Portable Device and Method for Non-Invasive Blood Glucose Level Estimation

Abstract

The invention relates to a device (1) comprising a measuring unit (2) with a measuring module (4) for measuring the glucose level, a first computer module (5) for processing data from a first part of the process for measuring the glucose level, a first communications module (6), a first data storage module (7) and a pushbutton (8). The device also comprises a personal monitoring unit (3) with second and third communication modules (17, 20), a second computer module (18) for processing data from a second part of the process for measuring the glucose level, an interface module (19) and a second data storage module (22). Also described is a method for non-invasive blood glucose level estimation.

Claims

1. A device (1) for non-invasive blood glucose level estimation, comprising: a measuring unit (2), comprising the following modules: a measuring module (4) for measuring the glucose level, comprising a plurality of elements configured to perform a process for non-invasive blood glucose level measurement; a first computer module (5), configured to control the measuring module (4), and to process data on a first part of the process for measuring blood glucose level, starting from the data provided by the measuring module (4); a first communications module (6), configured to receive configuration commands and send the data associated with such commands to the first computer module (5); a first data storage module (7), configured to store the information from the measuring unit (2); a pushbutton (8), configured for activating the measuring unit (2); a personal monitoring unit (3), comprising: a second communications module (17) intended to establish bidirectional wireless communications with at least the measuring unit (2); a second computer module (18), configured to process data on a second part of the process for measuring the blood glucose level; an interface module (19), configured to display the information from the measuring unit (2) and the data provided by the second computer module (18), and to enable the user to interact; a third communications module (20), configured to establish bidirectional wireless communications with an external service provider (21); a second data storage module (22), configured to store data from the personal monitoring unit (3); and characterised in that the measuring module (4) comprises: a first light emitter E1 (9), activatable from the first computer module (5), and configured to emit with a wavelength corresponding to a maximum absorbance in the absorption spectrum of the glucose molecule within the near-infrared range, which strikes on the skin of a body area (10) irrigated by a vascular bed; a second E2 light emitter (11), activatable from the first computer module (5), and configured to emit with a wavelength corresponding to a minimum absorbance in the absorption spectrum of the glucose molecule, and arranged close to the light emitter E1 (9); a photodetector (12), sensitive to the wavelength of the first and second light E1 and E2 (9, 11), configured to generate an electrical current signal (S1), the amplitude of which depends on the intensity of light received in the sensitivity spectrum of the photodetector (12); a first amplification step (13), which generates an electrical voltage signal (S2), amplified from the electrical current signal (S1), when said electrical current signal (S1) is weak; a first filtering step (14) which abstracts the components of the electrical voltage signal (S2) which vary due to the arterial flow, generating a third signal (S3); a second amplification step (15), which generates an amplified signal (S4) from the third signal (S3); a second filtering step (16), which abstracts the components of the electrical voltage signal (S2) related to stationary properties in the measurement, as well as possible motion artefacts and other sources of low-frequency error, generating a fifth signal (S5).

2. The device according to claim 1, characterised in that the first and second light emitters E1 and E2 (9, 11) are arranged such that the light beams cross through a relatively translucent body area (10), such as a finger or an ear lobe, and are captured by the photodetector (12) located on the opposite side of said body area (10).

3. The device according to claim 1, characterised in that it is covered by a casing opaque to the light spectrum wherein the photodetector (12) is sensitive.

4. The device according to claim 3, characterised in that the opaque casing is configured to exert a constant pressure on the body area (10).

5. The device according to claim 1, characterised in that the measuring unit (2) and the personal monitoring unit (3) comprise a real-time timing system configured to manage instants of measurement and periods of time of the operations.

6. The device according to claim 1, characterised in that the measuring unit (2) comprises a temperature module (24) configured to measure the temperature in the body area (10) wherein the measurement is performed.

7. The device according to claim 1, characterised in that the personal monitoring unit (3) is configured to additionally measure physiological variables selected from: respiratory rate, heart rate, ECG, heart rate variability, body temperature, physical activity, falls, body composition, skin impedance and pulse oximetry.

8. The device according to claim 1, characterised in that the measuring unit (2) and the personal monitoring unit (3) are physically separated or integrated in a monolithic device (23).

9. A method for non-invasive blood glucose level estimation using the device described in any one of claims 1 to 8, performed in a distributed manner by the first computer module (5) and the second computer module (18), and comprising the following operations: performing a first estimation (28) of a first parameter (D1) as the average value of the fifth signal (S5) during a pre-configured time period P1 (25) wherein the light emitters E1 (9) and E2 (11) are deactivated; performing a second estimation (29) of a second parameter (D2) as the average value of the fifth signal (S5) during a second pre-configured time period P2 (26) wherein the emitter E1 (9) is activated, and the emitter E2 (11) is deactivated; performing a third estimation (30) of a third parameter (D3) during that second pre-configured time period P2 (26), wherein the third parameter corresponds to the average value of the differences between successive maxima and minima identified in the pulsating signal (S4); performing a fourth estimation (31) of a fourth parameter (D4) as the average value of the fifth signal (S5) during a third pre-configured time period P3 (27) wherein the emitter E2 (11) is activated, and the emitter E1 (9) is deactivated; performing a fifth estimation (32) of a fifth parameter (D5) during the third pre-configured time period P3 (27), wherein such fifth parameter (D5) corresponds to the average value of the differences between successive maxima and minima identified in the pulsating signal (S4); estimating the blood glucose level (33) starting from a model which depends on the parameters from the first to the fifth (D1, D2, D3, D4, D5) wherein the model isolates the influence of glucose by weighting the dependence with respect to these parameters (D1, D2, D3, D4, D5) according to two conditions: with the glucose molecules subjected to a light associated with a maximum absorbance in the second parameter (D2) and third parameter (D3) or subjected to a light associated with a minimum absorbance in the fourth parameter (D4) and the fifth parameter (D5) and wherein the influence of the ambient light in the measurement of the photodetector (12) is weighted in the dependence with respect to the first parameter (D1), and wherein the influence of the signal components related to stationary properties in the measurement, with motion artefacts and error sources generated by low-frequency signals, is weighted in the dependence with respect to the second and fourth parameters (D2, D4) and the model isolates the influence of the arterial blood in the estimation, and eliminates the influence of other tissues, weighting the dependence with respect to the third and fifth parameters (D3, D5).

10. The method of claim 8 wherein the dependence of the model for glucose level estimation with respect to the first to fifth parameters (D1, D2, D3, D4, D5) is performed based on coefficients which can be configured remotely by means of sending commands, and wherein the values of the coefficients generate a generalised model for use in different users, or a customised model for an individual use, or a generalised and customisable model including the dependency with other parameters related to the particular characteristics of the user.

11. The method of claim 8 incorporating the measurement of the temperature module (24) as a parameter of the model for glucose level estimation.

12. The method of claim 8 incorporating an operation which activates an alarm locally and remotely when the glucose estimation records a value considered unsuitable.

Description

DESCRIPTION OF THE DRAWINGS

[0044] As a complement to the description provided herein, and for the purpose of helping to make the features of the invention more readily understandable, in accordance with a preferred practical exemplary embodiment thereof, such description is accompanied by a set of drawings which, by way of illustration and not limitation, represent the following:

[0045] FIG. 1 shows a diagram of the basic architecture of the device object of the patent and the devices which make it up.

[0046] FIG. 2 shows a diagram of the basic architecture of the measuring unit.

[0047] FIG. 3 shows a diagram of the basic architecture of the measuring module.

[0048] FIG. 4 shows a diagram of the basic architecture of the personal monitoring unit.

[0049] FIG. 5 shows a diagram of the monolithic device which combines the measuring unit and the personal monitoring unit.

[0050] FIG. 6 illustrates the method for non-invasive blood glucose level estimation.

PREFERRED EMBODIMENT OF THE INVENTION

[0051] In a possible embodiment of a first aspect of the invention proposed here shown in FIG. 1 it has a device (1) for non-invasive blood glucose level estimation, which in a preferred embodiment comprises a device formed by two units: a measuring unit (2) and a personal monitoring unit (3). The device (1) is capable of communicating wirelessly and bidirectionally with an external service provider (21).

[0052] The measuring unit (2) is a portable device which is placed on the skin of a human area body irrigated by a vascular bed, and which emits light at two different wavelengths, one of them corresponding to a maximum absorbance in the absorption spectrum in the glucose molecule within the near-infrared range. The measuring unit (2) captures the light which crosses through the measuring area, and in conjunction with the personal monitoring unit (3), performs a blood glucose level estimation by means of a computational model based on the following conditions: 1) isolating the influence of the glucose from the relationship existing in the amount of light received at each of the wavelengths; 2) normalising the estimation with respect to the influence of the ambient light and with respect to stationary properties of the measurement such as the level of light emitted, the properties of the tissues, the arrangement and features of the light emitters and the photodetector, or the influence of the measuring area, as well as motion artefacts and other sources of low-frequency noise; 3) isolating the influence of the arterial blood considering the pulsating component of the received signals. In the preferred embodiment, the measuring unit (2) comprises the following modules, referred to in FIG. 2:

a) a measuring module (4), which incorporates the components for the non-invasive measurement of glucose level;
b) a first computer module (5), responsible for activating some components of the measuring module (4) and a first part of the processing associated with the glucose level estimation starting from the data provided by the measuring module (4);
c) a first communications module (6), which is responsible for receiving configuration commands and sending data associated with the first computer module (5);
d) a first data storage module (7), for the temporary storage of the information in the event of communication failure, or for the persistent recording of the information from the measuring unit (2);
e) a pushbutton (8), for activating the measuring unit (2);

[0053] In turn, the measuring module (4) comprises the following components, referred to in FIG. 3:

a) A first light emitter E1 (9), activatable from the first computer module (5), with a wavelength corresponding to a maximum absorbance in the absorption spectrum of the glucose molecule within the near-infrared range, which strikes on the skin of a human area body (10) irrigated by a vascular bed. In one embodiment of the invention the wavelength corresponding to 950 nm is used, although other wavelengths are possible.
b) A second light emitter E2 (11), also activatable from the first computer module (5) and with a wavelength corresponding to a minimum absorbance in the absorption spectrum of the glucose molecule, located in a close manner to the first emitter E1 (9), and which affects the same area of the skin (10). In one embodiment of the invention the wavelength corresponding to 660 nm is used, although other wavelengths are possible.
c) A photodetector (12) sensitive to the wavelength of the first and second emitters (9, 11), which generates an electrical current signal S1 the amplitude of which depends on the intensity of light received in the sensitivity spectrum of the photodetector (12). In a preferred embodiment, the sensitivity spectrum of the photodetector integrates the wavelengths corresponding to 660 nm and 950 nm.
d) When the signal S1 is very weak, a first amplification step (13) generates the electrical voltage signal S2 amplified from signal S1.
e) A first filtering step (14) which abstracts the components of signal S2 which vary as a consequence of the arterial blood flow in the vascular bed, generating signal S3. In a preferred embodiment, this step is performed by means of a high-pass filter with a cut-off frequency which enables the pulsating components related to cardiac activity to pass.
f) When the signal S3 is very weak, a second amplification step (15) which generates the amplified signal S4 starting from the signal S3.
g) A second filtering step (16) which abstracts the components of signal S2 related to stationary properties in the measurement (light level emitted, stationary properties of the tissues, arrangement and features of the light emitters and the photodetector (12), or the influence of the measuring area (10)), which may vary from one measurement to another, as well as possible motion artefacts and other low-frequency error sources, generating signal S5. In a preferred embodiment, this step is performed by means of a low-pass filter with a cut-off frequency which does not enable the pulsating components related to cardiac activity to pass.

[0054] The information generated by the measuring unit (2) is transmitted wirelessly to the personal monitoring device (3), with which it maintains a bidirectional communications link. The start time of the measurement can be activated locally by means of a pushbutton (8) on the measuring unit (2) or it can be activated remotely by means of sending a command from the personal monitoring unit (3). Also by means of another command, the time instants wherein the automatic glucose estimations would be performed could be previously configured.

[0055] In the personal monitoring unit (3), with greater capabilities, both in terms of hardware and software, than the measuring unit (2), the part of processing with the greatest computational load associated with the method for glucose level estimation is developed. The multilevel distribution of the processing favours energy saving and reduces the computational load. The personal monitoring unit (3) can also be responsible for the processing and the management of the information coming from other portable sensors connected to it, which can be related to other physiological variables (respiratory rhythm, heart rate, ECG, heart rate variability, body temperature, physical activity, falls, body composition, skin impedance and pulse oximetry, etc.). In the preferred embodiment, the personal monitoring unit (3) comprises the following modules, referring to FIG. 4:

a) A second communications module (17) intended to establish bidirectional wireless communications with at least the measuring unit (2).
b) A second computer module (18) responsible for the second part of the processing associated with glucose level estimation. Algorithms for the detection of alarm situations or situations which should be considered worthy of attention are also executed in it.
c) An interface module (19) for displaying the information from the measuring unit (2) and the results from the second computer module (18), and enabling the user to interact in an adapted manner: touch (19.a), visual (19.b), auditory (19.c), or voice control (19.d), etc. If an alarm event is detected, the interface (19) includes adapted warning means (light, acoustic, vibrations, etc.). The user could then deactivate or silence the alarm while he manages and reviews the information provided. The interface (19) can be used by two types of users: the monitored user, which could occur in a home environment, or the professional user, which could occur in a clinical environment.
d) A third communications module (20) intended to establish bidirectional wireless communications with an external service provider (21).
e) A second data storage module (22) which is responsible for the temporary storage of the information from the personal monitoring unit (3) in case of communication failure, or for the persistent recording of such information, which enables the future access thereof without needing a remote connection to an external database.

[0056] In a preferred embodiment of the invention, the personal monitoring unit (3) is portable, although in other possible embodiments it can also be a fixed installation. Such device can be implemented physically by means of a smartphone or a tablet.

[0057] The measuring unit (2) and the personal monitoring unit (3) maintain a real-time timing system in order to manage the instants of measurement and the time periods of the operations. This timing system is also responsible for assigning to each estimation the instant in time in which they are performed. The personal monitoring unit (3) is responsible for coordinating the realization of the glucose estimations according to a pre-established plan, which can be configured by an expert user locally through the interface (19) of the device or remotely through telematic services of the e-Health system. Such estimations will be activated in the measuring unit (2) by means of sending a command. A hierarchical procedure is established from the personal monitoring unit (3) to the measuring unit (2) based on the sending of commands for the synchronisation of the timing systems. The different users, both experts and monitored users, can also activate the instantaneous performance of an estimation. This instantaneous activation can be performed from the pushbutton (8) of the measuring unit (2) or from the interface (19) of the personal monitoring unit (3).

[0058] The personal monitoring unit (3) can manage the information in an autonomous manner, including alarm management, establishing communications in a seamless manner to the user with the measuring unit (2) and with an external service provider (21) in order to integrate information and the alarms in an e-Health system.

[0059] The structural and functional modularity of the device for the non-invasive blood glucose level estimation enables two possible configurations: a distributed one (1), wherein the measuring unit (2) is physically separated from the personal monitoring unit (3), and another monolithic one, shown in FIG. 5, wherein the measuring unit (2) is integrated together with the personal monitoring unit (3) in a single device (23). In this second case, the communications between both units can be performed directly or wired (not wireless). Furthermore, the measuring unit (2) and the personal monitoring unit (3) can share physical components in the monolithic configuration (device (23)), such as a single computer module.

[0060] In a preferred embodiment of the invention, the first and second light emitters E1 and E2 (9, 11) are arranged such that the light beams cross through a relatively translucent body area (10) (a finger, for example), and are captured by a photodetector (12) located on the opposite side of the body area. This first embodiment is focused on the incorporation of the measuring unit (2) in a casing opaque to the spectrum of light wherein the photodetector (12) is sensitive, which is configured to maintain a constant pressure on the measuring area (10).

[0061] In another embodiment, and as FIG. 1 also shows, the measuring unit (2) incorporates a temperature module (24), which is responsible for measuring the temperature of the measuring area (10), such that the glucose estimation model incorporates this data in order to adjust the coefficients as a function of the temperature.

[0062] In addition to the components and elements making up the device object of the patent (1), it is also characterised in the method used for the non-invasive blood glucose level estimation, which is performed in a distributed manner in two levels: a first level of processing in the measuring unit (2), and a second level of processing in the personal monitoring unit (3). Thus, a distributed processing architecture and methodology are established, which is advantageous in terms of computing and energy saving. In terms of computing, because such multilevel structure enables the processing load between the two devices to be compensated for in order to prevent computational overload. In terms of the energy, because the highest energy consumption in portable devices is related to sending data wirelessly. As multilevel processing reduces and abstracts the wireless information to be transmitted, energy saving is thus favoured.

[0063] Said method comprises the following operations, referring to FIG. 6:

a) During a pre-configured time period P1 (25) wherein the first and second light emitters E1 and E2 (9, 11) are deactivated, the estimation (28) is performed of the parameter D1 as the average value of the signal S5.
b) During a second pre-configured time period P2 (26) wherein the first emitter E1 (9) is activated, and the second emitter E2 (11) is deactivated, the estimation (29) is performed of the parameter D2 as the average value of the signal S5.
c) During that same time period P2 (26), the estimation (30) is performed of the parameter D3 as the average value of the differences between successive maxima and minima identified in the pulsating signal S4 related to the cardiac activity.
d) During a third pre-configured time period P3 (27) wherein the second transmitter E2 (11) is activated, and the first transmitter E1 (9) is deactivated, the estimation (31) is performed of the parameter D4 as the average value of the signal S5.

[0064] e) During that same time period P3 (27), the estimation (32) is performed of the parameter D5 as the average value of the differences between successive maxima and minima identified in the pulsating signal S4 related to the cardiac activity.

f) Estimation (33) of the blood glucose level starting from a model which depends on the parameters D1, D2, D3, D4 and D5. The model isolates the influence of the glucose by weighting the dependence with respect to the parameters according to two conditions: with the glucose molecules subjected to a light associated with a maximum absorbance in the parameters D2 and D3, or subjected to a light associated with a minimum absorbance in the parameters D4 and D5. The influence of the ambient light on the measurement of the photodetector (12) is weighted in the dependence with respect to the parameter D1. The influence of the signal components related to stationary properties in the measurement (light level emitted, stationary properties of the tissues, arrangement and features of the light emitters and the photodetector (12), or the influence of the measuring area (10)), as well as possible motion artefacts and other error sources generated by low-frequency signals, is weighted in the dependence with respect to the parameters D2 and D4. The model isolates the influence of the arterial blood on the estimation, and eliminates the influence of other tissues, weighting the dependence with respect to the parameters D3 and D5.

[0065] The dependence of the model for glucose level estimation with respect to the parameters D1, D2, D3, D4 and D5 is based on coefficients which can be remotely configured by means of sending commands. The value of the coefficients is fixed by means of a quantitative method (least squares methods, genetic algorithms, swarm intelligence or neural networks), which minimises the mean square error of the estimations in a reference study, which is used as a calibration method. There are three possible models for glucose level estimation as a function of the coefficients: 1) a generalised model, wherein the value of the coefficients is adapted for the use of the model in multiple users; 2) a customised model, wherein the value of the coefficients is adjusted in order to optimise the glucose estimations for a given user; 3) a generalised and customisable model, which includes the dependence with other parameters related to the particular characteristics of the user, such as age, sex, the type of diabetes or the context of the measuring.

[0066] It is also possible to select the method of representing the glucose level estimation in the user interface (19): text, graphic, auditory, etc. or a multiple selection thereof. Furthermore, this proposal adds the possibility of selecting the classification method of the user, based on the results of the estimation. The selected classification method will establish thresholds based on the blood glucose level, which will enable the user to be classified into different levels, for example: very high, high, normal, low or very low. The thresholds, levels and the result of the classification will be displayed in a manner related to the representation method selected for the estimation (text, graphic, auditory, etc. or a multiple selection thereof). The classification method assumes prior clinical knowledge and classification standards in order to provide direct information about the state of the user and thus facilitate their evaluation and diagnosis.

[0067] The possibility of performing a historical tracking of the glucose estimations in the different measurements of a user is further considered. Such historical record will be displayed in a manner related to the selected representation method (text, graphic, auditory, etc. or a multiple selection thereof). In each of the measurements, the date and time when the estimation was performed can be identified.

[0068] The object of the invention may comprise additional processing on the record of the measurements which has the object of automatically establishing trends, patterns and predictions in the history of the measurements, which may be notified to the user.

[0069] The second computer module (18) also implements a system for detecting undesirable situations, which, if detected, would generate a series of local and remote alarms which would enable preventive action on the user. Such system uses a library of locally or remotely configurable indicators and a table with critical values for the generation of alarms related to said indicators. These indicators can be associated with a specific glucose estimation, but also with an analysis of trends, patterns and predictions of the history of the estimations. The logic and the decision rules which govern the activation of the alarms can also be configured to relate one or more of the indicators.