly changing parameters include glycated hemoglobin.

[0077] According to an embodiment, the method further includes the steps of converting said at least two PPG signals from the time domain to the frequency domain and describing the converted PPG signals with respect to harmonics in the frequency domain, wherein the set of features characterizing rapidly and slowly changing parameters of blood components includes features determined in the time domain and in the frequency domain.

[0078] According to an embodiment of the method, the step of converting PPG signals from the time domain to the frequency domain is carried out by means of a Fourier transform.

[0079] According to an embodiment of the method, the step of describing the PPG signals in the frequency domain is carried out by selecting from all the PPG signals the most significant signal, characterized by the highest signal-to-noise ratio or the highest amplitude, and using it as a reference signal, extracting harmonics from the reference signal, determining the coordinate of the fundamental harmonic (f ₀ ) and calculating the coordinates of the other harmonics (f _l ... f _n ), decomposing all the PPG signals according to the said coordinates, measuring for each harmonic the amplitude, the area of the curve, the width at half the peak height and/or the signal-to-noise ratio with the receipt of a set of amplitude-phase characteristics (t _λ ⁰ ) of the harmonics, which represent features determined in the frequency domain.

[0080] According to an embodiment of the method, the features determined in the time domain include critical points of the PPG signals, such as the amplitude of the systolic peak, the amplitude of the diastolic peak, the size of the dicrotic notch, the pulse rate, and additionally include the starting points of the sliding time window of the PPG signals, such as the step length of the window, the length of the window, the length of the segment.

[0081] According to an embodiment of the method, the comparison step is carried out using a numerical method or a prediction algorithm, and the prediction model is pre-trained using a compiled database using a machine learning algorithm or a neural network.

[0082] According to an embodiment of the method, the database is configured to be stored on a storage device, a remote server and/or in cloud storage.

[0083] The proposed invention further relates to a system for determining the level of glycated hemoglobin in the blood of a user, comprising a wearable device according to any of the above-mentioned variants and a remote server and/or cloud storage, wherein the system comprises means for exchanging information between the wearable device and the remote server and/or cloud storage.

[0084] According to an embodiment of the system, the database is intended for storage on a storage device, a remote server and/or in cloud storage and includes user profile data and results of previous determinations of the glycated hemoglobin level, wherein the system further comprises means for accessing the user profile data and results of previous determinations of the glycated hemoglobin level from external devices via the remote server and/or cloud storage.

[0085] The proposed wearable device, method and system provide the ability to non-invasively, accurately continuously monitor the level of glycated hemoglobin (HbA1c) and are suitable for non-professional use.

[0086] Features of the present disclosure and exemplary embodiments of the invention will now be described with reference to the drawings. It will be understood by those skilled in the art that the various exemplary embodiments should not be construed as limiting the scope of the claimed invention, and that other material and technical means equivalent or similar to those listed below may be applied by those skilled in the art to perform the various operations, functions, method steps, etc. described below. This detailed description is not intended to limit the scope of the claimed invention, which is defined only by the appended claims.

[0087] It should be noted that the principles of the present invention are given with respect to measuring the level of glycated hemoglobin. However, the present invention is not limited to this and can be used to determine other parameters of the user's blood, such as, for example, total hemoglobin, oxygenated hemoglobin, by training and using an appropriate predictive model. The input data for the predictive model are the characteristics of the molecular structure of the blood (see Table 1), which include at least a set of spectral-temporal features (v _λ1 ^tl , v _λ2 ^tl , ..., v _λn ^tl , ... v _λ1 ^tk , ... v _λn ^tk ) signals and a set of spectral-frequency features ( a _λ1 ^ƒ0 , a _λ2 ^ƒ0 , a _λn ^ƒ0 , ... a _λn ^ƒk ) signals.

[0088] The set of spectral-temporal features for each of at least two wavelengths represents at least one of the intensity of scattered radiation, the total signal power, the power in a given frequency band of the signal, critical points of the signal waveform characterizing the dynamics of blood components, and combinations of said spectral-temporal features, absolute values of said spectral-temporal features, relative values of said spectral-temporal features, linear combinations of said spectral-temporal features at different wavelengths, nonlinear combinations of said spectral-temporal features at different wavelengths.

[0089] The set of spectral-frequency features includes, for example, amplitude-phase characteristics of harmonics (amplitude and total power of at least two harmonics, including the magnitude of the fundamental harmonic (ƒ ₀ ), within a spectral band specified relative to the central frequency, the position of which is determined by the reference frequency corresponding to the heart rate, of at least two PPG signals with different wavelengths (λ). Additional input data may be user profile data, at least one parameter of the user's body impedance, such as the magnitude (Z) of the user's body impedance modulus, the phase angle (ϕ) of the user's body impedance at certain frequencies (at least two frequencies, such as 20-40 kHz and 200-300 kHz), the total amount of fluid in the body (TBW), the ratio of the amount of intracellular fluid (ICW) to the amount of extracellular fluid (ECW), obtained using a bioimpedance (BIA) sensor, etc.

[0090] The amplitude-phase characteristics of the harmonics are the original characteristics extracted from the PPG signal at each wavelength used for measurement, namely:

a ₀ , a ₁ , a ₂ , ... a _n - the total power (e.g., the area under the curve) of harmonics and/or the amplitude of harmonics at the corresponding frequencies ƒ ₀ , ƒ ₁ =2ƒ ₀ , ƒ ₂ =3ƒ ₀ , …, where ƒ ₀ is the reference frequency (fundamental harmonic) corresponding to the heart rate.

In addition, various derivative characteristics of the above-mentioned features can be used to determine the level of glycated hemoglobin.

[0091] Knowledge of the above-mentioned sets of spectral-temporal and spectral-frequency features determined by PPG signals in the time and frequency domains, as well as the presence of a pre-compiled database in which the sets of features combined into matrices are compared with specific blood components (values of blood component parameters, in particular, the level of glycated hemoglobin), measured by laboratory methods, provides the ability to accurately determine the level of glycated hemoglobin taking into account the influence of other blood components. In this case, in the database, the comparison of specific values of blood component parameters with the matrix of features is performed in the form of functional dependencies and the values of the coefficients included in them are determined in the calibration process by at least one of the regression methods, including linear regression, logistic regression, successive approximation methods, differential methods, including gradient methods such as gradient descent, stochastic gradient descent, and modifications of the said methods, based on the values of the blood component parameters measured in advance for a plurality of subjects.

[0092] Table 1 below shows an example of a set of input data used to determine the level of glycated hemoglobin according to the present invention and the source of their determination. This set of input data is provided as an example and is not limited to this.

[0093] The specifics of determining glycated hemoglobin levels will be discussed below with reference to the drawings.

[0094] Referring to Fig. 1A and 1B, a wearable device 201 with a function for determining glycated hemoglobin is schematically illustrated.

[0095] Fig. 1A schematically illustrates a block diagram of a preferred embodiment of a wearable device 201 comprising a housing 202, a processor (e.g., a microprocessor) 203, a battery 204, at least one PPG sensor 208, an input module 209, an output module 210, a display module 211 including a display, a memory (SRAM) 212 including a volatile SRAM and a non-volatile SRAM, a communication module 213 including a wireless communication module 223 and a wired communication module 224, and a sensor module 218. The input module 209 and the output module 210 may be implemented as a single input and output module. The sensor module 218 of the wearable device 201 may comprise sensors such as a bioimpedance (BIA) sensor, inertial displacement sensors, gyroscopic sensors, accelerometers (acceleration sensors), an electrocardiogram (ECG) measurement sensor, an atmospheric pressure sensor, a humidity sensor, a temperature sensor, etc. In embodiments, the wearable device 201 may be a smart device, in particular, a smart watch or a fitness bracelet, a medical wearable device for monitoring health status with a function for determining glycated hemoglobin, etc. The wearable device is an electronic wearable device and may be configured to establish a wired or wireless communication channel with external devices, for example, devices 225, 226, such as smartphones, wearable fitness bracelets, voice assistants, smart TVs, smart watches, headphones, etc., or a server 227 and with the ability to transmit data via a network 228 or cloud storage 229. In some embodiments, at least one of the components may be excluded from the electronic device 201, or one or more other components may be added to the electronic device 201.

[0096] The processor 203 may execute software for controlling at least one other component (e.g., the sensor module 218) of the electronic device 201 and may perform various data processing or calculations. As at least part of the data processing or calculation, the processor 203 may load a command or data received from the other component (e.g., the sensor module 218 or the communication module 213) into a volatile memory, process the command or data stored in the volatile memory, and store the resulting data in a non-volatile memory. The processor 203 may include a main processor (e.g., a central processing unit (CPU), an application processor (AP)), an auxiliary processor, and the like, where the auxiliary processor may operate independently of the main processor or in conjunction with it. The auxiliary processor may be implemented separately from or as part of the main processor.

[0097] The processor 203 of the wearable device 201 is configured to transmit information about the values of the parameters of the components of the user's blood, determined based on the readings of the sensors 208, to the output device 210 to inform the user.

[0098] The processor 203 is configured to download data received via the communication module 213 to the memory device 212 and/or download data from the memory device 212 to the communication module 213 for transmitting them to an external device (for example, devices 225, 226, server 227 or cloud storage 229).

[0099] According to an embodiment, the main processor or auxiliary processor (e.g., a data processing device in a neural network) may include a hardware structure designed to process an artificial intelligence model. The artificial intelligence model may be created using machine learning. Such training may be performed, for example, in the electronic device 201 itself, on which the artificial intelligence model is executed, or may be performed through a separate server (e.g., server 227). Learning algorithms may include, for example, supervised learning, unsupervised learning, semi-supervised learning, or reinforcement learning, but are not limited to this. The artificial intelligence model may include multiple layers of an artificial neural network. The artificial neural network may be any of a deep neural network (DNN), a convolutional neural network (CNN), a recurrent neural network (RNN), a restricted Boltzmann machine (RBM), a deep belief network (DBN), a bidirectional recurrent DNN (BRDNN), a deep Q-network, or a combination of two or more of the above networks, but is not limited to the above examples. In addition to the hardware structure, the artificial intelligence model may additionally or alternatively include a software structure.

[0100] The memory device 212 may store various data used by at least one component (e.g., the processor 203 or the sensor module 218) of the electronic device 201. The various data may include, for example, software (e.g., a program) and input data or output data for an associated command. The program may be stored in the memory device 212 as software and may include, for example, an operating system (OS), middleware, and/or applications.

[0101] The wearable device 201 with the PPG sensor 208 is capable of detecting, in the reflection mode, radiation backscattered and/or backreflected from the biological tissues of the user, including skin, bones, blood, blood vessels. Detection and measurement of the backscattered and/or backreflected radiation is performed using at least one photodetector, which is at least one of a photodiode, a photoresistor, a phototransistor, a charge-coupled device (CCD), or a device based on a complementary metal-oxide-semiconductor (CMOS) structure, or another light-sensitive element, and a photocurrent-voltage converter.

[0102] In a preferred embodiment, the photodetector is a broadband photodetector configured to detect and amplify at least one response of backscattered and/or backreflected radiation from dynamic blood components and static tissue components, representing a signal from each of at least two radiation sources.

[0103] In wearable devices with a PPG sensor, a reflective mode is used to conveniently place the device, for example, on the user's arm. When worn on the arm or other part of the user's body, the PPG sensor is in contact with the skin. However, a PPG sensor that takes measurements in the reflective mode can be affected by motion artifacts and pressure fluctuations. Any movement, such as physical activity, can affect the accuracy of measurements due to motion artifacts that distort the PPG signal and limit the accuracy of measuring physiological parameters. Pressure fluctuations acting on the sensor, such as the contact force between the PPG sensor and the measurement site, that is, the force with which the device is pressed against the user's skin, can deform the geometry of the artery due to compression. Thus, in the reflective mode, the amplitude of the AC component of the PPG signal can be affected by the pressure exerted on the skin. More detailed information on such influences on PPG signal measurement is provided in the article by Toshiyo Tamura, Current Progress of Photoplethysmography and SpO ₂ for Health Monitoring, Biomedical Engineering Letters (2019) 9, pp. 21–36 (https://doi.org/10.1007/sl3534-019-00097-w).

[0104] The electrical circuit of the PPG sensor may include an amplifier, a high-pass filter (about 0.1 Hz) to cut off the constant component and obtain pulsating changes in the signal, a low-pass filter (about 30 Hz) to eliminate high-frequency noise, and a microprocessor. The frequency range used depends on the design of the circuit. The PPG sensor may have a wireless module for transmitting data to an external device.

[0105] A BIA sensor (bioimpedance measurement module) may include a biocontact circuit (a circuit for contact with the user's body) containing 4 electrodes, an impedance measurement module containing a current-voltage measurement module, a current-voltage circuit configuration module, an impedance calculation module, at least one leakage current compensation module, and, optionally, a microprocessor.

[0106] In a preferred embodiment of the present invention, illustrated in Fig. 1B, the wearable device 201 is made in the form of a smart watch, which is intended to be placed on the user's wrist. In other embodiments of the present invention, the wearable device can be made to be placed on other parts of the user's body, for example, on a finger.

[0107] Fig. 1B schematically illustrates a wearable device 201 in a top view and a bottom view. According to this embodiment, the body 202 of the wearable device 201 includes a first side 214 (or front surface), a second side (or back surface) 215 and a side surface 216 surrounding the space between the first side 214 and the second side 215, and fastening elements 217 intended for detachably fastening the wearable device 201 to the wrist, for example, using a watch strap. In addition, as an example, it is shown that the wearable device 201 can have four BIA sensor electrodes, two of which 251, 253 are integrated into the control buttons of the device, and the other two 252, 254 are located on the back side 215 of the wearable device.

[0108] According to an embodiment of the present invention, the processor 203 is configured to receive said at least two PPG signals from said at least one PPG sensor 208, store them in the memory device 212 and load them from the memory device 212 to perform processing operations.

[0109] The processor 203 is configured to filter said at least two PPG signals of different wavelengths from motion artifacts, separate said at least two PPG signals of different wavelengths into AC and DC components of time series and transform these time series (signal in the time domain) into the frequency domain with subsequent extraction of a set of features of the PPG signals in the time and frequency domains characterizing rapidly and slowly changing parameters of the blood components, wherein in the frequency domain, the amplitude-phase characteristics of the harmonics are distinguished as such features. In this case, the transformation of said time series into the frequency domain may comprise the Fourier transform or the Hilbert-Huang transform.

[0110] The battery 204 is configured to supply power to at least one component of the wearable device 201. The battery may be a primary cell or a secondary (rechargeable) cell and may be built into the wearable device or replaceable.

[0111] With reference to Fig. 1B, it is illustrated that said at least one PPG sensor 208 of the wearable device 201 is located on its rear surface. Thus, during operation, the radiation surface of the PPG sensor comes into contact with the user's wrist. Several PPG sensors may be provided in the wearable device 201.

[0112] The at least one PPG sensor 208 is configured to measure at least two PPG signals with different wavelengths. In the embodiment of the present invention shown in Fig. 1B, the at least one PPG sensor 208 is located on the side of the back surface 215 of the housing 202 (the surface facing the user's skin). In one embodiment of the present invention (not shown), the wearable device 201 may comprise two PPG sensors 208, each of which comprises one radiation source 236 and one radiation receiver 237, wherein the radiation sources 236 of these two PPG sensors emit light of different wavelengths. In this embodiment of the present invention, the two PPG sensors 208 are capable of providing registration of at least two PPG signals with different wavelengths. In the embodiment of the present invention shown in Fig. 1B, the wearable device 201 comprises one PPG sensor 208 comprising two radiation sources 236 and two radiation receivers 237, wherein these radiation sources 236 emit light of different wavelengths. In this embodiment of the present invention, one PPG sensor 208 is capable of providing detection of at least two PPG signals of different wavelengths. In the embodiment of the present invention, the radiation source 236 may comprise a veto-emitting diode (LED). Photo barriers may be provided between the radiation sources and/or receivers to reduce noise and eliminate crosstalk.

[0113] The input device 209 is configured to receive data that can be used by another component (for example, the processor 203) of the wearable device 201, entered from the outside (for example, by a user of the wearable device 201). In one embodiment of the present invention, the input device 209 can be configured to enter a user profile. The processor 203 of the wearable device 201 can receive the user profile from the input device 209, store the user profile in the memory device 212, read it from the memory device 212 and use it if necessary. Also, voice input and output of data can be provided in the wearable device 201.

[0114] The user profile data may include the user's gender, age, height, weight, and data on the values of blood component parameters for previous periods. The output device 210 may be configured to output information to the user on the first side 214 of the wearable device 201. The output device 210 may include a display 211 that may be configured to display the output information and is located on the first side 214 of the wearable device 201. Additionally, the display 211 may be configured to use an on-screen keyboard to enter data, such as, for example, a user profile.

[0115] The memory device 212 in a preferred embodiment may be configured to store:

databases with sets of PPG signal features characterizing rapidly and slowly changing parameters of blood components, compared with various blood parameters measured by laboratory methods;

user profile data;

sets of measurement data containing sets of features characterizing rapidly and slowly changing parameters of blood components, and including:

features of at least two PPG signals with different wavelengths in the time domain:

features of the mentioned at least two PPG signals in the frequency domain, including the magnitude of the first harmonic (f ₀ ) and the magnitudes of the amplitude-phase characteristics (t ^λ ₀ ) of other harmonics;

dates and times of collection of each set of measurement data;

glycated hemoglobin values determined for each set of measurement data, etc.

[0116] Also, the memory device 212 can store various additional data, for example, taken from additional sensors of the wearable device 201. In addition, the memory device 212 can store various instructions that, when executed on the processor 203, cause the processor 203 to control the components of the wearable device 201 associated with the processor 203 and to perform various data processing or calculations.

[0117] The above-mentioned data that can be stored by the memory device 212 can also be stored on a remote server and/or in a cloud storage. A communication module 213 can be used to transmit data from or to the wearable device. According to an embodiment, the communication module 213 can include a wireless communication module 223 (for example, a cellular communication module, a short-range wireless communication module, or a global navigation satellite system (GNSS) communication module) or a wired communication module 224 (for example, a local area network (LAN) communication module or a power line communication module). Thus, the communication module 213 can support establishing a wireless communication channel between the wearable device 201 and an external device (for example, device 225, device 226, server 227, or cloud storage 228) and performing communication via the established communication channel. The communication module 213 may include one or more communication processors that operate independently of the processor 203 and support direct (e.g., wired) communication or wireless communication. A corresponding one of these communication modules may communicate with an external electronic device, a remote server or cloud storage via a first network 228 (e.g., a short-range communication network such as Bluetooth™, wireless communication (Wi-Fi) or data transmission via the infrared data acquisition standard (IrDA)) or a second network 229 (e.g., a long-range communication network such as a cellular network, the Internet or a computer network such as a local area network or a wide area network (WAN)). These different types of communication modules may be implemented as a single component (e.g., a single chip) or may be multi-component (e.g., a plurality of chips).

[0118] Fig. 2A and B illustrate the absorption spectra of total hemoglobin (tHb), oxygenated hemoglobin (HbO ₂ ) and glycated hemoglobin (HbA1c), and more specifically, the dependences of the molar absorption coefficient (attenuation) on the radiation wavelength are shown. In this case, Fig. 2A illustrates the theoretically calculated molar absorption coefficients of tHb and HbO ₂ , and Fig. 2B illustrates the theoretically calculated and practically obtained molar absorption coefficients of HbA1c at different levels.

[0119] The graphs are given according to the publications Hossain S., Satter S., Kwon T.-H., Kim K.-D., Optical Measurement of Molar Absorption Coefficient of HbA1c: Comparison of Theoretical and Experimental Results, Sensors 2022, 22, 8179, https://doi.org/10.3390/s22218179 /1/ and Hossain S., Kwon T.-H., Kim K.-D., Estimation of molar absorption coefficients of HbA1c in near UV-Vis-SW NIR light spectrum, JKICS 2021, 46, 1030-1039 121. Since diabetes is a serious group of diseases that can cause many complex problems including sudden death, a reliable diagnostic tool for diabetes, such as glycated hemoglobin testing, is needed. In order to optically estimate the amount of HbA1c present in red blood cells, it is necessary to know the molar absorption coefficient of HbA1c. In the referenced publications, the molar absorption coefficient of HbA1c was calculated based on the transmission spectrum of glycated hemoglobin dissolved in water, and the range values were calibrated with the molar absorption coefficients of glycated hemoglobin at 535 nm (green light) and 593 nm (yellow light). The molar absorption coefficient in /1/ was evaluated in the range from 300 to 1100 nm and a characteristic peak of HbA1c was obtained at 411 nm and two more peaks at 540 nm and 576 nm (see Fig. 2A).

[0120] Table 2 below shows the molar absorption coefficients of HbA1c obtained according to publications /1/ and /2/.

The characteristic and secondary peaks described in publications /1/, /2/ and /3/ are confirmed by other studies, and the calculated molar absorption coefficient in the range from near UV to shortwave near IR radiation helps to ensure the possibility of non-invasive assessment of HbA1c.

[0121] It is clearly seen from Fig. 2A that in the range of 500-600 nm, the molar absorption coefficients of oxyhemoglobin (HbO ₂ ) and total hemoglobin (tHb) are very similar. It is clearly seen from Fig. 2B that the visible light - near infrared (VIS-NIR) absorption bands of HbO ₂ and glycated hemoglobin (HbA1c) coincide (see peaks at 545 nm and 579 nm for level 1, at 544 nm and 577 nm for level 2). Therefore, there is a strong mutual influence of the absorption spectra of oxygenated and glycated hemoglobin. Therefore, separate identifiers are needed for _HbO2 and HbA1c, they must be measured in parallel and decoupled during processing, and the glycated hemoglobin estimation algorithm must have multi-output prediction that takes into account glycated hemoglobin as well as at least oxygenated hemoglobin and total hemoglobin.

[0122] Figure 3A illustrates the composition of blood and provides examples of some of the major components of blood (figure taken from the article available at https://www.cellgs.com/blog/what-is-in-blood-and-where-is-it-made.html). The appearance of blood in a test tube after addition of an anticoagulant and centrifugation is shown in the left part of Figure 3A. The yellowish plasma is located on top of the red blood cells and is separated from the red blood cells by a layer of white blood cells known as the buffy coat. The components shown are arranged roughly in order of size, ranging from cells to small molecules including water. Figure 3B schematically illustrates graphs of changes in the concentration of the so-called fast-changing blood components (e.g., HbO ₂ , glucose) and slowly changing blood components (e.g., HbA1c). The graphs show that a noticeable change in the concentration of the former occurs within minutes, and the latter within weeks. Therefore, to detect glycated hemoglobin, it is necessary to separate the features of pulse signals characterizing rapidly and slowly changing parameters of the blood component. Separation of the features of pulse signals characterizing HbA1c from those characterizing other blood components using a photoplethysmographic (PPG) device operating in the VIS-NIR range is possible only by analyzing the response of the characteristics (features) of PPG signals to trigger events.

[0123] A trigger event is an event in a person's daily activities that changes some blood parameters, such as the level of oxygenated hemoglobin, glucose level, total hemoglobin, etc. Trigger events can be intentionally created or natural. Intentionally created trigger events include an occlusion test, holding your breath, or very tight temporary fixation of a strap of a wearable device on a part of the user's body. An occlusion test is a type of functional load that consists of a short-term blockade of arterial and/or venous blood flow in the vessels of a limb with simultaneous observation of their condition before, during, and after occlusion. For example, an occlusion test with a cuff on the arm leads to a disruption of blood microcirculation, i.e., the absence of microcirculation and an increase in heart rate. Events occurring in everyday activities or so-called natural trigger events can include physical activity, eating, moving the user from a room to fresh air, being at a high altitude, sleeping, etc.

[0124] Figure 4 illustrates the design of the intentional breath-hold experiment. This experiment is described in Michael Chan, Venu G. Ganti, J. Alex Heller, Calvin A. Abdallah, Mozziyar Etemadi and Omer T. Inan, Enabling Continuous Wearable Reflectance Pulse Oximetry at the Sternum, Biosensors 2021, 11, 521, pp. 1-19, https://doi.org/10.3390/bios11120521/4/, and investigated breath-holding to induce hypoxemia and significant changes in SpO ₂ . While holding the breath, blood oxygenation was monitored (sensor on the index finger of the right hand), PPG signals, electrocardiogram (ECG) and seismocardiogram (SCG) were recorded. Blood oxygenation or the degree of blood oxygen saturation is the percentage of oxyhemoglobin in the blood, that is, the ratio of the amount of oxyhemoglobin to the total amount of hemoglobin. Oxygenation of 95-100% is normal, but may vary if a person has lung disease. Oxygenation of less than 90% is considered low and is called hypoxemia, in which case the body requires oxygen supplementation. The causes of low blood oxygen saturation are anemia, sleep apnea, smoking, lung diseases (asthma, pulmonary emphysema, etc.), viral infections (COVID-19, etc.), air quality, reduced heart capacity, taking strong painkillers.

[0125] In the breath-hold test, subjects performed 10 end-expiratory breath holds while sitting upright, with a one-minute rest between breath holds. One minute was found to be sufficient for SpO ₂ to return to baseline. Important oxygenation/deoxygenation events were noted after data collection. Figure 4 illustrates a patient undergoing a breath-hold test. The locations of the wearable patch biosensor and the main Biopac sensors for obtaining electrophysiological data are shown. The relative size of the biosensor is shown with a pre-made ECG electrode and a coin, the photodiode (PD) has an area of 4.5 mm ² . PD1 is on the left and PD2 is on the right of the subject. In parallel, a biosensor on a wearable patch was also attached to the middle of the subject's sternum and ECG was collected from a single lead, and two sets of multi-wavelength PPG sensors with LEDs of red, IR and green light wavelengths, i.e. 660 nm, 950 nm and 526 nm respectively, were used.

[0126] Fig. 5, taken from publication /4/, illustrates filtered high-quality physiological signals obtained during breath holding and normal breathing. As noted above, the following information was analyzed: respiratory flow data, blood oxygenation (SpO ₂ ), electrocardiogram (ECG), seismocardiogram (SCG), PPG signals at different wavelengths.

[0127] It should also be noted that oxygenation/deoxygenation events do not occur simultaneously at different sites in the body, and, for example, SpO ₂ measured at a fingertip typically lags behind SpO ₂ measured at central sites. This delay can be explained in part by the oxygen conservation effect of breath holding. Like the diving response, breath holding also results in bradycardia and peripheral vasoconstriction, which reduces peripheral oxygen consumption and redistributes blood flow to vital organs such as the brain and heart. The combined effect results in a delay in measuring deoxygenation with a fingertip pulse oximeter compared to a pulse oximeter located closer to the heart or brain. This is why SpO ₂ does not immediately fall during a breath hold, but may initially rise slightly after the breath hold. Further, during the breath hold, the blood oxygen saturation level remains unchanged for a certain period of time. The end of this phase should be considered a change in saturation by 1% ("oxygenation stability phase"). It is believed that the duration of this phase reflects the intensity of the oxidative processes. Then the hypoxemic phase begins, lasting from the moment of a decrease by 1% until the end of breath holding, i.e. until the first inhalation, and characterized by a progressive decrease in blood oxygen saturation. The duration of this phase and the depth of the decrease in HbO ₂ saturation characterize the body's resistance to an increasing decrease in oxygen and an increase in carbon dioxide in the blood. Note that after the start of inhalation, blood oxygen saturation not only does not increase, but even decreases for some time. This is due to the fact that after inhalation, arterial blood does not immediately reach the photometric area, but only after a few seconds, which corresponds to the blood flow velocity (see the graph of Sp0 ₂ changes in Fig. 5).

[0128] From the respiratory flow (upper waveform in Fig. 5), the segments of normal breathing (oscillating part, segments on the right and left) and breath-holding (flat part, segment in the center) are easily distinguished. Based on the actual SpO ₂ (second waveform from the top on the left side of Fig. 5), three separate time stamps were recorded for each deoxygenation event: the onset, the nadir, and the end. The onset of deoxygenation is defined as the point at which SpO ₂ begins to fall sharply (the rate of decrease in SpO ₂ exceeds 0.5%/cardiac cycle for 3 consecutive cardiac cycles). The nadir of deoxygenation is defined as the lowest SpO ₂ during the deoxygenation period. The end of deoxygenation is defined as the point at which SpO ₂ returns to the baseline level. In general, the nadir and the end of deoxygenation can be easily determined.

[0129] The given example of a breath-holding test confirms the possibility of a controlled effect on blood parameters, for example, the SpO ₂ level. Simultaneously with the change in blood parameters, the signs of signals characterizing blood parameters change, in particular, the change in PPG signals during breath-holding was studied in the indicated publication.

[0130] Fig. 6A, 6B are presented to confirm the possibility of practical implementation of a wearable device for determining separate levels of HbO ₂ and HbA1c.

[0131] The top graph of Fig. 6A schematically illustrates the basic principles of performing an occlusion test. Occlusion at full pressure is typically maintained for 60 seconds, then the pressure is released. The bottom graph of Fig. 6A illustrates an example of a testing protocol using a signal characterizing the level of oxygenated hemoglobin. The y-axis scale is presented in optical density units. The asterisks * indicate the approximate times when the pulse value was calculated at rest, immediately after ischemia, and during reactive hyperemia. Data are given from the publication A. A. Sann, K. K. McCully, Interpretation of Near-Infrared Spectroscopy (NIRS) Signals in Skeletal Muscle, J. Funct. Morphol. Kinesiol. 2019, 4(2), 28, 2019.

[0132] Fig. 6B illustrates graphs of changes in (a) oxygenated hemoglobin concentration, (b) deoxygenated hemoglobin concentration, (c) total hemoglobin concentration, and (d) blood oxygen saturation as a function of time for a typical subject under a forearm occlusion and reactive hyperemia protocol. Data taken from X. Chen, W. Lin, C. Wang, S. Chen, J. Sheng, B. Zeng, M Xu, In vivo real-time imaging of cutaneous hemoglobin concentration, oxygen saturation, scattering properties, melanin content, and epidermal thickness with visible spatially modulated light, Biomed Opt Express, 8, 8(12):5468–5482, 2017.

The graphs shown in Fig. 6B confirm that blood parameters such as oxygenated hemoglobin, deoxygenated hemoglobin, total hemoglobin, and blood oxygen saturation change significantly during the occlusion period and therefore can be subject to controlled change in the same manner as during breath holding.

[0133] The key features of the claimed invention will be disclosed below.

[0134] Fig. 7 illustrates a first key feature (KOI) consisting in detecting features of PPG signals characterizing rapidly and slowly changing parameters of blood components. Using a wearable device equipped with at least one PPG sensor configured to irradiate biological tissues of the user, including skin, bones, blood, blood vessels, with radiation of at least two different wavelengths and detect at least two PPG signals at said at least two different wavelengths in a reflection mode, the PPG signals are detected. Then, characteristic features of the detected PPG signals are extracted and changes in the features of the waveform of the PPG signals characterizing rapidly and slowly changing parameters of the blood components are determined. As noted above, slowly changing parameters of the blood components are parameters of components whose level noticeably changes over several weeks, and these include, for example, the level of glycated hemoglobin. At the same time, rapidly changing parameters of blood components are the parameters of components characterized by a change in their level within a few seconds to several minutes, and include, for example, the glucose level, the oxygenated hemoglobin level or the oxygenation of the blood. The determination of the change in the signs of PPG signals is possible due to the continuous detection of signals by the wearable device during the user's daily activities and the events that occur that cause changes in the rapidly changing parameters of blood components.

[0135] Fig. 8 illustrates the second and third key features (KO2 and KO3), which consist, respectively, of compiling a database based on trigger events and estimating the level of glycated hemoglobin based on information from the database, taking into account the influence of other blood components. Compiling the database includes collecting a matrix of sets of rapidly and slowly changing features of the waveform of PPG signals, correlated with the values of rapidly and slowly changing parameters of blood components. For non-invasively estimating the level of a slowly changing parameter of a blood component, such as glycated hemoglobin, information from the compiled database and sets of spectral-temporal and spectral-frequency features of PPG signals obtained using the wearable device of the present invention are used. The level of glycated hemoglobin is estimated using a predictive model. In this case, a slowly changing parameter of a blood component, such as glycated hemoglobin, is characterized by one set of features {z ^DC ₀ , z ^DC _l …z ^DC _n }, and a rapidly changing parameter of a blood component, such as oxygenated hemoglobin, is characterized by another set of features {z ^AC ₀ , z ^AC _l …z ^AC _n }. Thus, the invention provides for separate accounting of rapidly and slowly changing parameters of blood components, which means that the level of glycated hemoglobin is assessed by a wearable device taking into account the influence of other blood components, for example, as in this case, oxygenated hemoglobin.

[0136] Fig. 9, 10A, 10B and 10B illustrate in more detail the first stage KO1, which consists of detecting the parameters of PPG signals characterizing the rapidly and slowly changing parameters of blood components.

[0137] It should be noted that the irradiation of biological tissue is carried out continuously, for example, for a day or more, three hours or more, two hours or more, 1 hour or more, 1 minute or more, but is not limited to this. Accordingly, the detection of PPG signals of different wavelengths (λ _n ) in the VIS-NIR spectral region is also carried out continuously in the reflection mode using a wearable device equipped with at least one PPG sensor. The duration of irradiation is determined by the need to identify the trigger event and register PPG signals during the trigger event in order to separate rapidly and slowly changing signal features.

More specifically, the wearable device according to the invention is capable of providing radiation of several wavelengths in the range of 400-1100 nm, i.e. in the VIS-NIR range. In particular, the radiation scattered/reflected from the analyzed volume of biological tissue is measured at specified target wavelengths, such as 670 nm and 940 nm. In another embodiment, the wavelength of the radiation used to irradiate the biological tissue is in the range of 495-570 nm (green light). However, the present invention is not limited to the specific wavelengths indicated and in embodiments, radiation of several wavelengths in the range of 400-1100 nm can be used. Thus, since the wearable device provides radiation and detection of signals in a continuous mode during the user's normal activity, it is possible to ensure the removal of a plurality of sets of signals at different wavelengths in different periods of time.

[0138] At the first stage, the features of the detected PPG signals presented in the time domain (time series) are extracted, obtaining a set of spectral-temporal features of the signals. One of the traditional approaches to processing the PPG signal in the time domain is to measure the area under the curve of the AC component of the PPG signal, i.e. to determine the integral area of the PPG signal shape. It is believed that the intensity of the AC component measured in this way is proportional to the optical transmittance of the measured tissue volume and can be recalculated into absorption in this area according to the Bouguer-Lambert-Beer law. This is the simplest approach to measurements, usually applied in the transmission mode. However, the signals can have approximately the same integral area, but different curve shapes. Since this approach does not take into account the information about the pulse shape, the measurement of optical transmittance and absorption in the reflection mode, which is usually used in a wearable device, is performed with a large error.

[0139] Fig. 10A schematically illustrates a second conventional approach to processing a PPG signal in the time domain, which consists of estimating the waveform of the PPG signal. The upper graph of Fig. 10A shows an exemplary pulse waveform. Typically, the features extracted in the time domain correspond to critical points of the PPG signals. The first peak of the pulse wave, corresponding to the anacrotic period of the pulse wave, is formed during the systole period (the amplitude of the "systolic peak" is designated as X). The amplitude value of the anacrotic phase is also called the pulse wave amplitude and corresponds to the stroke volume of blood during cardiac output, thus providing indirect information on the degree of the inotropic effect. The second peak of the pulse wave, corresponding to the dicrotic period of the pulse wave, is formed due to the reflection of the blood volume from the aorta and large trunk vessels and partially corresponds to the diastolic period of the cardiac cycle (the amplitude of the "diastolic peak" is designated as Y). The dicrotic phase provides information on the vascular tone. The peak of the pulse wave corresponds to the largest volume of blood, and its opposite part - to the smallest volume of blood in the examined area of biological tissue. On the descending front of the pulse wave, a depression is noticeable - a dicrotic incisure, which corresponds to the closure of the aortic valve ("dicrotic notch", designated as Z). In addition, the pulse repetition period is ( _{t5 -t4} ). The nature of the pulse wave depends on the elasticity of the vascular wall, pulse rate, volume of the examined area of biological tissue, the width of the lumen of the vessels, etc. It is believed that the frequency and duration of the pulse wave depend on the characteristics of the heart, and the size and shape of its peaks - on the state of the vascular wall. A pulse wave in which the systolic and diastolic peaks are clearly expressed, as well as a depression between them, is characteristic of young, healthy people with a strong cardiovascular system. In most cases, these features of the pulse wave are quite clearly expressed, as illustrated in Fig. 10A, and such a wave form is characteristic of relatively healthy adults. However, in some patients, the systolic and diastolic peaks, as well as the trough between them, may be difficult to determine or even indistinguishable, which is typical for elderly people with a weak cardiovascular system or young unhealthy people. Nevertheless, in all cases, the shape of the wave signal curve is characterized by the same parameters that have a clear physiological meaning, as described above.

[0140] Thus, as an example, the features of the detected PPG signals extracted in the first stage are (see Fig. 10A), but are not limited to:

X - amplitude of the systolic peak;

Y - amplitude of the diastolic peak;

Z - dicrotic notch;

(t ₅ -1 ₄ ) - period of the pulse signal.

[0141] The detected PPG signals are then transformed from the time domain to the frequency or time-frequency domain. In general, several representations of the signal are possible: in the time domain, when f(t) is expressed as a function of time, in the frequency domain, when the spectrum (i.e., the amplitudes of the various harmonics) is determined, or in the time-frequency domain. The signals can be transformed from the time to the frequency or time-frequency domain using various types of transformations. According to an embodiment of the invention, the tool for representing the detected signal by exponential components is the Fourier transform. The function F(ω) is the direct Fourier transform of the signal f(t). F(ω) represents the signal f(t) in the frequency domain. The time representation defines a certain signal at each moment in time, while the frequency representation characterizes the relative amplitudes of the frequency components of the signal, i.e., the harmonics. After the Fourier transform and the representation of the signal in the frequency domain, its amplitude-phase characteristics of harmonics (instantaneous spectrum, spectral density, etc.) can be obtained. For example, the Fourier transform allows each PPG pulse to be transformed into a set of amplitudes and phases, so the resulting set of features contains information necessary for the predictive model used to determine the level of glycated hemoglobin. In this case, for calculations, the set of signal features can be conveniently described by complex numbers. Any of the specified representations completely determines the signal. However, the invention is not limited to the use of the Fourier transform and other transforms can be used. For example, the transformation of signals from the time to the frequency domain can also be performed by the Hilbert-Huang transform, and into the time-frequency domain - by the wavelet transform or bilinear time-frequency distribution. The transformation of the signal into the frequency domain is necessary, among other things, to overcome the above-described problems that may arise when processing signals in the time domain.

[0142] Thus, in the second stage, the features of the detected PPG signals in the frequency domain are extracted, which are, for example, the amplitudes of the harmonics (see Fig. 10B), but are not limited to this;

I _0Red , I _0Green , I _0IR - intensities of the raw optical signal from the PPG sensor at a given wavelength (red, green, PC);

f ₀ - the amplitude of the PPG signal at the heart rate frequency (i.e. the 1st peak in the PPG spectrum or the fundamental harmonic);

f ₁ - the amplitude of the PPG signal at double the pulse frequency (2nd peak in the spectrum);

f ₂ - the amplitude of the PPG signal at a triple heart rate (3rd peak in the spectrum).

[0143] Finally, since, as noted earlier, the signal detection is performed in a continuous mode, in the third stage it is possible to extract the starting points of the sliding window in the time domain, as shown in Fig. 10B, which represent the following features of the detected PPG signals:

S ₁ - window step length;

W ₁ - window length;

K ₁ - segment length;

S _n - window step length;

W _n - window length;

k _n - segment length.

In particular, the sliding time window method makes it possible to determine the pulse rate and detect changes, including pulse wave anomalies. For example, in the publication of S. M. Isuru Niroshana, Satoshi Kuroda, Kazuyuki Tanaka, Wenxi Chen, Beat-wise segmentation of electrocardiogram using adaptive windowing and deep neural network, Scientific Reports, 2023, 13:11039, https://doi.org/10.1038/s41598-023-37773-у /5/, methods for timely detection of anomalies and automatic interpretation of the electrocardiogram (ECG) are described. Rhythm segmentation is one of the important steps in ensuring the reliability and accuracy of many automatic ECG classification methods. The publication discloses a robust method for ECG rhythm segmentation using a convolutional neural network (CNN) model with an adaptive (sliding) window algorithm. The proposed adaptive window algorithm can recognize cardiac cycle events and perform segmentation, including regular and irregular beats, from an ECG signal with sufficiently accurate boundaries. Over the past decades, various methods for automatic computer classification of ECG have been proposed and implemented. Typically, three stages are performed: (i) pre-processing, (ii) heartbeat segmentation, (iii) classification by heart rhythms. Automatic detection and segmentation of ECG rhythm using R-peak (a critical event in detecting one heart rhythm) is one of the important stages in many ECG-based algorithms, including cardiac diagnosis, heart rate variability analysis, and ECG-based authentication. The principles set forth in publication /5/ are applicable to the processing of PPG signals according to the present invention. Thus, by dividing continuous cycles of PPG signals into segments using a sliding window, it is possible to extract from the signals the starting points of the adaptive window, as shown in Fig. 10B, in different periods of time, i.e., in different sections of the pulse signal, and therefore it is possible to determine the beginning of the trigger event.

The duration of the sliding time window can be adjusted during operation if, for example, the features corresponding to slowly changing component parameters change too slowly, or the features corresponding to rapidly changing component parameters change too quickly. In general, the photoplethysmographic signal is received continuously, and the sliding time window moves along this signal, determining the presence of events (trigger events) in each of the sections during which it is possible to separate the rapidly and slowly changing features corresponding to rapidly and slowly changing parameters of blood components. In this way, it is possible to track how the features corresponding to changes in the glycated hemoglobin level change when moving from one time window (PPG signal section) to another.

[0144] After extracting the features of the detected PPG signals, the second stage of KO1 is carried out, which consists of extracting changes in the features of the waveform of the PPG signals, characterizing the rapidly and slowly changing parameters of the blood components.

[0145] As noted previously, rapid changes in PPG waveform features characterize rapidly changing blood component parameters, such as when blood component levels or concentrations change over periods of less than a minute to several minutes, while slow changes in PPG waveform features characterize slowly changing blood component parameters, such as when blood component levels or concentrations change over periods of several days or weeks.

[0146] Fig. 11 illustrates a graph schematically showing the change in the levels of slowly changing features (f^(t)) of the PPG signal and rapidly changing features (f*(t)) of the PPG signal during continuous monitoring of the PPG signals, in particular, the dependence of the signal amplitude on time.

[0147] To process the signals and obtain the results, a machine learning model (ML model) is used and a predictive model is trained. The input data for the pre-trained ML model are raw PPG signals at different wavelengths (λ _n ) [x, y, t ₅ , z, f _0Red , f _0Green , f _0IR ]. The output data are fast and slowly changing features of the PPG signal waveform, characterizing the fast and slowly changing parameters of the blood components {z ^AC ₀ , z ^AC _l …z ^AC _n }.

The present invention provides for the extraction of PPG signal features corresponding to rapidly and slowly changing parameters of blood components. The effect consists in obtaining unique sets of rapidly and slowly changing waveform features of PPG signals characterizing rapidly and slowly changing parameters of blood components.

[0148] Fig. 12 schematically illustrates a second key feature of the present disclosure, consisting in collecting a data set based on trigger events. The reference database is formed by data provided by a wearable device with PPG sensors, reference data provided by laboratory blood tests, data on trigger events, including events occurring in the daily activities of users, and metadata. Natural trigger events or events occurring in daily activities can include physical activity, sleep, eating, moving from indoors to outdoors, being at a high altitude (e.g., above 2000 meters), etc. Intentional trigger events can include an occlusion test, holding your breath, fixing the strap of the wearable device very tightly. Intentional trigger events are typically used to calibrate the wearable device, for example, during its first use, since they provide controlled changes in PPG signals and blood components. Metadata may include gender, age, height, weight of the wearable device user, etc.

[0149] During compilation of the reference database, data is collected from a plurality of wearable device users during natural and intentional trigger events. In this case, for each user, a laboratory test is performed on blood, which is taken during the trigger event, and PPG signals are collected and processed to obtain PPG signal features for each wavelength λ _m in response to the events that occur, and PPG signals are collected and processed to obtain PPG signal features for each wavelength λ _m in response to intentional trigger events, after which the obtained PPG signal features are compared with the levels or concentrations of rapidly and slowly changing parameters of blood components measured during the blood test.

The end result is a reference database in which sets of PPG signal features are matched with levels or concentrations of rapidly and slowly changing blood component parameters during various events in the daily activities of multiple wearable device users.

[0150] Fig. 13 illustrates the third key feature of the present invention, which consists in estimating the level of glycated hemoglobin taking into account the influence of other blood components. At this stage, a predictive model is constructed for accurately estimating the level of glycated hemoglobin as a blood component with a slowly changing parameter using a wearable device. For this purpose, a compiled database is used in which rapidly and slowly changing features of the PPG signal waveform are compared with concentrations of rapidly and slowly changing parameters of blood components. In this case, sets of slowly changing features of the PPG signals correspond to specific values of slowly changing parameters of blood components, for example, HbA1c, measured by a laboratory method, while sets of rapidly changing features of the PPG signals correspond to specific values of rapidly changing parameters of blood components, for example, HbO ₂ , measured by a laboratory method. The database is used to identify functional dependencies between data sets and the values of blood component parameters, and the values of the coefficients included in them are determined in the calibration process using machine learning methods, after which a predictive model is constructed and trained.

[0151] According to the present invention, by separating the fast and slow changing features of the PPG signals, it is possible to separate, for example, the level of glycated hemoglobin from the level of oxygenated hemoglobin. In this case, a machine learning model (ML model) is used, the input data for which are the fast and slow changing features of the waveform of the PPG signals, characterizing the fast and slow changing parameters of the blood components, and the output data are the levels of glycated hemoglobin.

[0152] Fig. 14 illustrates the features of determining the events occurring in the daily activities of the user of the wearable device according to the present invention. The daily activities or ordinary life of the user of the wearable device, like any person, consists of periods of sleep, rest, eating, reading, playing sports, etc., that is, some physiological events occur from time to time. As disclosed above, with almost any event occurring in the daily activities of a person, changes in the values of some parameters of blood components occur, for example, after eating, the glucose level increases, during intense physical activity, the glucose level, on the contrary, falls and blood oxygenation decreases sharply. Blood oxygenation also decreases during sleep. Thus, during periods of activity, a change in the levels of rapidly changing parameters of blood components occurs, while the level of glycated hemoglobin remains virtually unchanged. Because the wearable device continuously monitors PPG signals during the user's normal life, it can detect the onset of an activity period or trigger event in which a rapidly changing blood component parameter will change and provide PPG signals at that time to automatically estimate the glycated hemoglobin level taking into account the influence of other blood components.

[0153] Fig. 15A schematically illustrates embodiments of an exemplary display of the HbA1c level measurement results during the day and the deviation of the HbA1c measurement results over several months on the screen of the wearable device according to the invention in the S-Health application. As illustrated in Fig. 15A, the user of the wearable device can see various data with the HbA1c level measurement results for different periods of time on the screen of the wearable device, including, for example, specific results of single measurements and average values for the day, month, year, etc. Fig. 15B schematically illustrates a graph displayed on the screen of the wearable device in the S-Health application, allowing monitoring of the deviation of the glycated hemoglobin level from the target value. With long-term monitoring of the HbA1c level, for example, over up to several months, it is possible to track the deviation from the target zone of the normal glycated hemoglobin level, as well as the crossing of the alert thresholds for high and low HbA1c levels. It is important to monitor any deviations of the specified level from the target zone. In this case, when monitoring the level of glycated hemoglobin, the following situations are possible. When determining a high or low level of HbA1c, going beyond the target zone, the user of the device is informed of the result and a preliminary diagnosis, for example, a predisposition to diabetes, and general recommendations can be displayed on the screen: changing the diet, increasing activity, walking in the fresh air, or a recommendation to see a doctor.

[0154] At the same time, short-term deviations of the glycated hemoglobin level from the norm are not critical and can be corrected by the user of the device using the recommendations displayed on the display screen. On the contrary, long-term deviations of the glycated hemoglobin level from the norm can indicate serious health problems and recommendations on the mandatory visit to a specialist are displayed on the display screen. In addition, the data obtained from the monitoring of the glycated hemoglobin level can be sent to the doctor directly from the measuring device.

The advantages of such monitoring include relatively low research costs, provision of a long-term period of continuous monitoring of blood parameters, which ensures obtaining reliable measurement data and establishing a more accurate diagnosis of a person based on the data obtained.

[0155] Fig. 16A illustrates an example of practical implementation of the claimed invention, namely, a diagram of an experiment with breath holding, in which changes in PPG signals were monitored for different users and the level of glycated hemoglobin was determined using an experimental sample of the claimed wearable device. As disclosed above, breath holding was used as a trigger event to cause changes in rapidly changing parameters of blood components, and therefore changes in PPG signals. PPG signals were recorded continuously for several minutes. In this case, user 1 breathed calmly for 120 seconds, then held his breath on inhalation for 30 seconds, then breathed calmly again, user 2 breathed calmly for 120 seconds, then held his breath on exhalation for 30 seconds, then breathed calmly again, and user 3 breathed calmly for 60 seconds, then held his breath on inhalation for 30 seconds, then breathed calmly again.

[0156] The collected clinical data set contained information on 100 measurements performed using a sample of the wearable device of the present invention on 50 volunteers during a breath-holding experiment and was used to train a predictive model and further estimate the level of glycated hemoglobin. When collecting reference data, a laboratory colorimetric method was used to determine the reference level of glycated hemoglobin. In addition, the said data set contained volunteer metadata, as well as information obtained using the AngioScan-01M device, which allows monitoring the state of the cardiovascular system, including heart function (heart rate, rhythm), the state of the vascular wall (elasticity, rigidity, wave type), blood oxygen saturation, and stress level.

[0157] Fig. 16B illustrates the results of predicting the glycated hemoglobin level in the blood of a wearable device user based on a clinical data set (reference database) using an experimental sample of the wearable device of the present invention. The upper graph of Fig. 16B is a bar chart illustrating the number of users with specific HbA1c levels with distribution by gender. The range of reference values for the glycated hemoglobin level was 4.4-5.8 g/L. In the lower graph of Fig. 16B, the X-axis shows the HbA1c level of the test subject measured by the laboratory method, and the Y-axis shows the HbA1c level measured by the predictive model. The X=Y line (passing in the middle) shows the reference data when the reference and measured values coincide. The lines plotted above and below the X=Y line show the measurement error range of +/- 0.6 g/L. The graph also shows the data used for training, the data obtained by blood testing, and the supporting data, i.e. the results of measuring the glycated hemoglobin level using the experimental sample of the wearable device of the present invention. In this case, a gradient descent machine learning regression model was used to train the predictive model, and k-fold cross-validation was used to evaluate the predictive performance.

[0158] The Pearson correlation coefficient for the measurement results shown in Fig. 16B is R=0.7, which indicates a weak positive relationship between the true HbA1c level and the predicted HbA1c level. It should be noted that this experimental example involved volunteers with a normal HbA1c level of 4.4-5.8 g/L. It is obvious that the accuracy of the prediction will increase both as the prediction model is refined and as information with measurement results on test subjects whose HbA1c level deviates from normal values accumulates in the database. Thus, it is expected that the wearable device of the present invention will be able to determine the glycated hemoglobin level with high accuracy.

[0159] The solution to the problem of assessing the values of the parameters of the user's blood components based on the characteristics of PPG signals is based on the identification of empirical patterns in training data placed in a common database using machine methods.

[0160] In an embodiment of the present invention, the predictive model is created for a wearable device located on the wrist, since the training of the predictive model was based on data recorded by sensors of the wearable device located on the wrist. In other embodiments of the present invention, it is possible to train the predictive model taking into account the placement of the wearable device on other parts of the body, for example, on a finger. The predictive model described herein can be implemented as software including one or more instructions that can be executed by the processor 203 of the wearable device 201 or external processors and/or processing devices.

[0161] In the above-described process of training the predictive model, profile data of a plurality of users (test subjects), sets of measurement data previously obtained for a plurality of users, and sets of reference data previously measured by a laboratory method were used, as described in detail above.

[0162] As discussed below, the database may be continuously updated and, in addition, the predictive model is refined through machine learning.

In the process of machine learning, methods of selection and processing of characteristics and corresponding discrete frequencies can be used:

- in the Relief-F method, the vector of feature weights (characteristics) is calculated and normalized, and then features whose weight exceeds the value of a given threshold are selected;

- the Correlation-based Feature Selection (CFS) method combines an evaluation formula with an appropriate correlation measure and a heuristic search strategy;

- the Fast Correlation Based Filter method starts working with a full set of features, uses a measure of symmetric uncertainty to determine dependencies between features and allows selecting a subset by searching and sequentially eliminating uninformative features;

- the Sequential Forward Feature Selection (SFFS) method at each iteration adds to the set a feature that provides the best recognition efficiency for a given iteration;

- the Mutual Information method defines a nonlinear correlation dependence instead of calculating the Pearson correlation “feature-feature” and “feature-label”;

- basic algorithms of machine learning (artificial intelligence);

- a combination of the specified regression methods and any derivative regression methods based on the basic algorithm can be used as a working solution:

- decision trees (Decision Tree) / random forest (Rnom Forests);

- Support Vector Machines;

- linear analysis;

- deep learning methods - artificial neural networks.

[0163] The advantageous effects of using the machine learning method are the ability to determine the glycated hemoglobin level for a user whose data is not included in the database and an increase in the accuracy of determining the glycated hemoglobin level by taking into account the influence of other blood components on the measurement results.

[0164] Once the predictive model is constructed, it can be used to estimate the level of glycated hemoglobin. In this case, according to the present disclosure, when estimating HbA1c, the influence of other blood components on the measurement results is taken into account. As a result, a more accurate estimate of the level of glycated hemoglobin is possible than in conventional methods of the prior art.

[0165] In summary, the present invention provides an accurate assessment of the glycated hemoglobin level taking into account the influence of other blood components without the need for blood sampling, i.e., in a non-invasive manner, due to the following features:

KO1: Continuous signal acquisition using PPG sensors. I. Irradiation of biological tissue and detection of PPG signals during daily activities, including periods of ongoing events.

II. Extraction of PPG signal features.

III. Determination of the event occurring in everyday activities, during which the values of the parameters of blood components change.

IV. Identification of rapidly and slowly changing parameters of blood components, characterizing rapidly and slowly changing parameters of blood components.

V. Extraction from PPG signals detected during an ongoing event of rapidly and slowly changing features characterizing rapidly and slowly changing parameters of blood components, respectively, and formation from them of a set of rapidly and slowly changing features of PPG signals.

KO2: Compiling a reference database based on events occurring and intentional trigger events.

VI. Obtaining, using a wearable device, sets of PPG signal features recorded during events occurring in everyday activities and intentional trigger events.

VII. Determination of blood component parameter (BPV) values (reference data) by conducting laboratory tests on blood collected during events occurring in daily activities and intentional trigger events.

VIII. Compilation of a database in which the feature sets of PPG signals (VI) are compared with BPV (VII).

KO3: Assessment of the level of glycated hemoglobin taking into account the influence of other blood components.

IX. Building and training a predictive model using machine learning.

X. Accurately estimate the glycated hemoglobin level taking into account the influence of other blood components for new users without conducting a blood test by comparing a set of features of the pulse signal for a new user, obtained using the mentioned PPG sensors of the wearable device, with matrices of sets from a pre-compiled reference database.

It should be noted that, according to the present invention, a set of features of PPG signals corresponding to slowly and rapidly changing parameters of blood components, recorded over a certain period of time (time window), characterizes in the only possible way the features of rapidly changing and slowly changing parameters of blood components of a certain user under certain conditions.

Industrial applicability

[0166] Worldwide, approximately 0.5 billion people suffer from diabetes mellitus. This number is projected to more than double to 1.3 billion in the next 30 years. The disease can affect men, women, and children of all ages in all countries. The following target groups should monitor the course of the disease or their predisposition to diabetes:

- healthy people for the purpose of preventive control;

- children with diabetes;

- patients with diabetes with abnormal renal glucose threshold;

- patients with type I diabetes mellitus, insulin-dependent;

- pregnant women with type II diabetes;

- people over 45 years old;

- people with a body mass index over 25;

- people with a hereditary risk of disease;

- other people when changing their diet or other habits.

Also at risk are people who:

- high blood pressure;

- vision problems;

- cardiovascular disease and/or history of stroke;

- kidney disease.

To prevent the risk of developing type II diabetes, you should follow a healthy diet, lose excess weight, and exercise regularly. Patients who exercise regularly can reduce their risk of developing diabetes by half. Patients who have lost at least 5% of their body weight can significantly reduce their risk of developing diabetes.

[0167] As noted, blood glucose levels are monitored using two measures: glycated hemoglobin (HbA1c) and instantaneous glucose. HbA1c is a long-term measure of glucose over the past 1-3 months and is not affected by short-term changes in blood glucose levels caused by food intake, exercise, etc. on the day of testing. Instantaneous glucose is a short-term measure of glucose at a specific time during testing, such as fasting or postprandial blood sugar, and is subject to fluctuations due to food intake and stress.

[0168] Thus, the wearable device disclosed in the present application, capable of non-invasive, continuous and/or on-demand determination of glycated hemoglobin levels, suitable for non-professional use, is expected to be in demand in the market.

Implementation option 1

[0169] The present embodiment relates to the use of a wearable device configured to determine the level of glycated hemoglobin when performing a breath-holding test, which is described earlier. Intentionally holding the breath leads to a decrease in the supply of oxygen to the blood, causing a change in rapidly changing parameters of blood components, which facilitates an accurate measurement of the level of glycated hemoglobin.

Implementation option 2

[0170] The present embodiment relates to the use of a wearable device configured to determine the level of glycated hemoglobin by performing a test with a tight fixation of the strap of the wearable device. The intentional fixation of the strap causes a change in the rapidly changing parameters of the blood components, which facilitates an accurate measurement of the level of glycated hemoglobin.

Implementation option 3

[0171] The present embodiment relates to the use of a wearable device configured to determine the level of glycated hemoglobin when events occur. As an example, it should be noted the breathing of a user of the wearable device inside an office space and outside it, that is, in the fresh air. In this case, a change in external conditions, such as, for example, a greater amount of oxygen, causes a change in rapidly changing parameters of blood components. In response to such an event, the wearable device detects a change in the blood parameters, which facilitates an accurate measurement of the level of glycated hemoglobin.

Cases of application of a wearable device capable of determining the level of glycated hemoglobin.

[0172] User use cases include a solution for monitoring average glucose using HbA1c monitoring using a wearable device.

Business Impact and Benefits of Using

The wearable device disclosed herein provides the following advantages and benefits from its use:

- control of the average glucose level over a certain period (2-3 months) using HbA1c control;

- monitoring the development and phase of microvascular complications associated with diabetes;

- control of nutrition or other habits;

- no action is required from the user;

- possibility of non-professional use;

- no additional equipment required;

- providing new data for the S-Health application.

[0173] The user of the wearable device has the ability to automatically determine the HbA1c level. In this case, if the user wears the device for a long period of time, the wearable device provides automatic determination of the user's activity. For example, as shown in Fig. 14, continuous monitoring of the PPG signals of the wearable device provides the ability to determine the type of activity of the user, such as periods of exercise, sleep, rest, eating, etc. This facilitates the automatic determination of the glycated hemoglobin level, since almost any activity of the user, such as light or intense exercise, eating, resting, sleeping, etc. leads to a change in rapidly changing parameters of blood components. As schematically illustrated in Fig. 14, rapidly changing parameters of blood components are, for example, oxygenation (SpO ₂ ) and blood glucose level, while the HbA1c level changes over a significantly longer period of time, which makes it possible to separate the rapidly and slowly changing parameters of blood components, and therefore, to more accurately determine the glycated hemoglobin level.

[0174] However, it should be noted that if the user is just starting to use a new wearable device, it should be calibrated. When the user first puts on the wearable device, a breath-holding test or a very tight strap of the wearable device should be performed. In this case, during the test, the oxygenated hemoglobin concentration, the total hemoglobin concentration, and the oxygen saturation of the blood change, while the HbA1c level does not change. Therefore, it is possible to accurately measure the glycated hemoglobin level for a new user of the wearable device of the present invention.

Advantageous effects of the invention

[0175] The ability to determine the level of glycated hemoglobin expands the functionality of the S-Health application in the smartwatch, providing additional parameters for a comprehensive analysis of the user's health. The smartwatch can provide daily fluctuations, current values, and a continuous result of assessing the level of glycated hemoglobin over a specified period of time.

As noted earlier, the target groups for which monitoring of glycated hemoglobin levels is necessary are:

- healthy people for the purpose of preventive control;

- children with diabetes;

- patients with diabetes with abnormal renal glucose threshold;

- patients with type I diabetes mellitus, insulin-dependent;

- pregnant women with type II diabetes;

- other people when changing their diet or other habits.

[0176] Continuous monitoring of glycated hemoglobin levels provides an assessment of the average glucose level over a period of time (2-3 months), allows for a direct link between poor control and the development of complications, and for predicting the development and phase of microvascular complications associated with diabetes. These capabilities are provided by the fact that glucose binds to hemoglobin continuously and irreversibly throughout the life of red blood cells (about 120 days), and the HbA1c level is proportional to the average blood glucose level over the past 6-12 weeks.

[0177] The advantages of HbA1c control also include the fact that the disclosed invention does not depend on short-term changes in food (nutrition), physical exercise, hypoglycemic agents, stress. Therefore, preventive glycemic control and feedback in the treatment of diabetes mellitus can be provided.

[0178] As noted earlier, blood glucose levels are characterized by several indicators, in particular HbA1c, which is a long-term average blood glucose level over a period of one to three months, and short-term or instantaneous blood glucose, which is the blood glucose level at a specific point in time, determined either on an empty stomach or after a meal. HbA1c is not affected by short-term changes in blood glucose levels caused by food intake, exercise, etc. on the day of testing, while instantaneous glucose is subject to fluctuations due to food and stress.

[0179] The determination of the HbA1c level using the wearable device disclosed herein is carried out in a manner that is simple for the user. Moreover, as noted earlier, with long-term use of the wearable device, the user's activity is automatically determined and the glycated hemoglobin level is accurately estimated in a non-invasive manner.

[0180] Thus, the technical result of the present invention consists in the possibility of automatic non-invasive assessment of the level of glycated hemoglobin (a long-term indicator of glucose in the blood over the past one to three months), which is not affected by short-term changes in the level of glucose in the blood caused by food intake, physical exercise, etc. on the day of testing.

[0181] The level of glycated hemoglobin has diagnostic value for determining diabetes mellitus. The results of the HbA1c analysis allow us to determine the average blood glucose level over the past 1-3 months. The HbA1c level of less than 5.7% indicates that there is currently no risk of developing diabetes mellitus. With glycated hemoglobin values in the range from 5.7 to 6.5%, there are no signs of diabetes mellitus, but there is a predisposition to diabetes (prediabetes). The risk of overall mortality and mortality from stroke and myocardial infarction increases. A glycated hemoglobin concentration of more than 6.5% indicates the need to exclude or confirm insulin-dependent (type I) or insulin-independent (type II) diabetes mellitus and indicates the need for additional research methods.

Possible causes of physiological increase in blood glucose levels:

- sedentary lifestyle;

- high-carbohydrate diet;

- not drinking enough water;

- inflammatory process in the body;

- taking certain medications (neuroleptics, steroids);

- menstrual period in women.

[0182] In patients with non-insulin-dependent diabetes mellitus, with elevated levels of glycated hemoglobin, mortality from malignant neoplasms increases, especially in the colorectal area. An increase in the level of glycated hemoglobin in the blood always indicates a long period of hyperglycemia and an increased risk of complications in the form of retinopathy, nephropathy, polyneuropathy, micro- and macroangiopathies. A patient suffering from hyperglycemia should strive to achieve the level of glycated hemoglobin of a healthy person - 5.7%. However, this is not always possible. In such cases, the goal of therapy is considered to be a decrease in the concentration of HbA1c to 6.5%. If this goal is achieved, then it can be argued that diabetes is sufficiently well compensated, the likelihood of the consequences of hyperglycemia in the form of all kinds of complications is reduced to a minimum.

[0183] Glycated hemoglobin testing is also performed on women with diabetes mellitus when planning pregnancy. It has been established that high HbA1c levels six months before gestation and during the first trimester are directly related to the likelihood of various complications during pregnancy. Strict control over blood glucose levels reduces the incidence of fetal malformations from 30-40% to 2%.

[0184] Low glycated hemoglobin levels (less than 4%) can be found in a pancreatic tumor that produces insulin (insulinoma). The reason for the decrease is frequent hypoglycemic conditions. Low HbA1c concentrations are found in adrenal insufficiency and some rare hereditary diseases (Hers and Forbes diseases, hereditary fructose intolerance). In addition, possible causes of a physiological decrease in blood glucose levels include:

- excessive physical exertion;

- fasting;

- long-term diet;

- consumption of alcoholic beverages.

The approximate ratio of sugar (glucose) content in the blood and the level of glycated hemoglobin is presented in Table 3.

The following diseases and conditions also affect glycated hemoglobin levels:

- acute (recent) or chronic bleeding may underestimate the actual level of glycated hemoglobin;

- in case of iron deficiency anemia, the results of glycated hemoglobin tests may be overestimated;

- recent blood transfusions, hemolytic anemia may be the cause of low glycated hemoglobin levels.

[0185] Thus, the ability to accurately assess the glycated hemoglobin level using the wearable device of the present invention facilitates monitoring of the user's health status.

[0186] Although the invention has been described with certain illustrative embodiments, it should be understood that the spirit of the invention is not limited to these specific embodiments. On the contrary, it is intended that the spirit of the invention include all alternatives, modifications and equivalents that may be included within the spirit and scope of the claims.

[0187] Furthermore, the invention includes all equivalents of the claimed invention, even if the claims are changed during the consideration process.

Claims (99)

1. A wearable device for determining the level of glycated hemoglobin in the user's blood, comprising: at least one photoplethysmographic (PPG) sensor configured to irradiate the biological tissue of the user with radiation of at least two different wavelengths and detect backscattered PPG signals at said at least two different wavelengths, wherein the wearable device is designed with the ability to, during the user’s daily activities: detection of PPG signals in reflection mode, extraction of features of detected PPG signals; determining an event occurring in the user's daily activities, during which a change in the values of the parameters of blood components occurs, identification of rapidly and slowly changing features characterizing rapidly and slowly changing parameters of blood components according to detected PPG signals, extracting from the PPG signals detected during the event, rapidly and slowly changing features characterizing, respectively, rapidly and slowly changing parameters of blood components, and forming from them a set of rapidly and slowly changing features of PPG signals, and determining the level of glycated hemoglobin in the user's blood based on a comparison of a generated set of features and information from a pre-compiled database using a predictive model. 2. The wearable device according to item 1, further comprising a housing, a processor, a battery and a memory device located in the housing, wherein the storage device is designed with the ability to store a database. 3. The wearable device according to item 1 or 2, further comprising an input and output device configured to input user data and display the determined level of glycated hemoglobin, and a communication module capable of exchanging information with a remote server and/or cloud storage. 4. A wearable device according to any one of paragraphs 1-3, wherein said at least one PPG sensor provides radiation of different wavelengths in the visible (VIS) - near infrared (NIR) spectrum range. 5. A wearable device according to any one of claims 1 to 4, wherein said at least one PPG sensor comprises at least one radiation source comprising two or more light-emitting diodes. 6. A wearable device according to any one of claims 1 to 5, wherein said at least one PPG sensor comprises at least one radiation receiver, preferably a broadband photodetector. 7. A wearable device according to any one of paragraphs 1-6, wherein the database information contains a matrix of features including sets of features characterizing rapidly and slowly changing features of PPG signals, compared with reference values of rapidly and slowly changing parameters of blood components determined by laboratory methods for users, and user data, wherein the said sets of features are generated by the wearable device during events occurring in the daily activities of users and trigger events associated with changes in the values of rapidly changing parameters of blood components. 8. A wearable device according to any one of claims 1-7, wherein the wearable device is configured to identify rapidly and slowly changing features of PPG signals by comparing the features of the signals during the user's daily activities and an event occurring in the user's daily activities. 9. A wearable device according to any one of paragraphs 1-8, wherein the rapidly changing parameters of blood components include total hemoglobin, oxygenated hemoglobin, glucose level, and the slowly changing parameters of blood components include glycated hemoglobin. 10. A wearable device according to any one of claims 1-9, wherein the wearable device is configured to convert said at least two PPG signals from the time domain to the frequency domain and to describe the converted PPG signals with respect to harmonics, Moreover, the set of features characterizing rapidly and slowly changing parameters of blood components includes features of PPG signals determined in the time and frequency domains. 11. The wearable device according to claim 10, wherein the conversion of PPG signals from the time domain to the frequency domain comprises a Fourier transform. 12. A wearable device according to item 10 or 11, wherein the wearable device is configured to, when describing PPG signals in the frequency domain: selecting from all PPG signals the most significant signal, characterized by the highest signal-to-noise ratio or the highest amplitude, and using it as a reference signal, extracting harmonics from the reference signal, determining the coordinates of the fundamental harmonic (f ₀ ) and calculating the coordinates of other harmonics (f ₁ ... f _n ), decomposition of all PPG signals according to the mentioned coordinates, measurements for each harmonic of the amplitude and/or area of the curve to obtain a set of amplitude-phase characteristics of the harmonics, which are features determined in the frequency domain. 13. A wearable device according to any one of paragraphs 10-12, wherein the features determined in the time domain include critical points of the PPG signal waveform, such as the amplitude of the systolic peak, the amplitude of the diastolic peak, the size of the dicrotic notch, and the pulse rate. 14. The wearable device according to claim 13, wherein the features determined in the time domain include the starting points of the sliding time window of PPG signals, such as the length of the window step, the length of the window, the length of the segment. 15. A wearable device according to any one of claims 1-14, wherein the comparison in determining the level of glycated hemoglobin is based on the use of a numerical method or a prediction algorithm, and the prediction model is pre-trained on a compiled database using a machine learning algorithm or a neural network. 16. A wearable device according to any one of paragraphs 1-15, wherein the wearable device is designed to be placed on the wrist. 17. A wearable device according to any one of claims 1-16, wherein the wearable device is a smart device, preferably a smart watch or a fitness bracelet. 18. A method for determining the level of glycated hemoglobin in the blood of a user, comprising the steps of, during the daily activities of the wearable device user: irradiate the biological tissue of the user with radiation of at least two different wavelengths, detecting backscattered PPG signals at said at least two different wavelengths in reflection mode, extract features of detected PPG signals, determine the event that occurs in the user's daily activities, during which the values of the parameters of blood components change, identify rapidly and slowly changing features characterizing rapidly and slowly changing parameters of blood components according to detected PPG signals, extract from the PPG signals detected during the event, rapidly and slowly changing features characterizing the rapidly and slowly changing parameters of the blood components, respectively, and form from them a set of rapidly and slowly changing features of the PPG signals, and determine the level of glycated hemoglobin in the user's blood based on a comparison of a generated set of features and information from a pre-compiled database using a predictive model. 19. The method according to claim 18, wherein the biological tissue of the user is irradiated with radiation of various wavelengths in the visible (VIS) - near infrared (NIR) range of the spectrum. 20. The method according to claim 18 or 19, wherein the biological tissue of the user is irradiated using at least one radiation source containing two or more light-emitting diodes. 21. The method according to any one of claims 18-20, wherein the backscattered PPG signals are detected using at least one radiation detector, preferably a broadband photodetector. 22. The method according to any one of paragraphs 18-21, wherein the database information contains a matrix of features including sets of features characterizing rapidly and slowly changing features of PPG signals, compared with reference values of rapidly and slowly changing parameters of blood components determined by laboratory methods for users, and user data, wherein the sets of features are generated by the wearable device during events occurring in the daily activities of users and trigger events associated with changes in the values of rapidly changing parameters of blood components. 23. The method according to any one of claims 18-22, wherein the fast and slow changing features of the PPG signals are identified by comparing the features of the signals during the user's daily activity and an ongoing event in the user's daily activity. 24. The method according to any of paragraphs 18-23, wherein the rapidly changing parameters of the blood components include total hemoglobin, oxygenated hemoglobin, glucose level, and the slowly changing parameters of the blood components include glycated hemoglobin. 25. The method according to any one of paragraphs 18-24, wherein the method includes the steps of: converting said at least two PPG signals from the time domain to the frequency domain and describe the transformed PPG signals with respect to harmonics in the frequency domain, Moreover, the set of features characterizing rapidly and slowly changing parameters of blood components includes features determined in the time domain and in the frequency domain. 26. The method according to claim 25, wherein the step of converting PPG signals from the time domain to the frequency domain is carried out by means of a Fourier transform. 27. The method according to claim 25 or 26, wherein the step of describing PPG signals in the frequency domain is carried out by: selecting from all PPG signals the most significant signal, characterized by the highest signal-to-noise ratio or the highest amplitude, and using it as a reference signal, extracting harmonics from the reference signal, determining the coordinates of the fundamental harmonic (f ₀ ) and calculating the coordinates of other harmonics (f _l …f _n ), decomposition of all PPG signals according to the mentioned coordinates, measurements for each harmonic of the amplitude and/or area of the curve to obtain a set of amplitude-phase characteristics of the harmonics, which are features determined in the frequency domain. 28. The method according to any one of paragraphs 25-27, wherein the features determined in the time domain include critical points of PPG signals, such as the amplitude of the systolic peak, the amplitude of the diastolic peak, the size of the dicrotic notch, and the pulse rate. 29. The method according to claim 28, wherein the features determined in the time domain include the starting points of the sliding time window of PPG signals, such as the length of the window step, the length of the window, the length of the segment. 30. The method according to any one of paragraphs 18-29, in which the comparison step in determining the level of glycated hemoglobin is carried out using a numerical method or a prediction algorithm, and the prediction model is pre-trained using a compiled database using a machine learning algorithm or a neural network. 31. The method according to any of paragraphs 18-30, wherein the database is designed with the possibility of storage on a storage device, a remote server and/or in cloud storage. 32. A system for determining the level of glycated hemoglobin in the blood of a user, comprising a wearable device and a remote server and/or cloud storage, wherein the system comprises means for exchanging information between the wearable device and the remote server and/or cloud storage, wherein the wearable device comprises at least one photoplethysmographic (PPG) sensor configured to irradiate the biological tissue of the user with radiation of at least two different wavelengths and to detect backscattered PPG signals at said at least two different wavelengths, and is configured to, during the user's daily activities: detection of PPG signals in reflection mode, extraction of features of detected PPG signals; determining an event occurring in the user's daily activities, during which a change in the values of the parameters of blood components occurs, identification of rapidly and slowly changing features characterizing rapidly and slowly changing parameters of blood components according to detected PPG signals, extracting from the PPG signals detected during the event, rapidly and slowly changing features characterizing, respectively, rapidly and slowly changing parameters of blood components, and forming from them a set of rapidly and slowly changing features of PPG signals, and determining the level of glycated hemoglobin in the user's blood based on a comparison of a generated set of features and information from a pre-compiled database using a predictive model. 33. The system according to claim 32, wherein the wearable device comprises a housing, a processor, a battery and a memory device located in the housing, wherein the storage device is designed with the ability to store a database. 34. The system according to claim 32 or 33, wherein the wearable device comprises an input and output device configured to input user data and display a determined level of glycated hemoglobin, and a communication module capable of exchanging information with a remote server and/or cloud storage. 35. The system according to any one of paragraphs 32-34, wherein said at least one PPG sensor provides radiation of different wavelengths in the visible (VIS) - near infrared (NIR) range of the spectrum. 36. The system according to any one of paragraphs 32-35, wherein said at least one PPG sensor comprises at least one radiation source comprising two or more light-emitting diodes. 37. The system according to any one of paragraphs 32-36, wherein said at least one PPG sensor comprises at least one radiation receiver, preferably representing a broadband photodetector. 38. The system according to any one of paragraphs 32-37, wherein the database information contains a matrix of features including sets of features characterizing rapidly and slowly changing features of PPG signals, compared with reference values of rapidly and slowly changing parameters of blood components determined by laboratory methods for users, and user data, wherein the said sets of features are generated by the wearable device during events occurring in the daily activities of users and trigger events associated with changes in the values of rapidly changing parameters of blood components. 39. The system according to any one of claims 32-38, wherein the wearable device is configured to identify rapidly and slowly changing features of PPG signals by comparing the features of the signals during the user's daily activities and an event occurring in the user's daily activities. 40. The system according to any one of paragraphs 32-39, wherein the rapidly changing parameters of the blood components include total hemoglobin, oxygenated hemoglobin, glucose level, and the slowly changing components of the blood include glycated hemoglobin. 41. The system according to any one of paragraphs 32-40, wherein the wearable device is configured to convert said at least two PPG signals from the time domain to the frequency domain and describe the converted PPG signals with respect to harmonics, Moreover, the set of features characterizing rapidly and slowly changing parameters of blood components includes features of PPG signals determined in the time and frequency domains. 42. The system according to claim 41, wherein the conversion of PPG signals from the time domain to the frequency domain comprises a Fourier transform. 43. The system according to item 41 or 42, wherein the wearable device is configured to, when describing PPG signals in the frequency domain: selecting from all PPG signals the most significant signal, characterized by the highest signal-to-noise ratio or the highest amplitude, and using it as a reference signal, extracting harmonics from the reference signal, determining the coordinates of the fundamental harmonic (f ₀ ) and calculating the coordinates of other harmonics (f _l …f _n ), decomposition of all PPG signals according to the mentioned coordinates, measurements for each harmonic of the amplitude and/or area of the curve to obtain a set of amplitude-phase characteristics of the harmonics, which are features determined in the frequency domain. 44. The system according to any one of paragraphs 41-43, wherein the features determined in the time domain include critical points of the PPG signal waveform, such as the amplitude of the systolic peak, the amplitude of the diastolic peak, the size of the dicrotic notch, and the pulse rate. 45. The system according to item 44, wherein the features determined in the time domain include the starting points of the sliding time window of PPG signals, such as the length of the window step, the length of the window, the length of the segment. 46. The system according to any one of paragraphs 32-45, wherein the comparison in determining the level of glycated hemoglobin is based on the use of a numerical method or a prediction algorithm, and the prediction model is pre-trained on a compiled database using a machine learning algorithm or a neural network. 47. The system according to any one of paragraphs 32-46, wherein the wearable device is designed with the possibility of being placed on the wrist. 48. The system according to any one of paragraphs 32-47, wherein the wearable device is a smart device, preferably a smart watch or a fitness bracelet. 49. The system according to any one of paragraphs 32-48, wherein the database is intended for storage on a storage device, a remote server and/or in cloud storage and includes user data and the results of previous determinations of the level of glycated hemoglobin, the system includes means for accessing user data and results of previous determinations of glycated hemoglobin levels from external devices via a remote server and/or cloud storage.