Closed-loop blood glucose control systems and processes

Abstract

A control device (30) for determining a recommended value for a control parameter of an insulin infusion device (20). Figure for the abbreviation: Fig. 1

Description

Title of the invention: Closed-loop blood glucose regulation systems and methods. FIELD OF THE INVENTION

[0001] The present invention relates to the field of closed-loop blood glucose control systems for the regulated administration of insulin to a patient. Such systems are also known as artificial pancreas.

[0002] BACKGROUND OF THE INVENTION

[0003] An artificial pancreas is a system that automatically regulates the insulin supply of a diabetic patient, or a user, based on their blood glucose history, meal history and insulin history.

[0004] In particular, the present invention relates to the insulin sensitivity factor (ISF), which can be used to determine the appropriate insulin dose based on the patient's body's sensitivity to insulin, thus enabling more precise adjustment of insulin dosage to maintain optimal glycemic regulation.

[0005] It would be desirable to improve the performance of systems based on the insulin sensitivity factor by improving the accuracy and reliability of physiological models, for example. More accurate predictions of insulin sensitivity would allow for a better estimation of insulin requirements, thereby reducing the risks of hyperglycemia and hypoglycemia. This improvement in prediction accuracy is essential for optimizing insulin therapy and ensuring better overall management of diabetes.

[0006] US20090054753A1 describes insulin sensitivity as a general measure of The body's response to insulin dosage. This factor can change depending on the patient's physiological state and can be useful in determining the patient's response to therapy. A patient's insulin sensitivity can be determined in various ways, such as through input from a healthcare provider, interference from other conditions, or by determining it from information about previous insulin dosages and blood glucose measurements. Insulin sensitivity can be used as a parameter to determine the timing of the next blood glucose test.

[0007] The present invention aims to improve this situation.

[0008] The invention thus aims to address at least partially the technical problems presented above.

[0009] BRIEF SUMMARY OF THE INVENTION

[0010] The invention thus relates to a control device for determining a recommended value for a control parameter of an insulin infusion device, the control device comprising: • a recovery unit, the recovery unit being configured to recover user data, each piece of user data having a timestamp and the user data being associated with a unique user, the user data comprising at least: • a quantity of insulin infused to the single user; • the amount of carbohydrates ingested by the single user; • a plurality of physiological values ​​of the single user, the plurality of physiological values ​​of the single user including at least measured blood glucose levels; • a recommendation unit, the recommendation unit being configured to determine the recommendation value at least based on an insulin sensitivity factor (ISF); in which, the FSI is a function of at least one blood glucose level measured from the plurality of physiological values.

[0011] The use of an FSI that is a function of at least one measured glucose level allows the recommendation unit to more accurately determine a recommendation value tailored to the needs of the individual user, given that the FSI can vary from one individual user to another and from one measured glucose level to another.

[0012] According to the present invention, the FSI is representative of the effect of a determined dose of insulin on a single user's blood glucose level.

[0013] According to the present invention, the FSI is representative of the effect of one unit of insulin on the measured blood glucose level of a single user. One unit of insulin corresponds to the "biological equivalent" of 34.7 qg of pure crystalline insulin.

[0014] According to the present invention, the FSI being a function of at least one blood glucose level measured from the plurality of physiological values ​​implies that the FSI changes according to the blood glucose level measured and changes for this reason over time if the blood glucose level measured is not constant over time.

[0015] According to one embodiment, the control device also includes an insulin delivery unit such as a subcutaneous insulin delivery device configured to deliver exogenous insulin into subcutaneous tissue of the patient in response to an insulin delivery control signal, in particular continuous infusion insulin such as basal insulin and / or bolus insulin.

[0016] According to one embodiment, the subcutaneous insulin delivery device is an insulin pump.

[0017] According to one embodiment, the recommendation value is a recommendation value corresponding to a quantity of insulin to be injected at a future time step.

[0018] According to one embodiment, the future time step does not exceed a few seconds after the nearest timestamp of the unique user's physiological values. This delay, which does not exceed a few seconds, corresponds to the calculation time required to determine the recommendation value. The recommendation value can be of any type, such as a bolus recommendation or a basal recommendation, for example.

[0019] According to the present invention, the terms injected or injection should be understood as a virtual injection in the case where the training of the reinforcement learning algorithm is carried out using a simulation in which the single user is a virtual user.

[0020] In one embodiment, the control device includes a conversion unit, the conversion unit being configured to convert the recommended insulin value into a control parameter of the fluid infusion device. In one embodiment, the control parameter takes the form of a bolus and / or a basal dose. Such a configuration allows the infusion device to infuse the recommended value to the single user.

[0021] According to the present invention, a measured glucose level is a glucose level measured on the individual user. The blood glucose level can be measured by any means such as a continuous glucose monitor (CGM) or a blood glucose monitor (BGM).

[0022] According to the present invention, a quantity of carbohydrates ingested by the single user corresponds to a quantity of sugar ingested during a meal, for example.

[0023] According to one embodiment, the FSI is a function of measured blood glucose levels with a timestamp no older than one to three hours prior to the present time, and preferably two hours prior to the present time. Such a configuration allows the recommendation unit to determine the recommendation value more accurately, since measured blood glucose levels with older timestamps are not as representative of the individual user's condition as blood glucose levels measured with more recent timestamps.

[0024] According to one embodiment, the FSI is a decreasing function of at least one blood glucose level measured from the plurality of physiological values.

[0025] Such a configuration allows the regulating device to precisely determine a recommendation value and, for this reason, to help manage the blood glucose levels of the individual user. Indeed, an FSI that is a decreasing function of at least one measured blood glucose level from the plurality of physiological values allows the recommendation unit to accurately determine a recommendation value.

[0026] According to one embodiment, the FSI is a decreasing function of at least one measured blood glucose level from the plurality of physiological values ​​in which the at least one measured blood glucose level considered from the plurality of physiological values ​​is greater than 100 mg / dl.

[0027] According to one embodiment, the FSI is a function of at least one blood glucose level measured from the plurality of physiological values ​​as follows:

[0028] [Math. 1]

[0029] FSI = mx Gly+b

[0030] Gly being at least one measured blood glucose level considered from the plurality of physiological values ​​such that Gly being the closest measured blood glucose level, in time, for example;

[0031] m being a factor; and

[0032] b being a constant.

[0033] Such a configuration allows the regulating device to accurately determine a recommendation value and, for this reason, to help manage the blood glucose levels of the individual user. Indeed, a Functional Status Indicator (FSI) that is a decreasing function of at least one measured blood glucose level from a plurality of physiological values ​​allows the recommendation unit to accurately determine a recommendation value.

[0034] According to one embodiment, m varies as a function of Gly such that m is in [1; 1.5] for Gly less than 100 mg / dL, m is in [-0.8; -0.6] for Gly in [100; 160[ mg / dL, and m is in [-0.2; 0] for Gly greater than 160 mg / dL. Such a configuration allows the FSI to be a representation closer to the physiological state of the single user. According to one embodiment, m varies as a function of Gly by a smooth transition from 1 to -0.8 according to the changes in m described above.

[0035] In one embodiment, b varies as a function of Gly such that b is in [-50; -30] for Gly less than 100 mg / dL, b is in [150; 170] for Gly in [100; 160[ mg / dL, and b is in [50; 70] for Gly greater than 160 mg / dL. Such a configuration allows the FSI to be a representation closer to the physiological state of the individual user. In another embodiment, b varies as a function of Gly by a smooth transition from -50 to 170 according to the evolutions of b described above. In yet another embodiment, FSI is a piecewise linear function with an undefined number of pieces.

[0036] According to one embodiment, the FSI is a piecewise affine function of Gly with m=ml and b=bl for Gly less than 100 mg / dl, m=m2 and b=b2 for Gly in [100; 160[ mg / dl and m=m3 and b=b3 for Gly greater than 160 mg / dl. A This configuration allows the FSI to be a representation closer to the physiological state of the individual user. In another embodiment, the FSI is a piecewise affine function with an undefined number of pieces. The functions of the piecewise affine function are nonlinear exponential functions, for example FSI(Gly) = ax eA((cx Gly)).

[0037] The control device 30 is configured to modify m and b as well as ml, m2, m3, bl, b2 and b3 over time. Such a configuration allows the control device 30 to adapt more precisely to the specifics of the individual user; m and b as well as ml, m2, m3, bl, b2 and b3 can be modified over time using any known method such as the self-learning method described in detail above applied to the FSI, for example.

[0038] According to one embodiment, the functions of the piecewise affine function are logarithmic functions ax log(Gly)+ b.

[0039] It should be noted that the examples of piecewise affine functions described above are linear and nonlinear functions. Piecewise affine functions can be a mixture of linear and nonlinear functions such as linear for Gly <100 mg / dL and nonlinear for Gly >=100 mg / dL.

[0040] According to one embodiment, the recommendation unit is configured to determine the recommendation value at least as a function of a predicted glucose level determined by calculating a physiological model of the glucose-insulin system representative of the single user using at least a part of the user data and the FSI.

[0041] Such a configuration makes it possible to determine a precise recommendation value and, therefore, to help manage the blood glucose levels of the individual user. Indeed, the use of a Functional Stress Intake (FSI) allows the recommendation value to be adapted specifically to the individual user. Furthermore, since an FSI is a function of at least one measured blood glucose level from a plurality of physiological values, the physiological model allows the physiological model to faithfully reproduce the physiology of the individual user and, therefore, enables the recommendation unit to determine a precise recommendation value.

[0042] Such a configuration also makes it possible to test or refine an existing algorithm more precisely using a simulator, for example. Indeed, an improved physiological model more faithfully reproduces the physiology of the individual user and therefore allows for a more precise evaluation of the existing algorithm.

[0043] According to one embodiment, the physiological model can be of any type such as the Hovorka model, the minimal Bergman model, or the Dalla Man model, for example. According to a preferred embodiment, the physiological model is the Hovorka model or a model derived from the Hovorka model.

[0044] According to one embodiment, the recommendation unit is configured to determine the recommendation value based on at least one product integral derivative (PID) approach and one FSI. The PID is a function of at least one measured or predicted blood glucose level and at least one target blood glucose level; the FSI is then used as a weighting factor.

[0045] According to one embodiment, the recommendation unit is configured to determine the recommendation value based on at least one neural network and one FSI. The neural network uses as input at least one quantity of insulin infused into the single user, at least one quantity of carbohydrates ingested by the single user, and at least one physiological value of the single user, and outputs a raw recommendation value. The FSI is then used as a weighting factor. The neural network can be of any type and trained as follows: • collect data; • preprocess the data; • design a neural network architecture such as a regression model for example; • define the input characteristics; • define the output characteristics; • train the neural network using pre-processed training data; • validate the model's performance; and • evaluate the performance of the model using measures such as mean absolute error (MAE) or mean squared error (MSE) for example.

[0046] According to one embodiment, the recommendation unit is configured to determine the recommendation value at least as a function of a predicted glucose level determined by calculating a physiological model of the glucose-insulin system representative of the single user using a rate of action of insulin in which the rate of action of insulin is an increasing function of at least one blood glucose level measured from the plurality of physiological values.

[0047] Such a configuration makes it possible to obtain a better value of active insulin (IOB) and for this reason allows the recommendation unit to determine a recommendation value more precisely.

[0048] According to one embodiment, at least one blood glucose level measured from the plurality of physiological values ​​corresponds to at least one blood glucose level measured from the plurality of physiological values ​​or to the last few blood glucose levels measured from the plurality of physiological values, or to a average of some recent blood glucose levels measured from the plurality of physiological values.

[0049] According to one embodiment, the FSI is not a function of a total daily dose, nor, a fortiori, of an average of several previous total daily doses. The total daily dose is the total amount of insulin infused to the single user during a day. Indeed, the FSI is not a constant of the blood glucose level.

[0050] According to one embodiment, the FSI is a decreasing function of at least one blood glucose level measured from the plurality of physiological values ​​and depends on at least one previous FSI.

[0051] Such a configuration makes it possible to precisely determine a recommendation value and, for this reason, to manage the blood glucose levels of the individual user while reducing the risk to the individual user. Indeed, an FSI dependent on at least one previous FSI makes it possible to avoid abrupt changes in the determination of the recommendation value.

[0052] According to the present invention, a prior FSI is a previously calculated FSI. Such a prior FSI may, for example, be calculated in the same way as the FSI or may be an arbitrarily initialized prior FSI.

[0053] According to one embodiment, the FSI is a decreasing function of at least one blood glucose level measured from the plurality of physiological values ​​and depends on an average of several previous FSIs.

[0054] Such a configuration makes it possible to further mitigate changes in the determination of the recommendation value and for this reason to reduce the risk for the single user.

[0055] According to one embodiment, the FSI depends on a self-learned blood glucose factor such that FSI = FSImoy * Gly * glyF

[0056] in which:

[0057] FSImoy being an average of previous FSIs; and

[0058] glyF being a glycemic factor.

[0059] According to one embodiment, FSImoy is the average of all FSIs calculated during a week.

[0060] According to one embodiment, the FSI is a decreasing function of at least one measured blood glucose level of the plurality of physiological values ​​and depends on a self-learned FSI.

[0061] According to one embodiment, the FSI depends on a self-learned blood glucose factor such that FSI = FSIAA * Gly * glyF

[0062] in which:

[0063] FSIAA being the self-learning FSI;

[0064] Such a configuration makes it possible to precisely determine a recommendation value and, therefore, to manage the glucose levels of the individual user by adapting to the individual user over time. Such a configuration is particularly useful when the variability of the individual user's FSI as a function of the measured blood glucose level is high.

[0065] Such a configuration makes it possible to precisely determine a recommended value and, therefore, to manage the glucose levels of a single user by adapting to the individual user over time. Indeed, the individual user's FSI can vary depending on several factors, such as a non-diabetic illness, for example.

[0066] According to the present invention, the term "self-learning FSI" refers to an FSI determined by means of a self-learning process. The self-learning process can be of any type, such as a process by which a system, generally a computer program or artificial intelligence, autonomously acquires knowledge, skills, or behaviors without explicit programming or direct human intervention. Self-learning systems often rely on algorithms, data, and feedback mechanisms to improve their performance over time, adapting to changing conditions or tasks through self-adjustment and refinement.

[0067] According to one embodiment, the self-learning FSI is determined by the creation of a plurality of time segments, each time segment having an FSI, in order to group user data at least according to their timestamp; the creation of a segment correction value for each time segment, the segment correction values ​​being a function of at least one previous FSI and at least one physiological value from the plurality of physiological values ​​during a time segment of the plurality of time segments; the creation of the self-learning FSI for a future determined period of time, the self-learning FSI being a function of at least one segment correction value from the segment correction values ​​and the FSI of the determined period of time.The use of segments makes it possible to determine the self-learned FSI precisely, since the correction values ​​for each segment can be created differently, either time segment by time segment or based on a specific interest of a time segment, for example.

[0068] According to one embodiment, the FSI depends on a self-learned blood glucose factor.

[0069] According to one embodiment, the FSI depends on a self-learned blood glucose factor such that FSI = pFSI * Gly * glyFAA

[0070] in which:

[0071] pFSI being a prior FSI;

[0072] FSImoy is an average of previous FSIs such as the five closest previous FSIs, in time, for example;

[0073] glyFAA being the self-learned blood glucose factor.

[0074] Such a configuration makes it possible to precisely determine a recommendation value and, therefore, to manage the glucose levels of the individual user by adapting to the individual user over time. Such a configuration is particularly useful when the variability of the individual user's FSI as a function of the measured blood glucose level is high.

[0075] According to one embodiment, pFSI can be replaced by FSImoy. Indeed, FSI = FSImoy * Gly * glyFAA allows for a smoother transition in determining the recommendation value and therefore reduces the risk for the individual user.

[0076] According to one embodiment, glyFAA is a combination of a factor and a constant term.

[0077] According to the present invention, the term "self-learning blood glucose factor" refers to a blood glucose factor determined using a self-learning process. The self-learning process can be of any type, such as a process by which a system, generally a computer program or artificial intelligence, autonomously acquires knowledge, skills, or behaviors without explicit programming or direct human intervention. Self-learning systems often rely on algorithms, data, and feedback mechanisms to improve their performance over time, adapting to changing conditions or tasks through self-adjustment and refinement.

[0078] According to one embodiment, the self-learned blood glucose factor is determined by the creation of a plurality of time segments, each time segment having a blood glucose factor, in order to group user data at least according to their timestamp; the creation of a segment correction value for each time segment, the segment correction value being a function of at least one previous blood glucose factor and at least one physiological value of the plurality of physiological values ​​during a time segment of the plurality of time segments; the creation of the self-learned blood glucose factor for a future determined period of time, the self-learned blood glucose factor being a function of at least one segment correction value of the segment correction values ​​and the blood glucose factor of the determined period of time.The use of segments allows for the precise determination of the self-learned blood glucose factor since the correction values ​​for each segment can be created differently, time segment. by time segment or based on a specific interest of a time segment, for example.

[0079] According to one embodiment, the FSI is a decreasing function of at least one measured blood glucose level of the plurality of physiological values ​​and depends on a self-learned FSI and a self-learned glucose factor.

[0080] According to one embodiment, the FSI depends on a self-learned FSI and a self-learned glucose factor such that FSI = FSIAA * Gly * glyFAA.

[0081] Such a configuration makes it possible to precisely determine a recommended value and, therefore, to manage the glucose levels of a single user by adapting to the individual user over time. Indeed, the FSI and glyF of a single user can vary depending on several factors, such as a disease unrelated to diabetes, for example.

[0082] According to one embodiment, the recommendation unit is configured to determine the recommendation value at least as a function of the FSI and a correction factor.

[0083] According to one embodiment, the correction factor is a factor applied to the FSI and allows the recommendation unit to take into account particular situations of the single user such as a disease not related to diabetes for example.

[0084] According to one embodiment, the recommendation unit considers the FSI multiplied by the correction factor as the FSI to determine the recommendation value.

[0085] The invention also relates to a method for determining a recommended value for a control parameter of an insulin infusion device, the method being implemented by the control device as described above and comprising the following steps • retrieving user data; and • the determination of the recommendation value at least based on the FSI; in which the FSI is a function of at least one blood glucose level measured from the plurality of physiological values.

[0086] The embodiments, technical effects, and definitions disclosed herein with respect to the control device are also applicable to the process described herein. The process encompasses steps that comprehensively utilize the functionalities and characteristics of the control device described herein. Therefore, all embodiments, technical effects, and definitions relating to the device, including, but not limited to, the evolution of the FSI over time and how it is calculated, are equally applicable to the process. This ensures a complete and unified understanding. both aspects of the invention relating to the control device and the method, which facilitates the implementation and use of the disclosed technology across a whole range of applications.

[0087] The invention also relates to software comprising instructions to cause the control device described above to execute the steps of the process described above. Brief description of the drawings

[0088] Embodiments of the invention will be described below with reference to the drawings, briefly described below:

[0089] Figure 1 illustrates a method for validating a control algorithm according to one embodiment of the invention; and

[0090] Fig. 2 illustrates a regulation device according to one embodiment of the invention. DETAILED DESCRIPTION OF THE INVENTION

[0091] Figure 1 represents a control device 30 for determining a recommendation value for a control parameter of an insulin infusion device 20. The recommendation value is a recommended amount of insulin to be injected at a future time step. The future time step does not exceed a few seconds after the nearest timestamp of the single user's physiological values. This delay, which does not exceed a few seconds, corresponds to the computation time required to determine the recommendation value. The recommendation value can be of any type, such as a bolus recommendation or a basal recommendation, for example. The terms injected or injection should be understood as a virtual injection in the case where the training of a reinforcement learning algorithm is performed using a simulation in which the single user is a virtual user.Such a simulation could be useful for validating or improving an algorithm and therefore reducing the risk for the individual user. Otherwise, the terms injected or injection should be understood as a regular injection.

[0092] The control device 30 includes a retrieval unit 32, the retrieval unit 32 being configured to retrieve user data. Each user data item has a timestamp, and the user data is linked to a single user. The user data includes the amount of insulin infused into the single user, the amount of carbohydrates ingested by the single user, and a plurality of physiological values ​​of the single user, the plurality of physiological values ​​of the single user including at least measured blood glucose levels.

[0093] A measured blood glucose level is a blood glucose level measured on the individual user. The blood glucose level can be measured by any means such as a continuous glucose monitor (CGM) 12 or a blood glucose monitor (BGM). Preferably, the blood glucose level is measured using a CGM 12. A quantity of carbohydrates ingested by the individual user corresponds to a quantity of sugar ingested during a meal, for example.

[0094] The regulating device 30 also includes a recommendation unit 34. The recommendation unit 34 is configured to determine the recommendation value based on at least one insulin sensitivity factor (ISF). The ISF is representative of the effect of a given dose of insulin on the single user's blood glucose level. More precisely, the ISF is representative of the effect of one unit of insulin on the single user's measured blood glucose level. One unit of insulin corresponds to the "biological equivalent" of 34.7 mcg of pure crystalline insulin.

[0095] According to one embodiment, the recommendation unit 34 is configured to determine the recommendation value based at least on a predicted glucose level determined by calculating a physiological model of the glucose-insulin system representative of the individual user, using at least some of the user data and the FSI. Such a configuration allows the recommendation unit 34 to accurately determine a recommendation value and, for this reason, to help manage the individual user's blood glucose levels. Indeed, the use of an FSI allows the control device 30 to adapt the recommendation value specifically to the individual user.Furthermore, since an FSI is a function of at least one measured blood glucose level across a plurality of physiological values, the physiological model accurately reproduces the physiology of the individual user and, therefore, allows the recommendation unit to precisely determine a recommendation value. Such a configuration also allows for more precise testing or refinement of an existing algorithm using a simulator, for example. Indeed, an improved physiological model more accurately reproduces the physiology of the individual user and thus allows for a more precise evaluation of the existing algorithm. The physiological model can be of any type, such as the Hovorka model, the minimal Bergman model, or the Dalla Man model, for example. Preferably, the physiological model is the Hovorka model or a model derived from the Hovorka model.

[0096] According to one embodiment, the recommendation unit 34 can also be configured to determine the recommendation value at least as a function of a product integral derivative (PID) approach and an FSI. The PID is a function of minus a measured or predicted blood glucose level and at least one target blood glucose level, the FSI is then used as a weighting factor.

[0097] According to one embodiment, the recommendation unit 34 can also be configured to determine the recommendation value based on at least one neural network and one FSI. The neural network uses as input at least one quantity of insulin infused into the single user, at least one quantity of carbohydrates ingested by the single user, and at least one physiological value of the single user, and outputs a raw recommendation value. The FSI is then used as a weighting factor. The neural network can be of any type and trained as follows: • collect data; • preprocess the data; • design a neural network architecture such as a regression model for example; • define the input characteristics; • define the output characteristics; • train the neural network using pre-processed training data; • validate the model's performance; and • evaluate the performance of the model using measures such as mean absolute error (MAE) or mean squared error (MSE) for example.

[0098] The recommendation unit 34 is configured to determine the recommendation value based on at least one predicted glucose level determined by calculating a physiological model of the glucose-insulin system representative of the individual user, using an insulin action rate in which the insulin action rate is an increasing function of at least one measured blood glucose level from a plurality of physiological values. Such a configuration allows for a better active insulin (OIB) value and therefore enables the recommendation unit 34 to determine a recommendation value more accurately.At least one blood glucose level measured from the plurality of physiological values ​​corresponds to at least one blood glucose level measured from the plurality of physiological values, or to the last few blood glucose levels measured from the plurality of physiological values, or to an average of the last few blood glucose levels measured from the plurality of physiological values.

[0099] It must be understood that, in the present invention, the FSI is not a function of a total daily dose nor, a fortiori, of an average of several previous total daily doses. The total daily dose is the total amount of insulin infused to the single user during a day. Indeed, the FSI is not a constant blood glucose level.

[0100] The FSI is a function of at least one measured blood glucose level from a plurality of physiological values. Using an FSI based on at least one measured blood glucose level allows the recommendation unit 34 to more accurately determine a recommendation value tailored to the needs of the individual user, since the FSI can vary from one individual user to another and from one measured blood glucose level to another. The fact that the FSI is a function of at least one measured blood glucose level from a plurality of physiological values ​​implies that the FSI changes according to the measured blood glucose level and therefore changes over time if the measured blood glucose level is not constant over time. The FSI is a function of measured blood glucose levels with a timestamp no older than one to three hours before the present time, and preferably two hours before the present time.Such a configuration allows recommendation unit 34 to determine the recommendation value more accurately, given that blood glucose levels measured with older timestamps are not as representative of the individual user's condition as blood glucose levels measured with more recent timestamps.

[0101] The FSI is a decreasing function of at least one measured blood glucose level from a plurality of physiological values, in which at least one measured blood glucose level from the plurality of physiological values ​​is greater than 100 mg / dL. Such a configuration allows the regulating device 30 to accurately determine a recommendation value and, for this reason, to help manage the blood glucose levels of the individual user. Indeed, an FSI that is a decreasing function of at least one measured blood glucose level from a plurality of physiological values ​​allows the recommendation unit 34 to accurately determine a recommendation value.

[0102] The FSI can be calculated as a function of at least one blood glucose level measured from the plurality of physiological values ​​as follows:

[0103] [Math. 2]

[0104] FSI = mx Gly+b

[0105] Gly being at least one measured blood glucose level considered from the plurality of physiological values ​​such that Gly being the closest measured blood glucose level, in time, for example;

[0106] m being a factor; and

[0107] b being a constant.

[0108] Such a configuration allows the control device 30 to precisely determine a recommendation value and for this reason to help manage the blood glucose level of the single user. Indeed, an FSI which is a decreasing function of at least one measured blood glucose level of the plurality of physiological values ​​allows the recommendation unit 34 to precisely determine a recommendation value.

[0109] According to one embodiment, the FSI is a piecewise linear function of Gly with m=ml and b=bl for Gly less than 100 mg / dl, m=m2 and b=b2 for Gly in [100; 160[ mg / dl, and m=m3 and b=b3 for Gly greater than 160 mg / dl. Such a configuration allows the FSI to be a representation closer to the physiological state of the individual user. Consequently, b varies as a function of Gly such that bl in [-50; -30] for Gly less than 100 mg / dl, b2 in [150; 170] for Gly in [100; 160[ mg / dl, and b3 in [50; 70] for Gly greater than 160 mg / dl. Such a configuration allows the FSI to be a representation closer to the physiological state of the individual user. In one embodiment, b varies with Gly via a smooth transition from -50 to 170 according to the previously described evolutions of b. In another embodiment, FSI is a piecewise linear function with an undefined number of pieces.Furthermore, m varies according to Gly such that ml is in [1; 1.5] for Gly less than 100 mg / dl, m2 in [-0.8; -0.6] for Gly in [100; 160[ mg / dl, and m3 in [-0.2; 0] for Gly greater than 160 mg / dl. Such a configuration allows the FSI to be a representation closer to the physiological state of the individual user. According to one embodiment, m varies according to Gly by a smooth transition from 1 to -0.8 according to the changes in m described above.

[0110] According to another embodiment, the FSI is a piecewise affine function with an undefined number of pieces. The functions of the piecewise affine function are nonlinear exponential functions, for example FSI(Gly) = ax eA((cx Gly)).

[0111] According to one embodiment, the functions of the piecewise affine function are logarithmic functions ax log(Gly)+ b.

[0112] It should be noted that the examples of piecewise affine functions described above are linear and nonlinear functions. Piecewise affine functions can be a mixture of linear and nonlinear functions such as linear for Gly <100 mg / dL and nonlinear for Gly >=100 mg / dL.

[0113] The control device 30 is configured to modify m and b as well as ml, m2, m3, bl, b2 and b3 over time. Such a configuration allows the control device 30 to adapt more precisely to the specifics of the single user; m and b as well as ml, m2, m3, bl, b2 and b3 can be modified over time using any known method such as the self-learning method described in detail above applied to the FSI for example.

[0114] The FSI preferably depends on at least one previous FSI. Such a configuration allows the recommendation unit 34 to accurately determine a recommendation value and, for this reason, to manage the single user's blood glucose levels while reducing the risk to the single user. Indeed, an FSI dependent on at least one previous FSI avoids abrupt changes in the determination of the recommendation value. A previous FSI is an FSI calculated previously. Such a previous FSI may, for example, be calculated in the same way as the current FSI or may be a previous FSI arbitrarily initialized by a healthcare professional, for example.

[0115] According to this reasoning, the FSI depends on an average of several previous FSIs. Such a configuration allows the recommendation unit 34 to further mitigate changes in the determination of the recommendation value and for this reason reduce the risk for the individual user.

[0116] According to one embodiment, the FSI depends on a self-learned blood glucose factor such that FSI = FSImoy * Gly * glyF

[0117] in which:

[0118] FSImoy being the average of all FSIs calculated during a week; and

[0119] glyF being a blood glucose factor.

[0120] According to one embodiment, the FSI depends on a self-learned blood glucose factor such that FSI = FSIAA * Gly * glyF

[0121] in which:

[0122] FSIAA being the self-learned FSI;

[0123] Such a configuration allows the recommendation unit 34 to accurately determine a recommendation value and, for this reason, to manage the single user's glucose levels by adapting to the individual user over time. Such a configuration is particularly useful when the variability of the single user's FSI (Frequency Sensitivity Index) as a function of the measured blood glucose level is high. This configuration also allows the recommendation unit 34 to accurately determine a recommendation value and, for this reason, to manage the single user's glucose levels by adapting to the individual user over time. Indeed, the single user's FSI can vary depending on several factors, such as a non-diabetic condition, for example. According to the present invention, the term diabetes refers to dependence on external insulin.

[0124] The term “self-learning FSI” refers to an FSI determined by means of a self-learning process. The self-learning process can be of any type, such as a process by which a system, generally a computer program or artificial intelligence, autonomously acquires knowledge, skills, or behaviors without explicit programming. or direct human intervention. Self-learning systems often rely on algorithms, data, and feedback mechanisms to improve their performance over time, adapting to changing conditions or tasks through self-adjustment and refinement.

[0125] According to one embodiment, the self-learning FSI is determined by the creation of a plurality of time segments, each time segment having an FSI, in order to group user data at least according to their timestamp; the creation of a segment correction value for each time segment, the segment correction values ​​being a function of at least one previous FSI and at least one physiological value from the plurality of physiological values ​​during a time segment of the plurality of time segments; the creation of the self-learning FSI for a future determined period of time, the self-learning FSI being a function of at least one segment correction value from the segment correction values ​​and the FSI of the determined period of time.The use of segments makes it possible to determine the self-learned FSI precisely, since the correction values ​​for each segment can be created differently, either time segment by time segment or based on a specific interest of a time segment, for example.

[0126] The FSI can also depend on a self-learned glucose factor: FSI = pFSI * Gly * glyFAA

[0127] in which:

[0128] pFSI being a prior FSI;

[0129] FSImoy is an average of previous FSIs such as the five closest previous FSIs, in time, for example;

[0130] glyFAA being the self-learned blood glucose factor.

[0131] Such a configuration allows the recommendation unit to accurately determine a recommendation value and, therefore, to manage the single user's glucose levels by adapting to the individual user over time. This configuration is particularly useful when the variability of the single user's FSI as a function of the measured blood glucose level is high. The pFSI can be replaced by the pFSImoy. Indeed, FSI = FSImoy * Gly * glyFAA further smooths the changes in determining the recommendation value and, therefore, reduces the risk for the individual user.

[0132] According to the present invention, the term "self-learning blood glucose factor" refers to a blood glucose factor determined using a self-learning process. The self-learning process can be of any type, such as a process by which a system, generally a computer program or artificial intelligence, autonomously acquires knowledge, skills, or behaviors without explicit programming or intervention. direct human interaction. Self-learning systems often rely on algorithms, data, and feedback mechanisms to improve their performance over time, adapting to changing conditions or tasks through self-adjustment and refinement.

[0133] According to one embodiment, the self-learned blood glucose factor is determined by the creation of a plurality of time segments, each time segment having a blood glucose factor, in order to group user data at least according to their timestamp; the creation of a segment correction value for each time segment, the segment correction value being a function of at least one previous blood glucose factor and at least one physiological value of the plurality of physiological values ​​during a time segment of the plurality of time segments; the creation of the self-learned blood glucose factor for a future determined period of time, the self-learned blood glucose factor being a function of at least one segment correction value of the segment correction values ​​and the blood glucose factor of the determined period of time.The use of segments makes it possible to determine precisely the self-learned blood glucose factor since the correction values ​​for each segment can be created differently, segment by segment or according to a specific interest of a time segment for example.

[0134] According to one embodiment, the FSI is a combination of the embodiments described above, given that the FSI is a decreasing function of at least one measured blood glucose level from a plurality of physiological values ​​and depends on a self-learned FSI and a self-learned glucose factor. The FSI depends on a self-learned FSI and a self-learned glucose factor such that FSI = FSIAA * Gly * glyFAA. Such a configuration makes it possible to determine a precise recommendation value and, therefore, to manage the glucose levels of the individual user by adapting to the individual user over time. Indeed, the FSI and the glyF of the individual user can vary depending on several factors, such as a non-diabetic disease, for example.

[0135] Recommendation unit 34 is configured to determine the recommendation value based at least on the FSI and a correction factor. The correction factor is a factor applied to the FSI and allows recommendation unit 34 to take into account specific situations of the individual user, such as a disease unrelated to diabetes, for example. Recommendation unit 34 considers the FSI multiplied by the correction factor as the FSI to determine the recommendation value.

[0136] The regulating device 30 also includes an insulin delivery unit 20. The preferred insulin delivery unit 20 is a device A subcutaneous insulin delivery system configured to administer exogenous insulin into the patient's subcutaneous tissue in response to an insulin delivery regulatory signal, such as from an insulin pump. Specifically, continuous infusion and / or bolus insulin delivery.

[0137] According to one embodiment, the control device 30 includes a conversion unit not shown in the drawings. The conversion unit is configured to convert the recommended insulin value into a control parameter of the fluid infusion device 20. The control parameter takes the form of a bolus and / or a basal dose. Such a configuration allows the infusion device to infuse the recommended value to the single user.

[0138] As illustrated in [Fig. 2], the invention also relates to a method for determining a recommended value for a control parameter of an insulin infusion device 20 as described above. The embodiments, technical effects, and definitions disclosed herein with respect to the control device 30 are also applicable to the method described herein. The method includes steps that fully utilize the functionalities and features of the control device 30 described herein. Therefore, all embodiments, technical effects, and definitions relating to the device, including, but not limited to, the evolution of the FSI over time and how it is calculated, are equally applicable to the method.This ensures a complete and unified understanding of both the aspects of the invention relating to the control device 30 and the method, which facilitates the implementation and use of the disclosed technology across a wide range of applications.

[0139] The process comprises the following steps: • the retrieval of 50 user data; and • the determination 58 of the recommendation value at least based on the FSI; in which the FSI is a function of at least one blood glucose level measured from the plurality of physiological values.

[0140] According to an embodiment in which the FS depends on the self-learned glucose factor such as FSI = FSIAA * Gly * glyF, the process also comprises the following steps: • the creation of a plurality of time segments 52, each time segment having an FSI, in order to group user data at least according to their timestamp; • the creation of a segment correction value for each time segment 54, the segment correction values ​​being a function of at least one previous FSI and at least one physiological value from the plurality of values physiological during a time segment of the plurality of time segments; • the creation of the self-learning FSI for a future determined time period 56, the self-learning FSI being a function of at least one segment correction value of the segment correction values ​​and the FSI of the determined time period.

[0141] The use of segments makes it possible to determine precisely the self-learned FSI since the correction values ​​of each segment can be created differently, time segment by time segment or according to a specific interest of a time segment for example.

[0142] The invention also relates to software comprising instructions for causing the control device 30 described above to execute the steps of the process described above.

[0143] The invention also relates to a regulation system 10 comprising an insulin infusion device 20, an MGC 12 and a regulation device 30 comprising a recommendation unit 34 and a recovery unit 32.

[0144] Although exemplary embodiments of the invention have been described, it will be understood by those skilled in the art that various changes, omissions, and / or additions may be made, and that equivalents may be substituted for elements thereof without departing from the spirit of the scope of the invention. Furthermore, numerous modifications may be made to adapt a particular situation or material to the teachings of the invention without departing from its scope. Therefore, it is intended that the invention is not limited to the particular embodiments disclosed for the implementation of this invention, but that the invention will include all embodiments falling within the scope of the supplementary claims. Moreover, unless specifically stated, any use of the terms first, second, etc.does not designate any order or importance, but instead the terms first, second, etc. are used to distinguish one element from another.

Claims

Demands

1. A control device (30) for determining a recommendation value for a control parameter of an insulin infusion device (20), the control device (30) comprising: • a retrieval unit (32), the retrieval unit (32) being configured to retrieve user data, each piece of user data having a timestamp and the user data being associated with a unique user, the user data comprising at least: • an amount of insulin infused to the unique user; • an amount of carbohydrates ingested by the unique user; • a plurality of physiological values ​​of the unique user, the plurality of physiological values ​​of the unique user comprising at least measured blood glucose levels;• a recommendation unit (34), the recommendation unit (34) being configured to determine the recommendation value at least as a function of an insulin sensitivity factor (ISF); in which, the ISF is a function of at least one blood glucose level measured from the plurality of physiological values.

2. Control device (30) for determining a recommendation value of a control parameter of an insulin infusion device (20) according to claim 1, wherein the FSI is a decreasing function of at least one blood glucose level measured from the plurality of physiological values.

3. A control device (30) for determining a recommended value for a control parameter of an insulin infusion device (20) according to claim 1, wherein the FSI is a function of at least one blood glucose level measured from the plurality of physiological values ​​as follows: FSI = mx Gly+b Gly being at least one measured blood glucose level considered from the plurality of physiological values ​​such that Gly being the closest measured blood glucose level, in time, for example; m being a factor; and b being a constant.

4. Control device (30) for determining a recommendation value of a control parameter of an insulin infusion device (20) according to any one of claims 1 to 3, wherein the recommendation unit (34) is configured to determine the recommendation value at least as a function of a predicted glucose level determined by calculating a physiological model of the glucose-insulin system representative of the single user using at least a portion of the user data and the FSI.

5. Control device (30) for determining a recommendation value of a control parameter of an insulin infusion device (20) according to claim 4, wherein the recommendation unit (34) is configured to determine the recommendation value at least as a function of a predicted glucose level determined by calculating a physiological model of the glucose-insulin system representative of the single user using an insulin action rate in which the insulin action rate is an increasing function of at least one blood glucose level measured from the plurality of physiological values.

6. Control device (30) for determining a recommendation value of a control parameter of an insulin infusion device (20) according to any one of claims 1 to 5, wherein the FSI is a decreasing function of at least one blood glucose level measured from the plurality of physiological values ​​and depends on at least one previous FSI.

7. Control device (30) for determining a recommendation value of a control parameter of an insulin infusion device (20) according to any one of claims 1 to 6, wherein the FSI is a decreasing function of at least one blood glucose level measured from the plurality of physiological values ​​and depends on an average of several previous FSIs.

8. Control device (30) for determining a recommended value for a control parameter of a device insulin infusion (20) according to any one of claims 1 to 7, wherein the FSI is a decreasing function of at least one blood glucose level measured from the plurality of physiological values ​​and depends on a self-learned FSI.

9. Control device (30) for determining a recommendation value of a control parameter of an insulin infusion device (20) according to any one of claims 1 to 8, wherein the FSI depends on a self-learned blood glucose factor.

10. Control device (30) for determining a recommendation value of a control parameter of an insulin infusion device (20) according to any one of claims 1 to 9, wherein the FSI is a decreasing function of at least one blood glucose level measured from the plurality of physiological values ​​and depends on a self-learned FSI and a self-learned glucose factor.

11. Control device (30) for determining a recommendation value of a control parameter of an insulin infusion device (20) according to any one of claims 1 to 10, wherein the recommendation unit (34) is configured to determine the recommendation value at least as a function of the FSI and a correction factor.

12. A method for determining a recommendation value for a control parameter of an insulin infusion device (20), the method being implemented by the control device (30) according to claim 1 and comprising the following steps: • retrieval (50) of user data; and • determination (52) of the recommendation value at least as a function of the FSI; wherein the FSI is a function of at least one blood glucose level measured from the plurality of physiological values.

13. Computer program comprising instructions to cause the control device (20) according to claim 1 to perform the steps of the process according to claim 12.

Images