• 专利附录
  • 专利说明
  • 权利要求
  • 类似技术
  • 同族专利
  • 引用文献
  • 法律状态
  • 优先权
    • 专利摘要:
    An automated system for controlling a patient's blood glucose, comprising: a blood glucose sensor (101); an insulin injection device (103); and a treatment and control unit (105) adapted to predict the future evolution of the patient's blood glucose from a physiological model and to control the insulin injection device (103) taking into account this prediction , wherein: the physiological model comprises a system of differential equations describing the evolution of a plurality of state variables as a function of time; and the processing and control unit (105) is adapted to implement an automatic calibration step of the physiological model comprising a step of estimating initial values of the state variables by minimizing a magnitude representative of the error, during a past observation period, between the glycemia estimated from the physiological model and the glucose measured by the sensor (101).
    公开号:FR3056094A1
    申请号:FR1658881
    申请日:2016-09-21
    公开日:2018-03-23
    发明作者:Pierre Jallon
    申请人:Commissariat a lEnergie Atomique CEA；Commissariat a lEnergie Atomique et aux Energies Alternatives CEA；
    IPC主号:
    专利说明:

    Field
    The present application relates to the field of automated blood sugar regulation systems, also called artificial pancreas.
    Presentation of the prior art
    An artificial pancreas is a system for automatically regulating the insulin intake of a diabetic patient based on his blood sugar history, his meal-taking history, and his insulin injection history.
    We are particularly interested here in MPC (Model-based Predictive Control) type regulation systems, also called predictive control systems, in which the regulation of the dose of insulin administered takes account of a prediction of the future evolution of the patient's blood sugar, carried out from a physiological model describing the assimilation of insulin by the patient's body and its impact on the patient's blood sugar.
    It would be desirable to be able to improve the performance of artificial pancreas with predictive control, and, more particularly, to be able to improve the quality of the prediction of the patient's future blood sugar level, so as to be able to
    B15018 - DD16959LP monitor insulin supplies with greater relevance and limit the risks of placing the patient in a situation of hyperglycemia or hypoglycemia.
    It would also be desirable to be able to limit the risks for the patient linked to a possible failure of the physiological model used to predict the patient's future blood sugar level.
    summary
    Thus, one embodiment provides an automated system for regulating the blood sugar level of a patient, comprising:
    a blood sugar sensor;
    an insulin injection device; and a processing and control unit adapted to predict the future development of the patient's blood sugar on the basis of a physiological model and to control the insulin injection device taking account of this prediction, in which:
    the physiological model includes a system of differential equations describing the evolution of a plurality of state variables as a function of time; and the processing and control unit is adapted to implement a step of automatic calibration of the physiological model comprising a step of estimating initial values of the state variables by minimizing a quantity representative of the error, during a past observation period, between the blood sugar estimated from the physiological model and the blood sugar measured by the sensor.
    According to one embodiment, the quantity is representative of the area between a first curve g representative of the temporal evolution of the glycemia estimated from the model over the observation period, and a second curve g representative of the evolution time of the blood glucose measured by the sensor over the observation period.
    According to one embodiment, the quantity is defined as follows:
    B15018 - DD16959LP to + ΔΤ m = δ Σ - ^ ι 2 t = t 0 where t is a discretized time variable, to is the start time of the observation phase, and tQ + tiT is the end time of the observation phase.
    According to one embodiment, the calibration method further comprises a step of estimating parameters of the system of differential equations by minimizing said quantity.
    According to one embodiment, the calibration method comprises a plurality of successive iterations of the following steps a) and b):
    a) estimating the parameters of the system of differential equations by minimizing said quantity by fixing the initial values of the state variables; and
    b) estimate the initial values of the state variables by minimizing said quantity by fixing the parameters of the system of differential equations.
    According to one embodiment, at the first iteration of step a), the initial values of the state variables are determined analytically by making the assumption that all the derivatives of the system of differential equations are zero.
    According to one embodiment, to simulate the evolution of the patient's blood sugar from the physiological model, the processing and control unit takes into account the history of insulin injected into the patient by the injection device and the history of glucose ingested by the patient.
    According to one embodiment, the physiological model is the Hovorka model.
    Another embodiment provides an automated regulation process for a patient's blood sugar, comprising:
    a step of calculating, by means of a processing and control unit, a prediction of the future evolution of the patient's glycemia from a physiological model comprising
    B15018 - DD16959LP a system of differential equations describing the evolution of a plurality of state variables as a function of time;
    a step of controlling an insulin injection device taking this prediction into account; and a step of automatic calibration of the physiological model comprising a step of estimating initial values of the state variables by minimizing a quantity representative of the error, during a past observation period, between the glycemia estimated from the physiological model and blood sugar measured on the patient by a blood sugar sensor.
    According to one embodiment, the method further comprises a step of estimating parameters of the system of differential equations by minimizing said quantity.
    According to one embodiment, the calibration step comprises a plurality of successive iterations of the following steps a) and b):
    a) estimating the parameters of the system of differential equations by minimizing said quantity by fixing the initial values of the state variables; and
    b) estimate the initial values of the state variables by minimizing said quantity by fixing the parameters of the system of differential equations.
    Another embodiment provides an automated system for regulating the glycemia of a patient, comprising:
    a blood sugar sensor;
    an insulin injection device; and a processing and control unit adapted to predict the future development of the patient's blood sugar on the basis of a physiological model and to control the insulin injection device taking account of this prediction, in which the unit processing and control is suitable for:
    a) implementing an automatic calibration step of the physiological model taking into account a history of
    B15018 - DD16959LP blood sugar measured by the sensor during a past observation period;
    b) at the end of the calibration step, determine whether the model is satisfactory or not from at least one digital indicator representative of the error between the blood sugar estimated from the model and the real blood sugar measured by the sensor; and
    c) if the quality of the model is not satisfactory, order the insulin injection device without taking into account the prediction made from the model.
    According to one embodiment, the digital indicator comprises an indicator m representative of the area between a first curve g representative of the temporal evolution of the glycemia estimated from the model over the observation period, and a second curve g representative of the temporal evolution of the glycemia measured by the sensor over the observation period.
    According to one embodiment, the indicator m is defined as follows:
    to + ΔΤ m = δϊ Σ t = t 0 where t is a discretized time variable, îq is the start time of the observation phase, and to + 4T is the end time of the observation phase.
    According to one embodiment, the digital indicator comprises a my indicator representative of the difference between the glycemia estimated from the model and the glycemia measured by the sensor at a given instant.
    According to one embodiment, the digital indicator comprises a mg indicator representative of the difference between the derivative of the glycemia estimated from the model and the derivative of the glycemia measured by the sensor at a given instant.
    B15018 - DD16959LP
    According to one embodiment, in step c), the control of the insulin injection device is a predictive control based on a simplified physiological model.
    According to one embodiment, in step c), the insulin injection device is controlled to deliver preprogrammed doses of insulin corresponding to a reference basal rate prescribed to the patient.
    According to one embodiment, the physiological model comprises a system of differential equations describing the evolution of a plurality of state variables as a function of time, and step a) of automatic calibration of the model comprises a step of estimation of parameters of the system of differential equations by minimization of a quantity representative of the error, during a past observation period, between the glycemia estimated from the physiological model and the glycemia measured by the sensor.
    According to one embodiment, step a) of automatic calibration of the model further comprises a step of determining initial values of the state variables.
    Another embodiment provides an automated regulation process for a patient's blood sugar, comprising:
    a step of calculating, by means of a processing and control unit, a prediction of the future evolution of the patient's glycaemia from a physiological model; and a step of controlling an insulin injection device taking account of this prediction, this method further comprising:
    a) an automatic calibration step of the physiological model taking into account a history of blood sugar measured by a blood sugar sensor during a past observation period;
    b) at the end of the calibration step, a step of determining the quality of the physiological model from at least one digital indicator representative of the error between the
    B15018 - DD16959LP blood sugar estimated from the model and the real blood sugar measured by the sensor; and
    c) if the quality of the model is deemed unsatisfactory, a step of ordering the insulin injection device without taking into account the prediction made from the model.
    Brief description of the drawings
    These characteristics and advantages, as well as others, will be explained in detail in the following description of particular embodiments made without implied limitation in relation to the attached figures among which:
    FIG. 1 schematically represents, in the form of blocks, an example of an embodiment of an automated system for regulating the glycemia of a patient;
    FIG. 2 is a simplified representation of a physiological model used in the system of FIG. 1 to predict the future evolution of the patient's blood sugar level;
    Figure 3 is a diagram showing in more detail an embodiment of the physiological model of Figure 2;
    FIG. 4 is a diagram illustrating an example of an automated method for regulating glycemia implemented by the system of FIG. 1;
    FIG. 5 is a diagram illustrating an example of an embodiment of an automated calibration method implemented by the system of FIG. 1; and FIG. 6 is a diagram illustrating an example of an embodiment of an automated method for regulating glycemia implemented by the system of FIG. 1.
    detailed description
    The same elements have been designated by the same references in the different figures. For the sake of clarity, only the elements which are useful for understanding the embodiments described have been shown and are detailed. In particular, the glycemia measurement device and the insulin injection device of the regulatory system described have not been detailed, the embodiments described being compatible
    B15018 - DD16959LP with all or most known blood glucose and insulin injection devices. In addition, the material implementation of the processing and control unit of the described regulation system has not been detailed, the realization of such a processing and control unit being within the reach of the skilled person. from the functional indications described.
    FIG. 1 schematically represents, in the form of blocks, an example of an embodiment of an automated system for regulating the glycemia of a patient.
    The system of Figure 1 includes a sensor 101 (CG) adapted to measure the patient's blood sugar. In normal operation, the sensor 101 can be permanently positioned on or in the patient's body, for example at the level of his abdomen. The sensor 101 is for example a CGM type sensor (from the English let's continue glucose monitoring), that is to say a sensor adapted to measure continuously (for example at least once per minute) the glycemia of the patient. The sensor 101 is for example a subcutaneous glycemia sensor.
    The system of FIG. 1 further comprises an insulin injection device 103 (PMP), for example a subcutaneous injection device. The device 103 is for example an automatic injection device of the insulin pump type, comprising an insulin reservoir connected to an injection needle implanted under the patient's skin, the pump being able to be electrically controlled to automatically inject doses of insulin determined at specified times. In normal operation, the injection device 103 can be permanently positioned in or on the patient's body, for example at the level of his abdomen.
    The system of FIG. 1 further comprises a processing and control unit 105 (CTRL) connected on the one hand to the blood glucose sensor 101, for example by wired link or by radio link (wireless), and on the other hand to the injection device 103, for example by wire or radio link. In operation, the processing and control unit 105 is
    B15018 - DD16959LP adapted to receive the patient's blood sugar data measured by the sensor 101, and to electrically control the device 103 to inject the patient with determined doses of insulin at determined times. In this example, the processing and control unit 105 is further adapted to receive, via a non-detailed user interface, data cho (t) representative of the evolution, as a function of time, of the amount of glucose ingested by the patient.
    The processing and control unit 105 is adapted to determine the doses of insulin to be injected into the patient, taking into account in particular the history of blood sugar measured by the sensor 101, the history of insulin injected by the device 103 , and the patient’s glucose intake history. For this, the processing and control unit 105 comprises a digital calculation circuit (not detailed), comprising for example a microprocessor. The processing and control unit 105 is for example a mobile device carried by the patient throughout the day and / or night, for example a device of the smartphone type configured to implement a method of regulation of the type described below.
    In the embodiment of FIG. 1, the processing and control unit 105 is adapted to determine the quantity of insulin to be administered to the patient by taking into account a prediction of the future evolution of his glycemia as a function of the time. More particularly, the processing and control unit 105 is adapted, on the basis of the history of insulin injected and the history of glucose ingested, and based on a physiological model describing the assimilation of insulin by the patient's body and its impact on blood sugar, to determine a curve representative of the expected evolution of the patient's blood sugar as a function of time, over a coming period, for example a period of 1 to 10 hours. By taking this curve into account, the processing and control unit 105 determines the doses of insulin to be injected into the patient during the coming period, so that the actual blood sugar (as opposed to the blood sugar estimated from the
    B15018 - DD16959LP physiological model) of the patient remains within acceptable limits, and in particular to limit the risks of hyperglycemia or hypoglycemia. In this operating mode, as will be explained in more detail below, the actual blood glucose data measured by the sensor 101 are used mainly for the purpose of calibrating the physiological model.
    Figure 2 is a simplified representation of a physiological MPC model used in the system of Figure 1 to predict future changes in the patient's blood sugar. In FIG. 2, the model is represented in the form of a processing block comprising:
    an input el to which is applied a signal i (t) representative of the evolution, as a function of time t, of the quantity of insulin injected into the patient;
    an input e2 to which is applied a signal cho (t) representative of the evolution, as a function of time t, of the amount of glucose ingested by the patient; and an output s providing a signal G (t) representative of the evolution, as a function of time t, of the patient's blood sugar level.
    The physiological model MPC is a compartmental model comprising, in addition to the input variables i (t) and cho (t) and the output variable G (t), a plurality of state variables corresponding to physiological variables of the patient, evolving over time. The temporal evolution of the state variables is governed by a system of differential equations comprising a plurality of parameters represented in FIG. 2 by a vector [PARAM] applied to an input pl of the block MPC. The response of the physiological model is further conditioned by the initial states or initial values assigned to the state variables, represented in FIG. 2 by a vector [INIT] applied to an input p2 of the block MPC.
    FIG. 3 is a diagram representing in more detail an example (nonlimiting) of the physiological model MPC used in the system of FIG. 1 to predict the future evolution of the patient's blood sugar level. This example model, known as
    B15018 - DD16959LP of the Hovorka model, is described in more detail in the article entitled Nonlinear model predictive control of glucose concentration in subjects with type 1 diabetes of Roman Hovorka et al. (Physiol Meas. 2004; 25: 905-920), and in the article titled Partitioning glucose distribution / transport, disposai, and endogenous production during IVGTT, by Roman Hovorka et al. (Am J Physiol Endocrinol Metab 282: E992-E1007, 2002).
    The physiological model of FIG. 3 comprises a first two-compartment sub-model 301 describing the effect of a food intake of glucose on the rate of appearance of glucose in the blood plasma. Sub-model 301 takes as input the quantity of glucose ingested cho (t), for example in mmol / min, and provides at its output a Ug absorption rate of glucose in the blood plasma, for example in mmol / min. Sub-model 301 includes two state variables Dg and Dg corresponding respectively to glucose masses, for example in mmol, in first and second compartments.
    The model of FIG. 3 also comprises a second bi-compartmental sub-model 303 describing the absorption, in the blood plasma, of the insulin administered to the patient. The submodel 303 takes as input the quantity of insulin i (t) injected into the patient, for example in mU / min, and provides at its output a rate Ug of absorption of insulin in the blood plasma, for example in mU / min. Sub-model 303 includes two state variables Sg and Sg corresponding respectively to insulin masses, for example in mmol, in first and second compartments.
    The model of Figure 3 further includes a third sub-model 305 describing the regulation of glucose by the patient's body. The sub-model 305 takes as input the absorption rates Ug of glucose and Ug of insulin, and supplies at its output the glycemia G (t), that is to say the glucose concentration in the blood plasma. , for example in mmol / 1. The sub-model 305 includes six state variables Qg, Qg, X3, xg, xg, I. The variables Q1 and Q2 correspond respectively to masses of glucose, for example in mmol, in first and second compartments.
    B15018 - DD16959LP
    The variables X] _, X2, X3 are unitless variables representing each of the actions of insulin on the kinetics of glucose. The variable I corresponds to insulinemia, that is to say the concentration of insulin in the blood plasma, for example in mU / 1.
    The Hovorka model is governed by the following system of equations:
    G (t) = dQi _ dt dQ 2 dt dx ± dt dx 2 dt dx 3 dt dS r dt dS 2 dt dl dt dD ± dt dD 2 dt
    Qi (0
    Vr.
    + * l (t)
    Lk G -G (t) * l (t) · Ql (0 - [fcl2 + ^ 2 (0] · Q 2 (0 = -k br · x / t) + k al · / (t) dx 2 - = -k b2 · x 2 (t) + k a2 -I (t) dx 3 - = -k b3 · x 3 (t) + k a3 -I (t) dSi - = i (t) - k a · S 1 (t) = k a S 1 (t) -k a -S 2 (t) k a -S 2 (t)
    V, ~ k e -I (t) = cho (t) -
    Dft)
    Di (t) P 2 (t) t-max t max
    Ur. =
    D 2 (t)
    Qi (t) + k 12 Q 2 (t) - F R + EGP 0 · [1 - x 3 (t)] + U G (t)
    With:
    , c F 01 · G (t) 01 0.85 · (G (t) + 1.0)
    B15018 - DD16959LP
    F = i R ( G ~ ^ ' v c if G> 9 R l 0 otherwise
    In this system of equations, the quantities Vq, Fqi, k] _2, Fr, EGPq, kpi, k a ] _, kb2, k a 2, kp3, k a 3, k a , Vj, k e and t max are parameters. Vq corresponds to the volume of distribution of glucose, for example in liters, Fqi corresponds to a non-insulin-dependent glucose transfer rate, for example in mmol / min, k] _2 corresponds to a transfer rate constant between the two compartments of the sub-model 305, for example in min _ l, ka] _, ka2, ka3 correspond to constants of insulin deactivation rate, for example in min _ l, Fr corresponds to a urinary excretion of glucose, for example in mmol / min, EGPq corresponds to an endogenous production of glucose, for example in min _ l, kp] _, kp2 and kp3 correspond to constants of insulin activation rate, for example in min _ l, ka corresponds to a constant rate of absorption of insulin injected subcutaneously, for example in min _ l, Vj corresponds to the volume of distribution of insulin, for example in liters, ke corresponds to a rate of elimination of l plasma insulin, for example in min _ l, and tmax corresponds to a time spent up to the peak of ab sorption of glucose ingested by the patient, for example in min. These fifteen parameters correspond to the vector [PARAM] of the representation of FIG. 2. The vector [INIT] comprises ten values corresponding to the initial values (at an instant tQ at the start of a phase of simulation of the patient's behavior at from the model) assigned to the ten state variables D] _, D2, Si, S2, Qi, Ç> 2 / x % x 2> X 1 I of the model.
    Among the parameters of the vector [PARAM], some can be considered as constant for a given patient. These are for example the parameters ki2, k a i, k a 2, k a 3, k a , k e , Vj, Vq and t max . Other parameters, hereinafter called time-dependent parameters, are however liable to evolve over time, for example the parameters kpi, kp2, kp3, EGPq, Fqi and Fr. Because of this variability of certain parameters of the system, it is in practice necessary to recalibrate or re3056094
    B15018 - DD16959LP regularly calibrate the model in use, for example every 1 to 20 minutes, to ensure that the predictions of the model remain relevant. This updating of the model, also called customization of the model, must be able to be carried out automatically by the system of FIG. 1, that is to say without it being necessary to physically measure the time-dependent parameters of the system on the patient and then transmit them to the treatment and control unit 105.
    FIG. 4 is a diagram illustrating an example of an automated method for regulating glycemia implemented by the system of FIG. 1.
    This method comprises a step 401 of recalibration or updating of the model, which can for example be repeated at regular intervals, for example every 1 to 20 minutes. During this step, the processing and control unit 105 implements a method for re-estimating the time-dependent parameters of the model, taking into account the insulin data actually injected by the device 103 and the actual blood sugar data. measured by the sensor 101 during a past observation period, for example a period of 1 to 10 hours preceding the calibration step. More particularly, during the calibration step, the processing and control unit 105 simulates the behavior of the patient over the past observation period from the physiological model (taking into account any glucose ingestion and injections of insulin during this period), and compares the glycemia curve estimated by the model with the actual glycemia curve measured by the sensor during this same period. The processing and control unit 105 then searches, for the time-dependent parameters of the model, a set of values leading to minimizing a quantity representative of the error between the glycemia curve estimated by the model and the actual glycemia curve. during the observation period. For example, the processing and control unit searches for a set of parameters leading to minimizing an indicator m representative of the area between the glycemia curve
    B15018 - DD16959LP estimated by the model and the actual blood sugar curve during the observation period, for example defined as follows:
    to + ΔΤ m = Â Σ - ^ i 2 t = t 0 where t is the discretized time variable, to corresponds to the start time of the past observation phase, to + 4T corresponds to the end time of the past observation phase (corresponding for example to the start time of the model calibration step), g is the time evolution curve of the real glycemia measured by the sensor 101 during the period [tO, to + 4T], and g is the glycemia curve estimated from the model during the period [to, to + 4T]. The algorithm for finding optimal parameters used during this step is not detailed in the present application, the described embodiments being compatible with the usual algorithms used in various fields to solve problems of optimization of parameters by minimization of a cost function.
    The method of FIG. 4 further comprises, after step 401, a step 403 of predicting, by the processing and control unit 105, the temporal evolution of the patient's glycemia over a coming period, starting from the physiological model updated in step 401 and taking into account the history of insulin injected into the patient and the history of glucose ingested by the patient.
    The method of FIG. 4 further comprises, after step 403, a step 405 of determination, by the processing and control unit 105, taking into account the future glycemia curve predicted in step 403, doses of insulin to inject into the patient for a future period. At the end of this step, the processing and control unit 105 can program the injection device 103 to administer the doses determined during the coming period.
    B15018 - DD16959LP
    The steps 403 of predicting the glycemia and 405 and determining the future doses of insulin to be administered can for example be repeated at each update of the physiological model (that is to say after each iteration of the step 401) , with each new ingestion of glucose reported by the patient, and / or with each new administration of a dose of insulin by the injection device 103.
    A problem which arises in the operation described above is that, when the physiological model is updated in step 401, the processing and control unit 105 must define a vector [INIT] of initial states (states to îq) of the state variables of the model, in order to be able to simulate the behavior of the patient from the model. These initial states are necessary not only to be able to predict the future evolution of the patient's blood sugar level (step 403), but also during the step of updating the model itself (step 401), to be able to simulate the evolution of the patient's blood sugar during the past observation period, so that the simulated blood sugar can be compared with the measured blood sugar.
    To define the initial states of the state variables of the model, a first possibility consists in assuming that, in the period preceding the observation period [tO, to + AT] on which the calibration of the model is based, the patient was in a stationary state, with a constant flow of insulin injected, and no food intake of glucose. Under this hypothesis, all the derivatives of the system of differential equations can be considered as zero at the initial time to- The values at îq of the state variables of the system can then be computed analytically. A disadvantage of this solution is that the output of the model (the estimated blood sugar) is not constrained. In particular, the glycemia estimated at instant îq may be different from the actual glycemia measured at instant îq. In this case, the algorithm used in step 401 to search for time-dependent parameters of the model by
    B15018 - DD16959LP minimizing the error between the simulated blood glucose and the measured blood sugar may have trouble converging.
    To improve the initialization, a second possibility consists in making the same assumptions as previously, but by constraining the variable Q] _ (tg) so that the glycemia estimated at the instant îq is equal to the actual glycemia measured by the sensor. . This improves the relevance of the initialization at time tQ. However, at time îq, the derivative of the estimated blood sugar and the derivative of the actual blood sugar may diverge. Consequently, the algorithm for finding time-dependent parameters of the system may again have difficulty in converging.
    In practice, the two aforementioned methods of determining the initial states of the physiological model are often unsatisfactory, which makes it difficult to find a set of values relevant to the time-dependent parameters of the model. One consequence is that the predictions of the future evolution of the patient's blood sugar from the model may be wrong, and lead to poor regulation of blood sugar by the system.
    To overcome this problem, according to one aspect of an embodiment, provision is made, during the calibration or updating phase of the model (step 401), to consider the initial states [INIT] of the model as random variables. , and to carry out, as is done to estimate the time-dependent parameters of the model, a search for an optimal set of initial state values by minimizing a quantity representative of the error between the estimated glycemia curve from the model and the actual blood sugar curve during the observation period on which the calibration is based.
    If the cumulative number of time-dependent parameters and state variables of the physiological model is sufficiently low, the optimal values of the time-dependent parameters and initial states of the state variables can be determined simultaneously, in the same step model optimization by
    B15018 - DD16959LP minimization of the error between the estimated blood sugar and the real blood sugar over the past observation period.
    In practice, in the Hovorka model, as well as in most physiological models describing the assimilation of insulin and glucose by the body and their impact on blood sugar, the cumulative number of time-dependent parameters and variables d states are relatively large, which can lead to numerical instability during the phase of search for optimal values. In other words, some values may be difficult or even impossible to estimate in a single search, the number of unknowns being too large. In this case, the problem can be broken down into two sub-problems, corresponding respectively to the estimation of the time-dependent parameters of the model and to the estimation of the initial states of the model, as will now be described in relation to the figure. 5.
    FIG. 5 is a diagram illustrating an example of an embodiment of an automated method for calibrating or updating the system of FIG. 1, corresponding to an example of implementation of step 401 of FIG. 4.
    This method comprises a step 501 during which the vector of parameters [PARAM] (here reduced to only time-dependent parameters of the model) is initialized to a first set of values PI. The PI set corresponds for example to the values taken by the parameters [PARAM] before the start of the model update phase. As a variant, the set of values PI is a predetermined reference set corresponding for example to the average values taken by the parameters [PARAM] over a reference period. During step 501, the vector of initial states [INIT] of the state variables is also initialized to a first set of value II. The set of values II is for example determined analytically as described above, by assuming a stationary state of the patient in the period preceding the calibration phase, and by making coincide the glycemia estimated at the instant îq and the actual blood sugar measured at that same time.
    B15018 - DD16959LP
    During a step 503 subsequent to step 501, the processing and control unit 105 searches, by fixing the set of initial states [INIT] at its current state, for a set of values of the time-dependent parameters of the model leading to minimize a quantity representative of the error between the glycemia curve estimated from the model and the actual glycemia curve during the observation period, for example the indicator m defines above. At the end of this step, the vector [PARAM] is updated with the new estimated values.
    During a step 505 subsequent to step 503, the processing and control unit 105 searches, by fixing the set of parameters [PARAM] at its current state, for a set of values of initial states of the variables of state leading to minimize a quantity representative of the error between the glycemia curve estimated from the model and the actual glycemia curve during the observation period, for example the indicator m defined above, or any other representative indicator the error between the two curves, for example an indicator based on the L1 standard. At the end of this step, the vector [INIT] is updated with the new estimated values.
    In this example, steps 503 and 505 are repeated a predetermined number N of times, where N is an integer greater than
    1. The values of the time-dependent parameters s and of the initial states of the updated model then correspond to the values of the vectors [PARAM] and [INIT] at the end of the Nth iteration of steps 503 and 505. As a variant the number of iterations of steps 503 and 505 may not be predetermined, and may be adjusted taking into account the evolution of the indicator of error m between the blood sugar estimated from the model and the real blood sugar over the period d 'observation.
    The algorithms for finding optimal values used during steps 503 and 505 are not detailed in the present application, the embodiments described being compatible with the usual algorithms used in fields
    B15018 - DD16959LP various to solve problems of optimization of parameters by minimization of a cost function.
    An advantage of the operating mode described above, in which the initial values of the state variables of the physiological model are determined by minimizing a quantity representative of the error between the measured blood glucose data and the estimated blood sugar during a period observation, is that it improves the quality of the prediction of the patient's future blood sugar level, and thus controls insulin intake with greater relevance.
    Another object of another embodiment is to limit the risks for the patient linked to a possible failure of the physiological model used to predict the patient's future blood sugar.
    For this, according to one aspect of an embodiment, the control and processing device 105 of the regulation system is adapted, after each update or recalibration of the physiological model (step 401), to estimate the quality of the physiological model updated using one or more digital quality indicators, and, if the quality of the model is deemed unsatisfactory, to stop using the model to regulate the patient's blood sugar.
    FIG. 6 is a diagram illustrating an example of an embodiment of an automated method for regulating glycemia implemented by the system of FIG. 1.
    This method comprises the same steps 401, 403 and 405 as in the example in FIG. 4. However, the method in FIG. 6 further comprises, after each step 401 of updating the physiological model used by the regulation system and before the implementation of the following steps 403 of predicting the patient's future blood sugar from the model and 405 of controlling the insulin delivery from the blood sugar prediction, a step 601 of checking the quality of the model up to date.
    During step 601, the processing and control unit 105 determines one or more digital indicators of
    B15018 - DD16959LP the quality of the model updated in step 401. As an example, the processing and control unit calculates a digital quality indicator representative of the area between the glycemia curve estimated from the model and the actual blood sugar curve measured by sensor 101 during a past observation period. This indicator corresponds for example to the quantity m defined above.
    In place of, or in addition to, an indicator representative of the surface between the curves of estimated blood sugar and real blood sugar during a past observation period, the processing and control unit 105 can calculate one and / either of the following mg and mg quality indicators:
    m l (tcourant) = d ^ current) -g (tcouranp) m 2 (tcouranp) = g '(t couran p) -g' (t couran p), where t couran - (- indicates a present instant of setting work of step 601 of verifying the quality of the model, g corresponds to the function of temporal evolution of the real glycemia measured by the sensor 101, g corresponds to the function of temporal evolution of the glycemia simulated from the model , g 'corresponds to the derivative of the temporal evolution function of the real glycemia, and g' corresponds to the derivative of the temporal evolution function of the simulated glycemia.
    For example, the quality of the model can be considered satisfactory by the processing and control unit 105 when the values m, mg and mg are below predefined thresholds. More generally, any other quality criterion or any other combination of quality criteria can be used in step 601 to determine whether the physiological model re-calibrated in step 401 can be considered reliable.
    If the physiological model is considered to be reliable in step 601 (O), steps 403 and 405 can be implemented in a similar manner to what has been described previously, that is to say that the unit of processing and control 105 continues to be based on the predictions made by the model
    B15018 - Physiological DD16959LP to regulate the administration of insulin to the patient.
    If the physiological model is judged to be insufficiently reliable in step 601 (N), the processing and control unit 105 stops using this model to regulate the administration of insulin to the patient, and implements a method of substitution regulation during a step 603.
    By way of example, during step 603, the processing and control unit 105 uses a simplified physiological model, for example a compartmental model comprising a number of state variables and a reduced number of parameters compared to the initial model, to predict the evolution of the patient's blood sugar and regulate insulin injection accordingly.
    As a variant, during step 603, the processing and control unit 105 ceases to implement a predictive command, that is to say that it ceases to use a physiological model to predict the patient's future blood sugar and regulate insulin injection accordingly. In this case, the processing and control unit 105 controls, for example, the insulin injection device 103 to administer pre-programmed doses of insulin, corresponding for example to a basal reference flow prescribed to the patient.
    Such a substitution method can for example be used for a predetermined period of time. At the end of this period, the steps 401 of calibrating the main physiological model and 601 of estimating the quality of the main physiological model can be repeated, so as to, if the quality of the main physiological model is deemed satisfactory, reactivate the use of this model to regulate the administration of insulin to the patient.
    It will be noted that the method of FIG. 6 is not limited to the embodiment described in relation to FIGS. 4 and 5, in which the calibration of the physiological model comprises a step of determining the initial values of the state variables of the model by minimization of a representative quantity
    B15018 - DD16959LP of the error between the measured blood glucose data and the estimated blood sugar during an observation period, but can be used regardless of the method chosen to determine the initial values of the state variables of the model.
    Particular embodiments have been described.
    Various variants and modifications will appear to those skilled in the art. In particular, the embodiments described are not limited to the particular example of a physiological model detailed in the present description, namely the model of
    Hovorka, but are compatible with any physiological model describing the assimilation of insulin by the body of a patient and its effect on the patient's blood sugar, for example the so-called Cobelli model, described in the article entitled A System Model of Oral Glucose Absorption: Validation on Gold Standard Data, of
    Chiara Dalla Man et al. (IEEE TRANSACTIONS ON BIOMEDICAL ENGINEERING, VOL. 53, NO. 12, DECEMBER 2006).
    B15018 - DD16959LP
    权利要求:
    Claims (12)
    [1" id="c-fr-0001]
    1. Automated system for regulating a patient's blood sugar level, comprising:
    a blood glucose sensor (101);
    an insulin injection device (103); and a processing and control unit (105) adapted to predict the future evolution of the patient's blood sugar from a physiological model and to control the insulin injection device (103) taking account of this prediction , in which :
    the physiological model includes a system of differential equations describing the evolution of a plurality of state variables (Dg, Dg, Sg, Sg, Qg, Qg, x l, x 2 < x 3> I) as a function of time ; and the processing and control unit (105) is adapted to implement a step of automatic calibration of the physiological model comprising a step of estimating initial values ([INIT]) of the state variables by minimization of a quantity representative of the error, during a past observation period, between the glycemia estimated from the physiological model and the glycemia measured by the sensor (101).
    [2" id="c-fr-0002]
    2. The system as claimed in claim 1, in which said quantity is representative of the area between a first curve g representative of the temporal evolution of the glycemia estimated from the model over the observation period, and a second curve g representative the temporal evolution of the glycemia measured by the sensor (101) over the observation period.
    [3" id="c-fr-0003]
    3. The system as claimed in claim 2, in which said quantity is defined as follows:
    to + ΔΤ
    m = Âr / East or t = t 0 t is a discretized time variable, îq The moment of start of the observation phase, and to + 4T East The moment of end of the observation phase.
    B15018 - DD16959LP
    [4" id="c-fr-0004]
    4. System according to any one of claims 1 to 3, wherein the calibration method further comprises a step of estimating parameters ([PARAM]) of the system of differential equations by minimizing said quantity.
    [5" id="c-fr-0005]
    5. System according to claim 4, in which the calibration method comprises a plurality of successive iterations of the following steps a) and b):
    a) estimate the parameters ([PARAM]) of the system of differential equations by minimizing said quantity by fixing the initial values ([INIT]) of the state variables; and
    b) estimate the initial values ([INIT]) of the state variables by minimizing said quantity by fixing the parameters ([PARAM]) of the system of differential equations.
    [6" id="c-fr-0006]
    6. The system as claimed in claim 5, in which, at the first iteration of step a), the initial values (INIT]) of the state variables are determined analytically by assuming that all the derivatives of the system differential equations are zero.
    [7" id="c-fr-0007]
    7. System according to any one of claims 1 to 6, in which, to simulate the evolution of the patient's blood sugar from the physiological model, the processing and control unit (105) takes account of the history insulin (i (t)) injected into the patient by the injection device (103) and the history of glucose (cho (t)) ingested by the patient.
    [8" id="c-fr-0008]
    8. System according to any one of claims 1 to 7, in which the physiological model is the Hovorka model.
    [9" id="c-fr-0009]
    9. Method for automated regulation of a patient's blood sugar level, comprising:
    a step (403) of calculation, by means of a processing and control unit (105), of a prediction of the future evolution of the patient's glycemia from a physiological model comprising a system of equations differentials describing the evolution of a plurality of state variables (Dg, Dg, S] _, Sg, Qg, Qg, x l, x 2 ' x 3> 7) as a function of time;
    B15018 - DD16959LP a step (405) of controlling an insulin injection device (103) taking into account this prediction; and a step (401) of automatic calibration of the physiological model comprising a step of estimating values
    5 initials ([INIT]) of the state variables by minimizing a quantity representative of the error, during a past observation period, between the glycemia estimated from the physiological model and the glycemia measured on the patient by a blood sugar sensor (101).
    [10" id="c-fr-0010]
    10. Method according to claim 9, further comprising a step of estimating parameters ([PARAM]) of the system of differential equations by minimizing said quantity.
    [11" id="c-fr-0011]
    11. The method of claim 9 or 10, wherein the calibration step comprises a plurality of iterations
    [12" id="c-fr-0012]
    15 successive steps a) and b) below:
    a) estimate the parameters ([PARAM]) of the system of differential equations by minimizing said quantity by fixing the initial values ([INIT]) of the state variables; and
    b) estimating the initial values ([INIT]) of the state variables by minimizing said quantity by fixing the parameters ([PARAM]) of the system of differential equations.
    B15018
    DD16959LP
    1/3
    101 cho (t)
    类似技术:
    公开号 | 公开日 | 专利标题
    FR3056094A1|2018-03-23|AUTOMATED SYSTEM FOR REGULATING THE GLYCEMIA OF A PATIENT
    FR3056095A1|2018-03-23|AUTOMATED SYSTEM FOR REGULATING THE GLYCEMIA OF A PATIENT
    FR2557445A1|1985-07-05|METHOD AND DEVICE FOR THE AUTOMATIC PROSPECTIVE DETERMINATION OF INDIVIDUALLY SPECIFIC PARAMETERS OF GLUCOSE REGULATION
    EP1881786A2|2008-01-30|Fully automated control system for type 1 diabetes
    CA3070296A1|2019-01-24|Automated system for regulating a patient&#39;s blood glucose level
    EP3768166B1|2022-02-09|System for predicting a patient&#39;s blood glucose level
    FR3083076A1|2020-01-03|AUTOMATED SYSTEM FOR CONTROLLING A PATIENT&#39;S GLYCEMIA
    EP3797426A1|2021-03-31|Automated system for monitoring a patient&#39;s blood sugar
    FR3081315A1|2019-11-29|AUTOMATED SYSTEM FOR CONTROLLING THE GLYCEMIA OF A PATIENT
    EP3899968A1|2021-10-27|Automated system for regulating the blood glucose level of a patient
    US20200402634A1|2020-12-24|Dynamic equivalent on board estimator
    FR3099043A1|2021-01-29|Automated blood sugar control system
    CA3071570A1|2019-02-07|Closed-loop blood glucose control systems and methods
    FR3103372A1|2021-05-28|Automated blood sugar control system
    FR2942387A1|2010-08-27|SYSTEM FOR REGULATING BLOOD GLUCOSE WITH LIVING DIABETIC SUBJECT AND CORRESPONDING COMPUTER PROGRAM PRODUCT.
    同族专利:
    公开号 | 公开日
    FR3056094B1|2018-10-12|
    BR112019005520A2|2019-06-18|
    JP2019534066A|2019-11-28|
    WO2018055283A1|2018-03-29|
    KR20190050834A|2019-05-13|
    US20190298918A1|2019-10-03|
    CA3037655A1|2018-03-29|
    EP3515307A1|2019-07-31|
    引用文献:
    公开号 | 申请日 | 公开日 | 申请人 | 专利标题
    GB2436873A|2006-04-07|2007-10-10|Univ Cambridge Tech|Blood glucose monitoring systems|
    US20140066884A1|2012-08-30|2014-03-06|Medtronic Minimed, Inc.|Sensor model supervisor for a closed-loop insulin infusion system|
    EP2583715A1|2011-10-19|2013-04-24|Unomedical A/S|Infusion tube system and method for manufacture|
    FR3083076A1|2018-06-29|2020-01-03|Commissariat A L'energie Atomique Et Aux Energies Alternatives|AUTOMATED SYSTEM FOR CONTROLLING A PATIENT'S GLYCEMIA|
    FR3090315A1|2018-12-21|2020-06-26|Commissariat A L'energie Atomique Et Aux Energies Alternatives|Automated system for regulating a patient's blood sugar|
    FR3099043A1|2019-07-25|2021-01-29|Commissariat A L'energie Atomique Et Aux Energies Alternatives|Automated blood sugar control system|
    FR3103372A1|2019-11-27|2021-05-28|Commissariat A L'energie Atomique Et Aux Energies Alternatives|Automated blood sugar control system|
    法律状态:
    2017-09-29| PLFP| Fee payment|Year of fee payment: 2 |
    2018-03-23| PLSC| Search report ready|Effective date: 20180323 |
    2018-09-28| PLFP| Fee payment|Year of fee payment: 3 |
    2019-09-30| PLFP| Fee payment|Year of fee payment: 4 |
    2020-09-30| PLFP| Fee payment|Year of fee payment: 5 |
    2021-09-30| PLFP| Fee payment|Year of fee payment: 6 |
    优先权:
    申请号 | 申请日 | 专利标题
    FR1658881A|FR3056094B1|2016-09-21|2016-09-21|AUTOMATED SYSTEM FOR REGULATING THE GLYCEMIA OF A PATIENT|
    FR1658881|2016-09-21|FR1658881A| FR3056094B1|2016-09-21|2016-09-21|AUTOMATED SYSTEM FOR REGULATING THE GLYCEMIA OF A PATIENT|
    BR112019005520A| BR112019005520A2|2016-09-21|2017-09-19|automated patient blood glucose regulation system|
    PCT/FR2017/052511| WO2018055283A1|2016-09-21|2017-09-19|Automated system for controlling the blood glucose level of a patient|
    US16/334,825| US20190298918A1|2016-09-21|2017-09-19|Automated system for controlling the blood glucose level of a patient|
    CA3037655A| CA3037655A1|2016-09-21|2017-09-19|Automated system for controlling the blood glucose level of a patient|
    JP2019515487A| JP2019534066A|2016-09-21|2017-09-19|Automated system for controlling a patient's blood glucose level|
    KR1020197010866A| KR20190050834A|2016-09-21|2017-09-19|Automatic system for blood glucose control of patient|
    EP17783918.0A| EP3515307A1|2016-09-21|2017-09-19|Automated system for controlling the blood glucose level of a patient|
    [返回顶部]