DYNAMIC DETERMINATION DEVICE FOR BASAL INSULIN INJECTION DOSE

Abstract

TITLE OF THE INVENTION: DEVICE FOR DYNAMIC DETERMINATION OF BASAL INSULIN DOSE TO BE INJECTED The device for dynamically determining the dose of basal insulin to be injected comprises: - a means (21) for capturing a blood glucose level at bedtime one day and a blood glucose level upon waking the following morning, - a means (26) for calculating the difference between blood glucose levels, - a means (22, 24, 25, 42, 43) for determining whether a predetermined event has occurred between bedtime and waking, - a means (26) for determining, if no predetermined event has occurred, hyperglycemia or hypoglycemia based on the difference between blood glucose levels, - a means (26) for determining whether at least a predetermined number of hyperglycemic or hypoglycemic episodes have occurred within a predetermined period of time, and - a means (26) for modifying the basal insulin to be injected, in this case of occurrence. Figure for the abbreviation: Figure 1.

Description

Title of the invention: DEVICE FOR DYNAMIC DETERMINATION OF BASAL INSULIN DOSE TO BE INJECTED Technical field of the invention

[0001] The present invention relates to a device for dynamic dose determination. of basal insulin to be injected. It applies, in particular, to the field of medical devices used by the patient. State of the art

[0002] Type 1 diabetes ("TID") affects approximately 350,000 people in France, and nearly six million people in Europe. For the past 20 years, the number of people with type 1 diabetes ("PDT1") has increased by 3 to 4% per year. The causes of this increase are currently unexplained, but since the factors involved are environmental, genetic, and nutritional, it is expected that this increase will continue and increase.

[0003] The body of type 1 diabetic patients does not produce insulin (essential for transforming carbohydrates from meals into energy and therefore regulating blood glucose levels). These patients must therefore administer insulin several times a day. This is manifested by:

[0004] Permanent monitoring of blood sugar (glucose level in the blood) by a device for measuring blood sugar, called a glucometer, or dextro, or blood sugar sensor.

[0005] At least four insulin injections per day are required using a device for administering insulin, called an insulin pen (or pump). These injections concern basal insulin, also called slow insulin (background insulin, to be administered once a day) and rapid insulin (to be administered before each meal, called bolus). Correction operations (reinjection of a low dose in the event of hyperglycemia) are also part of these injections.

[0006] The aim of these different steps is to maintain the patients' blood sugar levels within a range of 70 to 180 mg / dL. When blood sugar levels are below this range, this indicates hypoglycemia, which can cause, in advanced cases, loss of consciousness and coma. When blood sugar levels are above this range, this indicates hyperglycemia, which can cause, in the long term, an imbalance of diabetes.

[0007] The imbalance of diabetes causes harmful events for the organs, and serious long-term complications such as atherosclerosis, myocardial infarction, stroke or even arteritis of the lower limbs, but also retinopathies and nephropathies. In fact, diabetes multiplies by eight the risk amputation and nine times the risk of dialysis for kidney failure. It is the second leading cause of cardiovascular accidents and the leading cause of blindness in adults.

[0008] A large proportion of people with type 1 diabetes have poorly balanced treatment. For example, the SAGE study presented in 2019 at the EASD annual congress, conducted in 17 countries in Western Europe, Eastern Europe, Latin America, Asia and the Middle East, shows that more than 40% of patients had an HblAc level above 8%, reflecting poorly controlled diabetes (Glycated hemoglobin, which reflects average blood sugar levels over the last three months). In addition, 43% of patients do not adjust their insulin doses themselves because they are dependent on the doctor or medical team.

[0009] Functional insulin therapy can help control diabetes but is not very accessible. It has been shown that adjusting the dose can reduce HbAlc levels by 1%. This reduction can, in many cases, help maintain "controlled" diabetes. Furthermore, studies have also shown that this can reduce long-term complications by 30%.

[0010] Functional insulin therapy consists of calculating the optimal insulin dose to administer based on the food ingested and other physiological parameters, instead of taking a fixed dose. Today, this adjustment is cumbersome to implement. Patients wishing to use this method must undergo specific training, acquire a great deal of knowledge, and then, at each meal, perform complex calculations, which are experienced as risky and time-consuming. These calculations are then applied to each injection. Only 20% of patients can commit to this approach, and some of them perform the calculations by integrating personal shortcuts and simplifications, taking the risk of reduced effectiveness and therefore unbalanced diabetes.

[0011] Decision support systems for people with type 1 diabetes currently on the market only offer a classic bolus calculator, requiring prior knowledge of the calculation parameters and mastery of carbohydrate counting and functional insulin therapy by the user, reducing the number of people who can use this method. These systems also do not offer dynamic adaptation of the slow insulin to be injected, and the few devices dealing with slow insulin are only designed for people with type 2 diabetes who do not perform multiple injections. These tools are sometimes present in insulin pumps, capillary blood glucose meters, or smartphones. The few randomized clinical studies that have evaluated these tools show an improvement in HbAlc.However, their limited functionality does not allow them to truly support functional insulin therapy. Presentation of the invention

[0012] The present invention aims to remedy all or part of these drawbacks.

[0013] To this end, according to a first aspect, the present invention aims at a device for dynamically determining the dose of basal insulin to be injected into a user, which comprises: - a means of capturing the user's blood sugar level at bedtime on the first day, - a means of capturing the user's blood sugar level upon waking the day after the first day, - a means of calculating the difference in blood sugar levels between this bedtime and this waking up, - a means of determining whether a predetermined event has occurred between this bedtime and this waking up, - a means of determining, if no predetermined event has occurred, hyperglycemia or hypoglycemia based on the difference in blood sugar levels between going to bed and getting up, - a means of determining whether at least a predetermined number, greater than or equal to two, of hyperglycemias or hypoglycemias have occurred within a predetermined period of time and - a means of modifying the basal insulin to be injected if at least the predetermined number of hyperglycemias or hypoglycemias have occurred within a predetermined period of time.

[0014] Slow-acting insulin, or basal insulin, helps meet the body's minimum insulin needs in the absence of carbohydrate intake. Basal insulin that is too high leads to increased hypoglycemia during the day and night. And insulin that is too low leads to increased blood sugar levels.

[0015] Thanks to the implementation of the present invention, basal insulin intake is better adapted and precise for the patient who injects several doses of insulin per day, than methods using only fasting blood sugar taken by capillary method.

[0016] When the predetermined number is seven, each week, if no predetermined event has occurred between the user going to bed and getting up, it is recommended to increase, decrease or keep the dose of basal insulin to be injected. If an event has occurred and the predetermined time period is greater than one week, blood glucose differences over that time period are used.

[0017] In embodiments, the means for determining whether a predetermined event has occurred between this setting and this rising is configured to determine the occurrence of at least one of: - eating a meal between going to bed and getting up, - performing physical activity between going to bed and getting up and - taking insulin between going to bed and getting up.

[0018] The inventor has determined that these events can modify blood sugar levels and that their occurrence must inhibit the taking into account of the difference in blood sugar levels for the night in question.

[0019] In embodiments, the means for determining whether a predetermined event has occurred between this setting and this rising is configured to determine the occurrence of at least one of: - eating a meal between going to bed and getting up, - performing physical activity between going to bed and getting up and - taking insulin between going to bed and getting up.

[0020] In embodiments, the device further comprises: - a means of capturing the user's pre-prandial blood glucose level before a meal, - a means of capturing the user's postprandial blood glucose level after a meal, - a means of receiving a description of a meal, - a means of determining IOB insulin at the time of injection, - a means of determining a new carbohydrate ratio for a meal time interval between the times of capture of preprandial blood glucose and postprandial blood glucose, depending on the following factors: - preprandial blood sugar, - postprandial blood sugar, - a quantity of carbohydrates absorbed during the meal, - a bolus, - an active insulin level, - a value of a correction factor, - a higher glycemic target and - a lower glycemic target; and - a means of replacing a carbohydrate ratio used by the device to determine the dose of insulin to be injected, representative of a number of carbohydrates absorbed by the user, if at least a predetermined number, greater than or equal to two, of new carbohydrate ratios have been determined in a predetermined period of time, the replacement carbohydrate ratio being an average, possibly weighted, of the new carbohydrate ratios.

[0021] In embodiments, the carbohydrate ratio replacement means is configured to inhibit carbohydrate ratio replacement if the carbohydrate ratio replacement is within a predetermined range of values ​​surrounding the carbohydrate ratio to be replaced.

[0022] In embodiments, the device further comprises means for daily determining a new correction factor corresponding to a blood glucose drop factor for a unit of insulin received, the correction factor being a function of the following factors: - a pre-correction blood sugar level, - post-correction blood sugar, - a bolus, - active insulin, - a higher glycemic target, and - a lower glycemic target; and - a means of replacing a correction factor if at least a predetermined number, greater than or equal to two, of new correction factors have been determined within a predetermined period of time, the replacement correction factor being an average, possibly weighted, of the new correction factors.

[0023] In embodiments, the means for replacing a correction factor is configured to inhibit replacement of the correction factor if the replacement correction factor is within a predetermined range of values ​​surrounding the correction factor to be replaced.

[0024] In embodiments, the device further comprises: - a means of receiving a description of a meal, - a means of calculating the nutrients corresponding to the meal described, number of calories, carbohydrates, lipids and proteins, - a means of determining correlation between the presence of the foods described and a tendency to hypoglycemia or hyperglycemia after meals, and - a means of triggering an alert to the user, in the event of detection of a sufficiently marked correlation and - a means of determining a meal bolus dose, based on carbohydrate ratio (ICR) and correction factor (CF) values.

[0025] In embodiments, the device further comprises: - a means of determining the occurrence of an event among the presence of a disease of the user and the presence of menstruation of the user, - a means of determining a correction bolus, or a meal bolus based on the user's last measured blood glucose level, a correction factor (CF), an upper blood glucose target and a blood glucose target lower, providing different values ​​of correction bolus or meal bolus or slow insulin, depending on whether a said event occurs or not.

[0026] In embodiments, the device further comprises a means for determining whether the correction bolus is greater than a minimum injectable insulin dose and: - if yes, transmission to the user of the value of the correction bolus to be injected, and - if no, transmission to the user of a value representative of a physical activity allowing the same effect to be obtained as the correction insulin dose.

[0027] According to a second aspect, the present invention relates to a method for dynamically determining the dose of basal insulin to be injected into a user, which comprises: - a step of capturing a blood glucose level of the user at bedtime on a first day, - a step of capturing a blood glucose level of the user at waking the day after the first day, - a step of calculating the difference in blood glucose levels between this bedtime and this waking, - a step of determining whether a predetermined event has occurred between this bedtime and this waking, - a step of determining, if no predetermined event has occurred, hyperglycemia or hypoglycemia as a function of the difference in blood glucose levels between bedtime and waking, - a step of determining whether at least a predetermined number, greater than or equal to two, of hyperglycemias or hypoglycemias have occurred in a predetermined period of time and - a step of modifying the basal insulin to be injected if at least the predetermined number of hyperglycemias or hypoglycemias have occurred within a predetermined period of time.

[0028] The advantages, aims and particular characteristics of this method being similar to those of the device which is the subject of the invention, they are not recalled here. Brief description of the figures

[0029] Other advantages, aims and particular characteristics of the invention will emerge from the following non-limiting description of at least one particular embodiment of the device for dynamically determining the dose of insulin to be injected, which is the subject of the present invention, with reference to the appended drawings, in which:

[0030] [Fig.l] schematically represents a particular embodiment of a device which is the subject of the invention.

[0031] [Fig.2] represents, in the form of a flowchart, steps in calculating the carbohydrate ratio, or “ICR”,

[0032] [Fig.3] represents, in the form of a flowchart, steps for calculating a correction factor “CF”,

[0033] [Fig.4] represents, in the form of a flowchart, the steps for calculating basal insulin,

[0034] [Fig.5] represents, in the form of a flowchart, the steps for calculating a meal bolus,

[0035] [Fig.6] represents, in the form of a flowchart, steps in calculating a bolus correction,

[0036] [Fig.7] represents, in the form of a flowchart, steps for determining basal insulin and

[0037] [Fig.8] represents the course of a day and the blood sugar measurements implemented. Description of the embodiments

[0038] The present description is given without limitation, each characteristic of an embodiment being able to be combined with any other characteristic of any other embodiment in an advantageous manner.

[0039] The functional components of an embodiment of the device which is the subject of the invention are first described below, with reference to [Fig. 1].

[0040] This device 20 comprises a user interface 25, preferably on a communicating terminal 24 of the user, for example a computer, a tablet or a mobile telephone and, more preferably, an intelliphone (in English “smartphone”).

[0041] This user interface 25 preferably comprises a keyboard 42, possibly in the form of a touch screen and a voice recognition module 43, implementing a microphone of the user terminal 24.

[0042] The user terminal 24 implements means of communication with a blood glucose reader 21, for example by implementing the Bluetooth communication protocol (registered trademark), with a connected watch 22, for example equipped with an actimeter and / or with a connected insulin pen, for example via a server 23 of the supplier of this connected watch or of this connected insulin pen.

[0043] Terminal 24 verifies that the inputs and outputs are correct.

[0044] A calculator 26, for example in a web server, includes a module 27 for calculating IOB (acronym for "Insulin on board" or insulin on board, also called "active insulin"). IOB refers to the amount of insulin from the bolus that was recently administered and that acts still in the user's body. The IOB also reflects the amount of insulin present in the user's body but not yet consumed.

[0045] The calculator 26 also includes a correction bolus module 28, a meal bolus calculation module 29 and a meal nutrient calculation module 30. The calculator 26 calculates the nutrients consumed and provides bolus or bolus modification recommendations.

[0046] A database 36 comprises a local part 37, in the user terminal 24, which stores local data and a remote part 38, preferably hosted by a web server, which stores a complete replication of the data of the local part 37. The local part 37 of database 36 stores in particular data on the food consumed by the user, the history of consumption, in particular during meals, the exercises practiced by the user, the quantities of insulin delivered, the physical conditions of the user and the values ​​of the operating parameters of the device 20.

[0047] Usage algorithms 39, in communication with the database 36 and with the user terminal 24, include algorithms 40 for recognizing the user's voice, in natural language, and algorithms 41 for analyzing the user's eating patterns. The amounts of nutrients are calculated based on the nutrient compositions of the foods consumed and the amounts consumed.

[0048] Adjustment algorithms 31, in communication with the database 36, include an algorithm 32 for detecting time events (“TimePoints”), which analyzes the entries received by the terminal 24 with respect to time and associates attributes (in English “labels”) with them, for example “fasting”, “before meal” or “after meal”). The adjustment algorithms 31 also include an algorithm 33 for calculating the carbohydrate ratio, or ICR, an algorithm 34 for calculating basal insulin (slow insulin) and an algorithm 35 for calculating correction factors.

[0049] The objective of this new device is to support patients in the implementation of functional insulin therapy. This consists of supporting the patient in different aspects that current tools do not provide for:

[0050] A / The importance of adjusting parameter values. A large number of patients do not know how to adjust their parameters themselves, or wait for a consultation to review their functional insulin therapy parameters. These parameters are specific to each patient, and allow the patient's insulin therapy to be brought as close as possible to their functional needs. These parameters change over time and with seasonal variations, and are strongly influenced by certain events such as illness or the menstrual cycle. This evolution implies the importance of readjusting them and not waiting for a doctor's consultation for optimal glycemic control. These parameters are: - Basal insulin. This is the minimum insulin needed to survive, and finding the right daily dose is essential for controlled insulin therapy. - The carbohydrate ratio is the amount of insulin needed for each gram of carbohydrate ingested. This value varies throughout the day and is essential for accurately calculating rapid-acting insulin doses at mealtimes. - The correction factor corresponds to how much each unit of insulin lowers blood sugar. This value is used to calculate the doses of rapid insulin corrections, taken when the person is hyperglycemic, or in addition to meals.

[0051] B / Correctly determine the quantity of nutrients in your meal. Studies show that an error of + / -10% on the quantity of carbohydrates increases the risks of hypoglycemia and hyperglycemia after meals. A study showed that helping, through an application, to calculate carbohydrates and then the meal insulin dose reduced the number of meal boluses that were poorly evaluated and followed by corrective action. The tools available to patients require that the patient make this assessment by themselves and enter a quantity of carbohydrates, without providing the means to support the patient in this task and make it easy and safe.Furthermore, the impact of lipids and proteins is not supported in these decision support tools.

[0052] C / Determine the correct calculation parameters and consequently the correct doses of insulin to inject during phases of illness and during the menstrual cycle.

[0053] Operating algorithms for the device which is the subject of the invention are described below, represented in the form of flowcharts in Figures 2 to 7. The various methods described with reference to the figures take place in parallel.

[0054] The algorithm for calculating ICR ([Fig.2]), CF ([Fig.3]) and basal ([Fig.4]) only depends on the timepoint detection algorithm. Regarding this algorithm, the following definitions are noted.

[0055] For type 1 blood sugar: Pre / Post-Prandial: Blood glucose measurement before and after a prandial injection, Pre / Post-Correction: Blood glucose measurement before and after a correction injection, Pre / Post-Manual: Blood glucose measurement before and after a manual injection.

[0056] For “In” injections separated by a “Tn” duration (Tn is the duration separating the L injection and the In+i injection.) For each injection of the day: If Tn4 is between two o'clock and five o'clock, Post-I„ । = Pre In, If Tn.i is greater than five hours, Post-In i = min (value between 2h and 5h of In.i) If Tn.i is less than two hours, Post-In i = Post-In, If In4 and In are two prandial injections, they are considered as an entry and if Tn is less than two hours, the injections are rejected, If Tn is a manual injection and Tn_i and Tn are less than two hours, the injections are rejected, If In concerns physical activity, the In injection is rejected.

[0057] For type 2 blood sugar: Low Glycemia: Blood glucose measurement below 70 mg / dL, Last Glycemia: Blood glucose measurement recovered at the end of the day, and Fasting Glycemia: Blood glucose measurement recovered after a period of fasting (usually in the morning just after the patient gets up).

[0058] There is only one measurement of fasting glycemia and last glycemia per day.

[0059] The maximum time between a fasting glycemia measurement and a last glycemia measurement is 12 hours. The minimum time between a fasting glycemia measurement and a last glycemia measurement is four hours.

[0060] To break a fast: By default, the fasting period ends at noon at the latest. But the occurrence of one of these events: - low blood sugar requiring re-sugar (lowglycemia), - the start of a physical activity, or - taking an injection (with or without a meal) defines that the end of the fasting period (and therefore that the fasting glycemia) is the one that comes just before the start of this event.

[0061] The physical activity is either entered by the patient, via the user interface, or detected by the device 20, via the connected watch 22.

[0062] Injection:

[0063] LowGlycemia (resweetening), Last glycemia = last post-injection of the day, Fasting = first pre-glycemia of the day, Or low Glycemia during the night.

[0064] Each of the ICR ([Fig.2]), CF ([Fig.3]) and basal ([Fig.4]) calculation algorithms needs different timepoints, but it is more efficient, economical and accurate to make a timepoint algorithm that goes through the whole previous day and identifies the right attributes ("labels") based on the events. However, in the formula, to calculate the new ICR carbohydrate ratio, the last valid value of the CF correction factor is used.

[0065] The method generally comprises: - a step of capturing the user's blood sugar level at bedtime on the first day, - a step of capturing the user's blood sugar level upon waking the day after the first day, - a step of calculating the difference in blood sugar levels between this bedtime and this getting up, - a step of determining whether a predetermined event occurred between this bedtime and this getting up and - a step of determining whether a predetermined event occurred between this setting and this rising among: - eating a meal between going to bed and getting up, - performing physical activity between going to bed and getting up, - taking insulin between going to bed and getting up, and, - possibly, an episode of hypoglycemia.

[0066] Method 50, illustrated in [Fig.2], concerns the calculation of the carbohydrate ratio, or ICR.

[0067] During a step 51, the entries of the previous day are analyzed by putting in implements a time event detection algorithm (“TimePoint”). Step 51 consists of associating attributes with events, among “before meal”, “after meal”, “fasting” or “at bedtime”.

[0068] When capturing data, for example from capillary blood glucose meters and continuous blood glucose meters, and from connected pens, this data is processed to associate it with the meal data provided by the user. The same applies to the insulin values ​​calculated and proposed to the user. For example, the following steps are implemented:

[0069] Data is received from connected medical devices.

[0070] The information is recorded in the database according to the nature of the data received.

[0071] The data is analyzed using different algorithms to label certain entries based on other events entered, for better use of blood glucose and insulin levels. With: - detection of preprandial and postprandial blood glucose levels, - detection of fasting blood glucose levels and bedtime blood glucose levels, and - association of insulin doses from the connected pen with the suggested insulin doses.

[0072] This processing avoids false duplicates. For example, in the situation where the patient performs a calculation and is late in giving his injection, the above procedure makes it possible to avoid duplicate entries and to associate the calculated dose with the actual time of the injection.

[0073] During a step 52, a particular time slot is chosen, for example a meal time slot, between breakfast, lunch, snack and dinner.

[0074] During a step 53, it is determined whether all the necessary values ​​are available and valid. These are the following values: - a preprandial blood glucose value, - a postprandial blood glucose value, - a quantity of carbohydrates, - the description of a meal, - the active insulin IOB at the time of bolus injection and - secondarily, because the value may be zero or non-existent, a possible indication of the presence of activity, illness or menstruation.

[0075] Otherwise, the processing of the time interval is rejected, during a step 56. If yes, during a step 54, it is determined whether there is a disease or a menstrual cycle, by questioning the user. If yes, the processing of the time interval is rejected, during step 56. If no, during a step 55, a new carbohydrate ratio ICR, newICR, is calculated, depending on the following factors: - preprandial blood glucose, - postprandial blood glucose, - quantity of carbohydrates, - bolus, - active insulin, - CF value, - upper blood glucose target and - lower blood glucose target.

[0076] During a step 57, it is determined whether this new carbohydrate ratio is within an acceptable range of values, because despite the previous steps, there may sometimes be erroneous inputs (manual entry error, etc.) which may give an aberrant result.

[0077] If no, the processing of the time interval is rejected, during step 56. If yes, during a step 58, it is determined whether there has been a sufficient number of new carbohydrate ratios over a predetermined period. For example, 7 new carbohydrate ratios over the last 30 days. If no, during a step 61, the new time-stamped carbohydrate ratio is stored and the following day is waited for (processed as described in step 51). If yes, during a step 59, a new average carbohydrate ratio is calculated, equal to the average of the carbohydrate ratios over the predetermined period. This is a weighted average taking into account the average of these 7 new carbohydrate ratios, and also the old carbohydrate ratio used.

[0078] Then, during a step 60, if the new average carbohydrate ratio is in an interval of values ​​around the old carbohydrate ratio (for example, between 90% and 110% of the old ICR value, oldlCR), the carbohydrate ratio to be applied for the user is not changed and, if the new carbohydrate ratio is outside this interval, the old carbohydrate ratio is replaced by the new average carbohydrate ratio. It is therefore the new ICR carbohydrate ratio which is then applied to the user.

[0079] In [Fig.3], we observe a method 70 for calculating a correction factor CF, carried out once a day. The initial steps, 52 to 54 and 56 have already been described with regard to [Fig.2],

[0080] During a step 74, a new correction factor CF is calculated for the user's insulin sensitivity. The carbohydrate ratio ICR corresponds to how much carbohydrate was consumed by the user. The correction factor CF corresponds to the blood glucose drop factor for one unit of insulin received. The new correction factor, newCF, is a function of the following factors: - pre-correction blood glucose, - post-correction blood glucose, - bolus, - active insulin, - upper blood glucose target and - lower blood glucose target.

[0081] During a step 75, it is determined whether the new value of the correction factor CF is within an acceptable range of values ​​(for example between 0 and 2 times oldCF).

[0082] If no, the method stops. If yes, during a step 76, it is determined whether there have been a sufficient number of new correction factors CF over a predetermined period. For example, 7 new correction factors over the last 30 days. If no, during a step 77, the new time-stamped correction factor CF is stored and the following day is waited for (processed as described in step 51). If yes, during a step 78, a new average correction factor is calculated, equal to the average of the correction factors CF over the predetermined period.Then, during a step 79, if the new average correction factor is in an interval of values ​​around the old correction factor (for example between 90% and 110% of the old value), the correction factor CF to be applied for the user is not changed and, if the new correction factor is outside this interval, the old correction factor is replaced by the new average correction factor.

[0083] In [Fig.4], we observe a method 90 for calculating basal insulin. Steps 52 to 54 and 56 have already been described with regard to [Fig.2].

[0084] During a step 96, an attribute vector is associated with the last day spent. The consideration of the previous day is a function of the fasting blood sugar of the current day and at bedtime of the previous day and, possibly, of the events occurring between bedtime and waking up, for example nighttime meal, nighttime hypoglycemia, taking insulin, carrying out physical activity during the night.

[0085] This analysis makes it possible to give: - a “normal” attribute, if the analysis shows that the nighttime fasting period went well, indicating that the basal insulin dose was correct, - a “hyper” attribute, if the analysis shows that the basal insulin dose was low, and caused an increase in blood sugar levels greater than 20 mgl / dl (for example) and - a “hypo” attribute, if the analysis shows that the basal insulin dose was high, and caused a drop in blood sugar levels greater than 20 mg / dl.

[0086] During a step 97, it is determined whether a counter is equal to a predetermined value, for example the predetermined value implemented during steps 58 and 76. If not, the value of the basal insulin is not changed and the next day is waited for. If yes, during a step 98, it is determined whether a decision has been made (and stored as an operating parameter value of the device) to carry out an automatic adaptation. If no, the value of the basal insulin is not changed and the next day is waited for. If yes, during a step 99, it is determined whether there is a tendency towards hypoglycemia, output “o”, or towards hyperglycemia, output “y”. In the event of a tendency towards hyperglycemia, during a step 100, the basal insulin is increased by one unit and the counter whose value is compared to the predetermined value during step 97 is reset to zero. This counter is incremented by one if the result of step 97 is “No”.When this counter has reached the value of seven, as long as the basal insulin value is not changed (exit "yes" from step 98), it remains at this value of seven.

[0087] Then, during a step 101, the attribute vector is checked for the following two days: If a single “hypo” attribute occurs over the last two days, which represents a drop in blood sugar between bedtime and waking up greater than 20 mg / dl, then there is a risk of hypoglycemia. Therefore, in step 102, the insulin is reduced by one unit and the counter is reset. If not, the process is completed for the current day. If yes, during a step 102, the basal insulin is reduced by one unit and the counter is reset to zero, the value of which is compared to the predetermined value during step 97, and the process is completed for the current day. If the output of step 99 is “0”, step 102 is carried out and the process is completed for the current day.

[0088] In [Fig. 5], we observe a method 110 for calculating a meal bolus. Depending on the user's use of the user interface, either step 113 (manual entry) or steps 111 and 112 (entry by voice recognition) are carried out. In both cases, the user answers a questionnaire to describe the foods in his meal. In the case of voice dictation (step 111), natural language voice recognition is implemented during step 112.

[0089] During a step 114, following step 112 or step 113, the nutrients corresponding to the meal are calculated, i.e. the number of calories, carbohydrates, lipids and proteins.

[0090] During a step 115, recommendations relating to the meal are given via the user interface, by implementing a food pattern analysis algorithm. For example, each entry concerns a piece of data: - preprandial blood sugar, - postprandial blood sugar, - bolus, - food 1 and quantity 1, - food 2 and quantity 2 ...

[0091] For each food to be analyzed: we search for all occurrences of food 1 in the data present in the history.

[0092] If a correlation exists between the presence of these foods and a tendency to hypoglycemia or hyperglycemia after meals, then this warning will be triggered.

[0093] The correlation can be determined via a data processing algorithm, such as linear regression, logistic regression, k nearest neighbor, support vector machine, random forest.

[0094] For example, a recommendation warns a user against a food planned in the meal, which the algorithm has found tends to cause hypoglycemia in that user.

[0095] During a step 116, a measurement of the user's blood glucose level over a certain period of time, for example fifteen minutes preceding the start of the meal, is obtained from at least one connected device. During a step 117, it is determined whether there is a disease or a menstrual cycle present by questioning the user. If not, during a step 118, generic calculation parameters are obtained. The parameters are the ICR and CF and the upper and lower glycemic targets. The ICR carbohydrate ratio and the generic CF correction factor are those adjusted by the algorithms described above.

[0096] Those that are specific are determined using an adaptation of the three previous algorithms, but using a few different parameters (on the number of days, three instead of seven for example, the way of calculating the weighted averages), and taking into account only the values ​​concerned by the same state.

[0097] Thus, in the Cloud and local database, we have a generic ICR carbohydrate ratio, a carbohydrate ratio for a disease "ICRmaladie", a carbohydrate ratio for periods "ICRrègle", .... The same goes for the correction factor CF, and for the basal.). If so, during a step 119, we obtain calculation parameters specific to the disease or the menstrual cycle.

[0098] Following one of steps 118 or 119, during a step 120, the bolus dose of the meal is calculated, by implementing the calculation parameters obtained, as a function of the values ​​of the carbohydrate ratio ICR and correction factor CF parameters.

[0099] In [Fig.6], a bolus correction method 130 is observed. During a step 131, the last blood glucose level measured in the user's blood is obtained. During a step 132, it is determined whether there is a disease or a menstrual cycle present, by questioning the user. If not, during a step 133, generic calculation parameters are obtained, as explained above. If the result of step 132 is positive, during a step 134, calculation parameters specific to the disease or the menstrual cycle are obtained as explained above.

[0100] Following one of steps 133 or 134, during a step 135, a correction insulin dose is calculated. For example, a formula of the type Correction bolus is used, which is a function of the blood glucose, the correction factor, CF, the upper blood glucose target and the lower blood glucose target, by implementing the calculation parameters obtained. During a step 136, it is determined whether the calculated correction insulin dose is greater than or equal to a minimum injectable insulin dose. If yes, during a step 137, the user is recommended, via the user interface, to increase the injected insulin dose. If not, during a step 138, the user is recommended, via the user interface, to engage in physical activity that can achieve the same effect as the correction insulin dose.

[0101] In [Fig.7], we observe a method 140 for determining basal insulin, i.e. slow insulin. During a step 141, it is determined whether the current time is close to an insulin injection time, for example less than 15 or 30 minutes from such a time stored in memory. Otherwise, the method ends. If yes, during a step 142, it is determined whether an injection is carried out, via a connected pen or a declaration by the user implementing the user interface. If yes, during a step 147, the time of this injection, the dose injected, the food ingested, the blood glucose level and the values ​​of the operating parameters of the device are stored and the method ends. If not, during a step 143, it is determined if there is a disease or menstrual cycle present, by questioning the user. If no, during a step 145, the user is reminded of the basal insulin dose to be injected and returns to step 141. If yes, during a step 144, the user is reminded of the event-specific dose of basal insulin to be injected and returns to step 141. The specific insulin dose is the adjusted value for the algorithm from [Fig.4].

[0102] In [Fig.8], a schematic curve 150 of the blood glucose level of a patient during a typical day is shown, with the times indicated on the abscissa. Also shown in [Fig.8] are the levels of low blood glucose (less than 70 mg / dl) 151, high blood glucose (greater than 140 mg / dl) 152 and very high blood glucose (greater than 180 mg / dl) 153.

[0103] At time 159 of the previous day, the last blood glucose level of the day is measured and stored. The device's memory then contains the time-stamped data of the previous day and the carbohydrate ratio values ​​ICRn, ICRmoy, CFn, CFmoy and IB for each time interval.

[0104] At time 160 in the early morning, a preprandial measurement is taken, which corresponds to fasting glycemia, a measurement of blood sugar recovered after a period of fasting of at least four hours.

[0105] At a time 161 preceding the taking of the first meal of the day, the user makes a description of the first meal envisaged. The nutrients of this first meal are then determined, then meal recommendations are calculated, as explained above, and insulin intake recommendations, as explained above for the bolus calculation. These recommendations are provided to the user via the user interface. The user then takes insulin and takes the first meal.

[0106] At a predetermined time of day (not shown in [Fig.8]), the device performs the association of attributes (“labels”) with the instants of the previous day, and the calculation of the new values ​​of the morning carbohydrate ratio, ICRmatin, of the afternoon, ICRaprès-midi, of the evening, ICRsoirée and of the night, ICRnuit, the correction factor CF and the basal insulin.

[0107] At time 162 in the morning, the user takes a postprandial and pre-correction blood glucose measurement. The device then calculates a possible correction bolus, as explained above, and transmits it to the user via the user interface.

[0108] At a time 163 preceding the intake of the second meal of the day, the user takes a blood sugar measurement and a description of the planned second meal. The nutrients of this second meal are then determined, then meal recommendations are calculated, as explained above and intake recommendations insulin, as discussed above for bolus calculation. These recommendations are provided to the user through the user interface. The user then takes the insulin and the second meal.

[0109] At time 164, the patient has low blood sugar. The application reacts by recommending a sugar boost, and a check of their blood sugar 15 or 30 minutes later. The user can optionally set their preferred sugar boosts (their favorite sweet foods), for greater precision, and the application tells them how many such sugar boosts to take.

[0110] At a time 165, the user takes a postprandial blood glucose measurement. The device calculates, if necessary, a correction bolus as explained above and transmits a recommendation to the user, via the user interface. However, in [Fig.8], time 165 also corresponds to low blood glucose. Therefore, it is not a correction bolus that is recommended by the device but a sugar boost, as explained above.

[0111] At a time 166 preceding the intake of the third meal of the day, the user performs a blood glucose measurement and a description of the third meal envisaged. The nutrients of this third meal are then determined, then meal recommendations are calculated, as described above, and insulin intake recommendations are calculated, as described above for the bolus calculation. These recommendations are provided to the user via the user interface. The user then takes the insulin and the third meal.

[0112] At time 167, for example just before going to bed, the user takes a postprandial blood glucose measurement. The device calculates, if necessary, a correction bolus as described above and transmits a recommendation to the user, via the user interface.

Claims

Claims

1. Device (20) for dynamically determining the dose of slow insulin to be injected into a user treated by multi-injection, characterized in that it comprises: - a means (21) for automatically capturing a multitude of blood glucose levels of the user during a first day and the day after the first day, - a means (32) for automatically assigning each blood glucose level captured according to the time of the first day and the day after the first day, - a means (32) for determining a blood glucose level assigned at bedtime on the first day, - a means (32) for determining a blood glucose level assigned at waking up on the day after the first day, - a means (26) for calculating the difference in blood glucose levels between the blood glucose level at bedtime and the blood glucose level at waking up, - a means (22, 24, 25, 42, 43) for determining whether a predetermined event has occurred between bedtime and waking up, - a means (26, 90, 54, 99) for determination,if no predetermined event has occurred, of hyperglycemia or hypoglycemia depending on the difference in blood sugar levels between going to bed and getting up, - means (26, 90, 97, 99) for determining whether at least a predetermined number, greater than or equal to two, of hyperglycemias or hypoglycemias have occurred in a predetermined period of time, - means (26, 90, 100, 101, 102) for modifying the slow insulin to be injected if at least the predetermined number of hyperglycemias or hypoglycemias have occurred in the predetermined period of time and - means (25) for indicating to the user a numerical value of the dose of modified slow insulin to be injected.,

2. Device (20) according to claim 1, in which the means for determining whether a predetermined event has occurred between this going to bed and this getting up is configured to determine the occurrence of at least one event among: - the taking of a meal between going to bed and getting up, - performing physical activity between going to bed and getting up and - taking insulin between going to bed and getting up.

3. A device (20) according to one of claims 1 or 2, which further comprises: - means for capturing a preprandial blood glucose level of the user before a meal, - means for capturing a postprandial blood glucose level of the user after a meal, - means (24, 25, 42, 43) for receiving a description of a meal, - means for determining the active insulin at the time of injection, - means (26, 50) for determining a new carbohydrate ratio (ICR) for a meal time interval between the times of capturing the preprandial blood glucose level and the postprandial blood glucose level, as a function of the following factors: - a preprandial blood glucose level, - a postprandial blood glucose level, - an amount of carbohydrates absorbed during the meal, - a bolus, - an active insulin level, - a value of a correction factor (CF), - an upper glycemic target and - a lower glycemic target;and - means (26, 50) for replacing a carbohydrate ratio used by the device to determine the dose of insulin to be injected, representative of a number of carbohydrates absorbed by the user, if at least a predetermined number, greater than or equal to two, of new carbohydrate ratios have been determined in a predetermined period of time, the replacement carbohydrate ratio being an average, possibly weighted, of the new carbohydrate ratios.;

4. The device (20) of claim 3, wherein the carbohydrate ratio replacement means (26, 50) is configured to inhibit carbohydrate ratio replacement if the replacement carbohydrate ratio is within a predetermined range of values ​​surrounding the carbohydrate ratio to be replaced.

5. Device (20) according to one of claims 1 to 4, which further comprises means (26, 70) for daily determination of a new correction factor (CF) corresponding to a blood glucose drop factor for a unit of insulin received, the correction factor being a function of the following factors: - a pre-correction blood glucose, - a post-correction blood glucose, - a bolus, - an active insulin, - a higher glycemic target, and - a lower glycemic target; and - means (26, 70) for replacing a correction factor if at least a predetermined number, greater than or equal to two, of new correction factors have been determined within a predetermined period of time, the replacement correction factor being an average, possibly weighted, of the new correction factors.

6. A device (20) according to claim 5, wherein the means (26, 70) for replacing a correction factor is configured to inhibit replacement of the correction factor if the replacement correction factor is within a predetermined range of values ​​surrounding the correction factor to be replaced.

7. Device (20) according to one of claims 1, 2 or 4 to 6, which further comprises: - means (24, 25, 42, 43) for receiving a description of a meal, - means (26) for calculating the nutrients corresponding to the described meal, number of calories, carbohydrates, lipids and proteins, - means (26) for determining a correlation between the presence of the described foods and a tendency to hypoglycemia or hyperglycemia after meals, and - means (24, 25, 26, 43) for triggering an alert to the user, in the event of detection of a sufficiently marked correlation and - means (26) for determining a bolus dose of the meal, as a function of carbohydrate ratio (ICR) and correction factor (CF) values.

8. Device (20) according to claim 3, which further comprises: - means (26) for calculating the nutrients corresponding to the meal described, number of calories, carbohydrates, lipids and proteins, - a means (26) for determining correlation between the presence of the foods described and a tendency to hypoglycemia or hyperglycemia after meals, and - a means (24, 25, 26, 43) for triggering an alert to the user, in the event of detection of a sufficiently marked correlation and - a means (26) for determining a meal bolus dose, based on carbohydrate ratio (ICR) and correction factor (CF) values.

9. Device (20) according to one of claims 1 to 8, which further comprises: - means (24, 25, 42, 43) for determining the occurrence, indicated by the user, of an event among the presence of a disease of the user and the presence of menstruation of the user, - means (26, 130) for determining a correction bolus, or a meal bolus according to the last blood glucose measured in the blood of the user, a correction factor (CF), an upper glycemic target and a lower glycemic target, providing different values ​​of correction bolus or meal bolus or slow insulin, depending on whether a said event is indicated or not.

10. Device (20) according to claim 9, which comprises means (26, 130, 136, 137, 138) for determining whether the correction bolus is greater than a minimum injectable insulin dose and: - if so, for transmitting to the user the value of the correction bolus to be injected, and - if not, for transmitting to the user a value representative of a physical activity making it possible to obtain the same effect as the correction insulin dose.

Images