AUTOMATED BLOOD SUGAR LEVEL MONITORING SYSTEM

Description

[0001] This patent application claims priority from French patent application FR19 / 13336, which will be considered as forming an integral part of this description. technical field

[0002] This application relates to the field of automated blood glucose regulation systems, also known as artificial pancreases. Previous technique

[0003] Automated blood glucose regulation systems, also called artificial pancreases, have already been proposed, allowing for the automatic regulation of insulin intake for a diabetic user based on their blood glucose history (or blood glucose level), their meal history, and their insulin injection history.

[0004] Examples of such regulatory systems are described in particular in international patent applications No. WO2018 / 055283 (DD16959 / B15018), No. WO2018 / 055284 (DD17175 / B15267), No. WO2019 / 016452 (DD17609 / B15860) and No. WO2019 / 180341 (DD18479 / B16770), and in French patent applications No. 18 / 56016 of June 29, 2018 (DD18587 / B16893), No. 18 / 00492 of May 22, 2018 (DD18480 / B16894), No. 18 / 00493 of May 22, 2018 (DD18588 / B16895), No. 18 / 73812 of December 21, 2018 (DD18986 / B17521) and No. 19 / 08457 of July 25, 2019 (DD19664 / B18647), previously filed by the applicant.

[0005] Patent applications WO2017 / 035019 A1 and US2018 / 085532 A1 also refer to systems for regulating a patient's blood glucose.

[0006] It would be desirable to be able to improve the performance of known artificial pancreases, and in particular to be able to further limit the risks of placing the user in a situation of hyperglycemia or hypoglycemia.

[0007] We are particularly interested here in the management of meals not declared in advance by the user. Summary of the invention

[0008] The invention is defined by the attached claims. One embodiment provides an automated blood glucose regulation system, comprising: a blood glucose sensor; an insulin injection device; and a processing and control unit, in which the processing and control unit is configured to implement a process for managing undeclared meals, this process comprising the following steps: a) detect, at time t0, an event likely to correspond to a meal not declared by a user; b) when an event is detected in step a), determine, from a first table generated by machine learning from the user's data history, a probability that a meal was eaten by the user within a predetermined period T_ANT preceding time t0; c) if the probability determined in step b) is greater than a threshold TH, determine, from the first table, an estimated time slot for the user to eat the meal within the period T_ANT, and, from a second table generated by machine learning from the user's data history, an estimated size of the meal, then activate a meal management module of the regulation system and transmit the estimated time slot and meal size to said module.

[0009] According to one embodiment: The first table consists of a series of probability values, each corresponding to a percentage of times that a meal was declared by the user in a determined time interval of a determined time cycle, during a learning phase comprising a plurality of occurrences of said time cycle; and the second table consists of a series of meal size values, each corresponding to the average size of meals declared by the user in each time interval of said time cycle during the learning phase.

[0010] According to one embodiment, said time cycle is divided into a plurality of time intervals, the number of values ​​in the first table and the number of values ​​in the second table being equal to the number of time intervals in the time cycle.

[0011] According to one embodiment, the processing and control unit is configured to, if the probability determined in step b) is less than the TH threshold, implement a blood glucose regulation process that is not specific to meals.

[0012] According to one embodiment, the system further comprises a user interface device connected to the processing and control unit.

[0013] According to one embodiment, the processing and control unit is configured to, when an event is detected in step a), implement, before step b), a first step of querying the user, using the user interface device, to ask him or her if he or she has eaten an undeclared meal in the period T_ANT.

[0014] According to one embodiment, the processing and control unit is configured to: If the user answers negatively to the first question, implement a blood glucose regulation procedure not specific to meals; if the user answers positively to the first question, implement a second questioning step of the user, using the user interface device, to ask them for the time and size of said undeclared meal taken during the period T_ANT; and if the user does not answer the first question, implement step b) then step c).

[0015] According to one embodiment, the processing unit is configured to: If the user answers the second query, activate the meal management module of the regulation system and transmit to said module the time and size declared by the user in response to the first query; and if the user does not answer the second query, determine, from the first table, an estimated time slot for the user to take the meal in the period T_ANT, and, from the second table, an estimated size of the meal, then activate the meal management module of the regulation system and transmit to said module the estimated time slot and size of the meal.

[0016] According to one embodiment, the processing and control unit is configured to, in step c), determine, by means of the meal management module, an insulin bolus to be injected into the user based on the estimated meal size.

[0017] According to one embodiment, the processing and control unit is configured to, in step c), weight the bolus by an aggressiveness factor based on the probability determined in step b).

[0018] According to one embodiment, the first and second tables are stored in a memory circuit of the processing and control unit. Brief description of the drawings

[0019] These features and advantages, as well as others, will be described in detail in the following description of particular embodiments, given by way of non-limiting example, in relation to the attached figures, among which: there figure 1 schematically represents, in block form, an example of an automated blood glucose regulation system for a subject according to one embodiment; figure 2is a diagram illustrating an example of an automated blood glucose regulation process that can be implemented by the system of the figure 1 ; there figure 3 represents an example of a first lookup table that can be used to implement the process of the figure 2 ; and the figure 4 represents an example of a second lookup table that can be used to implement the process of the figure 2 . Description of the implementation methods

[0020] The same elements have been designated by the same reference numerals in the different figures. In particular, structural and / or functional elements common to the different embodiments may have the same reference numerals and may have identical structural, dimensional and material properties.

[0021] For the sake of clarity, only the steps and elements necessary for understanding the described embodiments have been shown and are detailed. In particular, the blood glucose monitoring and insulin delivery devices of the described control systems have not been detailed, as the described embodiments are compatible with all or most known blood glucose monitoring and insulin delivery devices. Furthermore, the implementation of the processing and control unit of the described control systems has not been detailed, as the implementation of such a processing and control unit is within the capabilities of a person skilled in the art, based on the functional specifications in this description.

[0022] Unless otherwise specified, the expressions "approximately", "roughly", "about", and "on the order of" mean within 10%, preferably within 5%.

[0023] There figure 1represents schematically, in block form, an example of an implementation of an automated blood glucose regulation system for a user.

[0024] The system of the figure 1It includes a sensor 101 (CG) adapted to measure a quantity representative of the user's blood glucose level, for example, the glucose concentration in the interstitial fluid, which will be referred to hereafter as blood glucose for simplicity. In normal operation, the sensor 101 can be permanently positioned on or in the user's body, for example, on their abdomen or arm. The sensor 101 is, for example, a CGM (Continuous Glucose Monitoring) type sensor, that is, a sensor adapted to continuously measure the user's blood glucose level or at a relatively high frequency (for example, at least once every twenty minutes and preferably at least once every five minutes). The sensor 101 is, for example, a subcutaneous blood glucose sensor.

[0025] The system of the figure 1It also includes an insulin delivery device 103 (PMP), for example, a subcutaneous injection device. Device 103 is, for example, an automatic insulin pump-type injection device, comprising an insulin reservoir connected to an injection needle implanted under the user's skin. The pump can be electrically controlled to automatically inject predetermined doses of insulin at predetermined times. In normal operation, the injection device 103 can be permanently positioned in or on the user's body, for example, on their abdomen.

[0026] The system of the figure 1It further comprises a processing and control unit 105 (CTRL) connected on one side to the blood glucose sensor 101, for example by wired or wireless connection, and on the other side to the injection device 103, for example by wired or wireless connection. In operation, the processing and control unit 105 is adapted to receive the user's blood glucose data measured by the sensor 101 and to electrically control the device 103 to inject the user with predetermined doses of insulin at predetermined times. In this example, the processing and control unit 105 is further adapted to receive, via a user interface 107 (USR), data cho(t) representing the evolution, over time, of the amount of glucose ingested by the user.

[0027] The processing and control unit 105 is designed to determine the insulin doses to be injected into the user, taking into account, in particular, the blood glucose history measured by the sensor 101, the history of insulin injected by the device 103, and the user's glucose intake history. To this end, the processing and control unit 105 includes a digital processing circuit (not detailed), comprising, for example, a microprocessor. The processing and control unit 105 is, for example, a mobile device carried by the user throughout the day and / or night, such as a smartphone configured to implement a regulation process of the type described below.

[0028] The processing and control unit 105 is configured, for example, to implement an automated MPC (Model-based Predictive Control) type regulation process outside of mealtimes, also called a predictive control regulation process, in which the regulation of the administered insulin dose takes into account a prediction of the future evolution of the user's blood glucose over time, made from a mathematical model, for example a physiological model describing the assimilation of insulin by the user's body and its impact on their blood glucose.More specifically, the 105 treatment and control unit can be configured to use the user's injected insulin history and ingested glucose history, and based on a predetermined mathematical model, to determine a curve representing the expected evolution of the user's blood glucose over time, over a future period called the prediction period or prediction horizon, for example, a period of 1 to 10 hours. Taking this curve into account, the 105 treatment and control unit determines the insulin doses that should be injected into the user during the upcoming prediction period so that the user's actual blood glucose (as opposed to the blood glucose estimated from the model) remains within acceptable limits, and in particular to limit the risks of hyperglycemia or hypoglycemia.

[0029] As an alternative, the processing and control unit 105 can be configured to, outside of mealtimes, implement an automated blood glucose regulation process of the decision matrix type, to determine the doses of insulin to be administered to the user according to various observed parameters such as the current blood glucose level measured by the sensor 101, or the rate of change (or slope) of blood glucose over a past period.

[0030] In another variant, the processing and control unit 105 can be configured to, outside of mealtimes, alternate between an MPC-type regulation process and a decision matrix-type regulation process.

[0031] In principle, the user declares each of their meals, and in particular the time of eating the meal and the approximate amount of glucose ingested during the meal (also called meal size), via the user interface 107.

[0032] The processing and control unit 105 is configured to activate a specific meal management module when a meal is declared by the user. This module implements a regulation process adapted to take into account the physiological specificities related to meal assimilation. The meal management module is, for example, implemented in software form by means of the processing and control unit 105. The regulation process implemented by the meal management module may include a calculation step, based on the data entered by the user, and in particular the declared time and size of the meal, to calculate an insulin bolus, that is, an additional dose of insulin to be injected into the user in addition to the normally injected basal insulin rate. The meal management module then commands the injection of the bolus by the injection device 103.The bolus can be administered as a single injection or in several successive injections, for example, two injections. The bolus can be administered at the time the meal is announced, or at the beginning of the meal, or even slightly before the start of the meal if the announcement is made before the meal begins. If the meal is announced late, the bolus can be administered after the meal has started. For example, the insulin bolus injected with a meal can be at least twice the dose of insulin normally injected in one hour between meals. For example, the basal rate of insulin normally injected by the user between meals is between 0.3 and 1.5 IU / h, where IU is an international unit of insulin, which is the biological equivalent of approximately 0.0347 mg of human insulin.The bolus determined by the meal management module and then injected by the injection device 103 is, for example, between 3 and 30 IU depending on the declared meal size and the subject's insulin sensitivity. After the bolus injection, the meal management module can modulate the basal insulin rate injected into the user, for example, using a PID filter or PID corrector (Proportional Integral Derivative), for a predetermined period, for example, three hours following the bolus injection, in order to bring the current blood glucose level back to a target value. At the end of this modulation period, the meal management process ends. The processing and control unit 105 can then implement another blood glucose regulation process, for example, an MPC-type process or a decision matrix-type process as described above.

[0033] One limitation of the functionality described above is that its effectiveness is heavily dependent on user input. If the user fails to report a meal, the meal management module is not activated. The user's blood glucose is then controlled by a non-meal-specific regulation process, such as a medication control panel (MCP) or decision matrix as described above. This can lead to relatively long periods of hyperglycemia, particularly due to the lack of aggressiveness of these processes in responding to meal-related hyperglycemic episodes.

[0034] There figure 2This diagram illustrates an example of a blood glucose regulation process adapted to manage meals not reported by the user. This process is based on the use of statistical data generated by machine learning from a history of data concerning the user.

[0035] The process of figure 2 is based more specifically on the use of a meal probability table or matrix M1, of the type illustrated by the figure 3 , and a table or matrix of average-sized meals M2, of the type illustrated by the figure 4 .

[0036] In the example of the figure 3Table M1 comprises an integer H of rows and an integer D of columns. In the example shown, the number D is equal to 7, with each column of table M1 corresponding to a day of the week. Furthermore, in this example, the number H is equal to 24, with each row of table M1 corresponding to an hour of the day.

[0037] In the example of the figure 4 Table M2 comprises the same number H of rows and the same number D of columns as table M1. Each column of table M2 corresponds to a day of the week, and each row of table M2 corresponds to an hour of the day.

[0038] Denoting by d the column ranks of tables M1 and M2, with d an integer from 0 to 6, and by h the row ranks of tables M1 and M2, with h an integer from 0 to 23, each value M1(d, h) of coordinates (d, h) in table M1 corresponds to the probability that a meal was eaten by the user on day d in the time slot h, and each value M2(d, h) of coordinates (d, h) in table M2 corresponds to the average size of meals usually eaten by the user on day d in the time slot h, for example in gCHO, i.e., grams of carbohydrates. In this example, the days with rank d=0 to d=6 correspond respectively to the seven days of the week, and each time slot with rank h corresponds to a one-hour slot from time h to time h+1.

[0039] Tables M1 and M2 can be generated through machine learning from historical data acquired for the user during a preliminary training phase, for example, a phase lasting from several days to several weeks. For example, each value M1(d, h) in table M1 corresponds to the percentage of times a meal was reported by the user on day d and in time slot h during the training phase. Each value M2(d, h) in table M2 corresponds, for example, to the average size of meals reported by the user on day d and in time slot h during the training phase.

[0040] Tables M1 and M2 can be stored in a memory circuit of the processing and control device 105.

[0041] Tables M1 and M2 can, for example, be updated as the system is used, each time the user declares a meal via user interface 107.

[0042] When custom M1 and M2 tables are not available for a given subject, generic M1 and M2 tables, obtained from historical population data, can be used as a first approach. Generic M1 and M2 tables can, for example, be determined for several population types, such as school children, adolescents, and adults, possibly with different meal patterns, for example, depending on the country of residence.

[0043] The process of figure 2includes a step 201 (DET) for detecting an event that may correspond to a meal not reported by the user. For example, the detection implemented in step 201 could be based on measurements provided by the system's glucose sensor 101. An event detected in step 201 might be, for example, a rise in the user's blood glucose level of the type normally observed after eating, but not explained by a previous report of a meal by the user. For example, the processing and control unit 105 could be configured to implement continuous monitoring of the user's blood glucose curve to detect such events.

[0044] When, at time t0, an event is detected in step 201, a first step 203 (USR1) is implemented, which queries the user using the user interface device 107. During this step, the user is asked, via device 107, whether they have eaten a meal within a predetermined period T_ANT preceding time t0, for example, within the three hours preceding time t0.

[0045] If, at step 203, the user responds, via device 107, that he or she has not taken (N) undeclared meals in the period T_ANT considered, the regulation continues in a step 205 (REGUL) with a non-meal-specific regulation process, for example an MPC type process or a decision matrix type process as described above.

[0046] If, in step 203, the user indicates that they (Y) ate an undeclared meal during the period T_ANT in question, a second step 207 (USR2) is implemented, which involves querying the user via the user interface device 107. During this step, the user is asked, through device 107, the size and time of the undeclared meal. In other words, during steps 203 and 207, the user is asked to retroactively declare the meal they had previously failed to report.

[0047] If, at step 207, the user responds (A), via device 107, by indicating the size and time of the undeclared meal, the processing and control unit 105 activates a specific meal management module, this module implementing, during a step 209 (PMM), a regulation process adapted to take into account the physiological specificities related to the assimilation of a meal, taking into account the size of the meal and the time of the meal declared by the user.

[0048] If, in step 203, the user does not answer (NA) to the question regarding the possible consumption of an undeclared meal during the period T_ANT, a step 211 (PMS1) is implemented, during which the processing and control unit 105 determines, from table M1, whether the probability that an undeclared meal was consumed by the user during the period T_ANT is greater than or less than a predetermined threshold TH. To do this, the processing and control unit 105 determines whether the matrix M1 contains a probability value M1(d, h) greater than the threshold TH in the column at rank d corresponding to the current day and in the rows corresponding to the smallest time range encompassing the period T_ANT.

[0049] If, at step 211, it is determined that table M1 does not include a meal intake probability value M1(d, h) greater than the TH threshold in the range T_ANT, step 205 is implemented, i.e. regulation continues with a non-meal-specific regulation process, for example an MPC type process or a decision matrix type process as described above.

[0050] If, in step 211, it is determined that the table M1 includes a meal-taking probability value M1(d, h) greater than the threshold TH in the range T_ANT, a step 213 (PMS2) is implemented, during which the processing and control unit 105 determines an estimated meal-taking time slot by the user in the period T_ANT, and an estimated size of the meal taken.

[0051] To determine the estimated mealtime slot for the user, the processing and control unit 105 can use table M1. The estimated time slot corresponds, for example, to the slot with coordinates d, h in the period T_ANT for which the mealtime probability value M1(d, h) from table M1 is the highest. Alternatively, the meal can be scheduled for the time when the event was identified in step 201.

[0052] To determine the estimated size of the meal consumed, the processing and control unit 105 can use table M2. The estimated meal size corresponds, for example, to the M2 (d, h) value from table M2 within the time slot with coordinates d, h estimated from table M1. Alternatively, the estimated meal size can be determined from the observed rise in blood glucose.

[0053] Following step 213, step 209 is implemented. This involves the activation of the processing and control unit 105, which activates a specific meal management module. This module implements a regulation process adapted to account for the physiological specificities related to meal assimilation. In this case, the regulation process implemented in step 209 takes as input parameters the estimated size and estimated meal time slot determined in step 213 from tables M1 and M2.

[0054] If, in step 207, the user does not answer (NA) to the question regarding the time and size of the unreported meal eaten during the period T_ANT, step 213 is implemented to estimate the time slot and size of the unreported meal from tables M1 and M2. Step 209 is then implemented in a similar manner to what has just been described.

[0055] Preferably, the insulin bolus to be injected into the user, determined by the regulation process implemented in step 209, is weighted by an aggressiveness factor. This factor can take a first value when step 209 is implemented following a meal declaration made by the user in step 207, and a second value lower than the first value when step 209 is implemented following an estimation of meal size and time from tables M1 and M2 in step 213. In the case where step 211 is implemented, the second value can be lower the lower the probability that a meal was taken by the user in the period T_ANT.

[0056] The process described in relation to the figure 2advantageously allows, by using statistical data representative of the times and sizes of meals usually taken by the user, to treat as meals, by means of a specific regulation module, events likely to correspond to meals but not declared as such by the user.

[0057] As an alternative, steps 203 and 207 of user inquiry can be omitted. In this case, when an event likely to correspond to a meal is detected in step 201, step 211 is implemented directly, followed by step 205 if it is determined in step 211 that the probability that an undeclared meal was taken by the user during the prior period T_ANT preceding the time t0 of event detection is less than the threshold TH, or followed by step 213 and then step 209 otherwise.

[0058] Various embodiments and variations have been described. Those skilled in the art will understand that certain features of these various embodiments and variations could be combined, and other variations will become apparent to them. In particular, the embodiments described are not limited to the examples described in relation to the figures 3 and 4The granularity of the temporal segmentation of the horizontal and vertical axes of tables M1 and M2 can be adjusted. As an alternative, instead of having a meal probability value and an average meal size value per hour and per day over seven consecutive days in tables M1 and M2, tables M1 and M2 could be designed to include a meal probability value and an average meal size value per hour and per working day (regardless of the day in question), and a meal probability value and an average meal size value per hour and per non-working day (regardless of the day in question). More generally, any other segmentation adapted to the user's lifestyle could be considered.

[0059] Furthermore, the embodiments described are not limited to the example described in relation to the figure 2in which the detection implemented in step 201 is based on blood glucose measurements provided by sensor 101. More generally, the detection of events likely to correspond to meals implemented in step 201 may be based on any other appropriate indicator, in addition to or instead of blood glucose measurements.

Claims

1. Automated blood glucose regulation system, comprising: - a blood glucose sensor (101); - an insulin injection device (103); and - a processing and control unit (105), wherein the processing and control unit is configured to implement a method of management of undeclared meals, this method comprising the steps of: a) detecting (201), at a time t0, based on measurements provided by the blood glucose sensor (101) or another indicator, an event likely to correspond to a meal non-declared by a user; b) when an event is detected at step a), reading (211), in a first table (M1) generated by training from a history of the user's data, a probability for a meal to have been taken by the user within a period T_ANT of predetermined duration preceding time t0; c) if the probability read at step b) is greater than a threshold TH, reading (213), in the first table (M1), an estimated time slot of meal ingestion by the user within period T_ANT and, reading, in a second table (M2) generated by training from a history of the user's data, an estimated size of the meal, and then activating (209) a meal management module of the regulation system and transmitting to said module the estimated time slot and size of the meal, wherein: - the first table (M1) is formed of a series of probability values, each corresponding to a percentage of times that a meal has been declared by the user within a determined time interval of a determined time cycle, during a training phase comprising a plurality of occurrences of said time cycle; and - the second table (M2) is formed of a series of meal size values, each corresponding to the average meal size declared by the user within each time interval of said time cycle during the training phase.

2. System according to claim 1, wherein the first time cycle is divided into a plurality of time intervals, the number of values of the first table (M1) and the number of values of the second table (M2) being equal to the number of time intervals of said time cycle.

3. System according to claim1 1 or 2, wherein the processing and control unit (105) is configured to, if the probability determined at step b) is lower than threshold TH, implement (205) a blood glucose regulation method non-specific to meals.

4. System according to claim 3, further comprising a user interface device (107) coupled to the processing and control unit (105).

5. System according to claim 4, wherein the processing and control unit (105) is configured to, when an event is detected at step a), implement, before step b), a first step (203) of interrogation of the user, by means of the user interface device (107), to ask them whether they have had an undeclared meal within period T_ANT.

6. System according to claim 5, wherein the processing and control unit (105) is configured to: - if the user answers negatively to the first interrogation, implement (205) a blood glucose regulation method non-specific to meals; - if the user answers positively to the first interrogation, implement a second step (207) of interrogation of the user, by means of the user interface device (107), to ask them the time and the size of said undeclared meal taken during period T_ANT; and - if the user does not answer to the first interrogation, implement step b) and then step c).

7. System according to claim 6, wherein the processing and control unit (105) is configured to: - if the user answers to the second interrogation, activate (209) the meal management module of the regulation system and transmit to said module the time and the size declared by the user as an answer to the first interrogation; and - if the user does not answer to the second interrogation, determine (213), based on the first table (M1), an estimated time slot of meal ingestion by the user within period T_ANT and, based on the second table (M2), an estimated size of the meal, and then activate (209) the meal management module of the regulation system and transmit to said module the estimated time slot and size of the meal.

8. System according to any of claims 1 to 7, wherein the processing and control unit (105) is configured to, at step c), determine, by means of the meal management module, an insulin bolus to be injected to the user according to the estimated meal size.

9. System according to claim 8, wherein the processing and control unit (105) is configured to, at step c), weigh said bolus by a coefficient which is a function of the probability determined at step b).

10. System according to any of claims 1 to 9, wherein the first (M1) and second (M2) tables are stored in a memory circuit of the processing and control unit (105).