Optimizing embedded formulations for drug delivery

Abstract

To provide a method implemented by a drug delivery device to determine an optimal dose of a liquid drug for current cycle of a medication delivery algorithm.SOLUTION: A method disclosed herein utilizes a stepwise evaluation of a model and a cost function across a coarse search space consisting of coarse discrete quantities of a drug and a refined search space consisting of refined discrete quantities of the drug.SELECTED DRAWING: Figure 1

Description

[Technical Field]

[0001] Related Applications This application is a continuation of U.S. Patent Application No. 17 / 539,270, filed December 1, 2021, the contents of which are incorporated herein by reference in their entirety. [Background technology]

[0002] Many conventional automated drug delivery systems are well known and include, for example, wearable drug delivery devices of the type shown in FIG. 2. The drug delivery device 102 can be designed to deliver any type of liquid medication to a user. In certain embodiments, the drug delivery device 102 can be, for example, an OmniPod® drug delivery device manufactured by Insulet Corporation of Acton, Massachusetts. The drug delivery device 102 can be a drug delivery device such as those described in U.S. Pat. No. 7,303,549, U.S. Pat. No. 7,137,964, and U.S. Pat. No. 6,740,059, the contents of each of which are incorporated herein by reference in their entirety.

[0003] Wearable drug delivery devices 102 are typically configured with a processor and memory and are often powered by an internal battery or power harvesting device that has a limited amount of power available to power the processor and memory. Furthermore, due to the size of the device, processing power and memory for storing software algorithms may also be limited. Due to these limitations, wearable drug delivery devices do not have on-board drug delivery algorithms that determine the timing and amount of liquid medication to be delivered to a user through a series of calculations based on feedback from sensors and other information. Such drug delivery applications are typically found on remote devices, such as remote personal diabetes management (PDM) devices or smartphones, either of which are configured to transmit drug delivery instructions.

[0004] The medication delivery algorithm may use an optimization algorithm to periodically calculate the amount of liquid medication to be delivered to the user. For example, in one embodiment, the medication delivery algorithm may operate on a five-minute cycle. The optimization algorithm may utilize a mathematical glucose model and minimize a cost function to determine the appropriate amount of liquid medication to be delivered. Such optimization algorithms often require a series of complex calculations with high computational and power consumption costs, making them difficult to implement in applications with low power consumption and efficient processing power available, such as embedded applications. This is particularly important when it is desired to implement such optimization algorithms in disposable, small-scale electronic devices, such as the wearable medication delivery device 102.

[0005] Therefore, it would be desirable to provide a method that reduces the computational cost and power consumption for running optimization algorithms, allows drug delivery algorithms to reside on-board the wearable drug delivery device 102, and improves the lifespan of the wearable drug delivery device 102 and the speed at which appropriate drug dosages can be calculated for delivery of drugs to a user or wearer of the wearable drug delivery device 102.

[0006] definition As used herein, the term "liquid drug" should be taken to include any drug in liquid form that can be administered by a drug delivery device via a subcutaneous cannula, including, for example, insulin, GLP-1, pramlintide, morphine, blood pressure medications, chemotherapy drugs, infertility medications, etc., or co-formulations of two or more of GLP-1, pramlintide, and insulin. Summary of the Invention [Means for solving the problem]

[0007] This Summary is provided to introduce a simplified selection of concepts that are discussed below in the Detailed Description. This Summary is not intended to identify key features or essential features of the claimed subject matter, nor is it intended as an aid in determining the scope of the claimed subject matter.

[0008] In certain embedded medical implementations, such as in the wearable drug delivery device 102, an exact solution to the optimization problem may not be required. This is largely due to the fact that the difference in the calculated dosage for each cycle between an exact solution to the optimization problem and an approximate solution to the optimization problem may be less than or equal to the minimum drug delivery resolution of the wearable drug delivery device 102. For example, in certain embodiments, the resolution of the delivery of a liquid drug may be limited to a certain amount (e.g., 0.05 U, or 0.0005 mL). That is, the wearable drug delivery device 102 may only be able to deliver certain discrete amounts of the liquid drug. Therefore, an exact solution to the optimization problem that provides a recommended dosage that falls between the certain discrete amounts cannot be delivered, and therefore, an approximation of the solution to the optimization problem is allowed. Furthermore, the difference between the calculated dosage and the dosage that the wearable drug delivery device 102 can deliver to the user may not result in a significant difference in the user's clinical outcome. In some embodiments, the drug delivery algorithm rounds the delivery to the nearest deliverable discrete amount and can add or subtract any difference to or from the calculated dose during the next cycle.

[0009] Therefore, the computational cost of running an optimization algorithm can be significantly reduced (99% or greater) by a simple implementation of the optimization algorithm that does not need to be accurate to the smallest decimal point, but allows the system to arrive at a solution that is close enough so as not to affect the user's treatment.

[0010] In the drawings, like reference characters generally refer to the same parts throughout the different views.In the following description, various embodiments of the present invention will be described with reference to the accompanying drawings. [Brief explanation of the drawings]

[0011] [Figure 1] FIG. 1 is a functional block diagram of an exemplary system suitable for implementing the systems and methods disclosed herein.

[0012] [Figure 2] FIG. 2 is a diagram of a prior art wearable drug delivery device of the type in which the invention disclosed herein may be used.

[0013] [Figure 3] FIG. 3 is a flow chart illustrating the steps comprising the disclosed method. DETAILED DESCRIPTION OF THE INVENTION

[0014] Detailed Description The present disclosure presents various systems, components, and methods for calculating the amount of liquid medication to be delivered to a user during the current execution cycle of a medication delivery algorithm. The embodiments described herein offer one or more advantages over conventional prior art systems, components, and methods, namely, reduced power consumption and required processing power due to simplified computational models used to solve the optimization problem.

[0015] Various embodiments of the present invention include systems and methods for delivering medication to a user using a drug delivery device (sometimes referred to herein as a "pod"), either autonomously or according to a wireless signal received from an electronic device. In various embodiments, the electronic device may be a user device including a smartphone, a smart watch, a smart necklace, a module attached to the drug delivery device, or any other type or kind of electronic device that can be carried by or worn on the user's body and that executes an algorithm to calculate the time and dosage of delivery of the medication.

[0016] For example, the user device may run an "artificial pancreas" algorithm that calculates the time and dosage of insulin delivery. The user device may also communicate with sensors, such as glucose sensors, that collect data about the user's physical attributes or conditions, such as glucose levels. The sensors are placed in or on the user's body and may be part of the drug delivery device or may be separate devices.

[0017] Alternatively, the drug delivery device can communicate with the sensor instead of, or in addition to, communication between the sensor and the user device. The communication can be direct (e.g., if the sensor is integrated with or otherwise part of the drug delivery device) or remote / wireless (e.g., if the sensor is located in a different housing than the drug delivery device). In these embodiments, the drug delivery device includes computing hardware (e.g., processor, memory, firmware, etc.) that executes some or all of the algorithms that calculate the time and dosage of medication delivery.

[0018] 1 is a functional block diagram of an exemplary drug delivery system 100 suitable for implementing the systems and methods described herein. Drug delivery system 100 can implement (and / or provide functionality for) a drug delivery algorithm, such as an artificial pancreas (AP) application, to govern or control the automated delivery of a drug or medication, such as insulin, to a user (e.g., to maintain euglycemia, which is a normal level of blood glucose). Drug delivery system 100 can be an automated drug delivery system including a drug delivery device 102 (which may be wearable), an analyte sensor 108 (which may also be wearable), and a user device 105.

[0019] The drug delivery system 100 may also include, in optional examples, an accessory device 106, such as a smart watch, personal assistant device, etc., which may communicate with other components of the system 100 via either wired or wireless communication links 191-193.

[0020] User Device The user device 105 may be a computing device such as a smartphone, tablet, personal diabetes management (PDM) device, dedicated diabetes care management device, etc. In one example, the user device 105 may include a processor 151, device memory 153, a user interface 158, and a communication interface 154. The user device 105 may also include analog and / or digital circuitry that may be implemented as the processor 151 for performing processing based on programming code stored in the device memory 153, such as a user application 160 for managing the user's blood glucose levels and controlling the delivery of drugs, medications, or therapeutic agents to the user, as described below, and for providing other functions such as calculating carbohydrate correction doses, correction bolus doses, etc. The user device 105 may be used to program, adjust settings, and / or control the operation of the drug delivery device 102 and / or analyte sensor 108 and optional smart accessory device 106.

[0021] The processor 151 may also be configured to execute programming code stored in the device memory 153, such as user apps 160. The user apps 160 may be computer applications operable to deliver medication based on information received from the analyte sensors 108, the cloud-based service 111, and / or the user device 105 or optional accessory devices 106. The memory 153 may also store programming code for operating, for example, a user interface 158 (e.g., a touchscreen device, a camera, etc.), a communication interface 154, or the like. When executing the user apps 160, the processor 151 may be configured to implement displays and notifications related to meal intake, blood glucose measurements, etc. The user interface 158 is under the control of the processor 151 and may be configured to present a graphical user interface that allows for input of meal announcements, adjustment of setting selections, etc., as described herein.

[0022] In a specific example, when the user app 160 is an AP application, the processor 151 is also configured to execute a diabetes treatment plan (which may be stored in memory) managed by the user app 160. In addition to the functions described above, when the user app 160 is an AP application, it may further provide functions to determine a carbohydrate compensation dose, a correction bolus dose, and a basal dose according to the diabetes treatment plan. Furthermore, as an AP application, the user app 160 provides functions to output signals to the drug delivery device 102 via the communication interface 154 to deliver the determined bolus and basal doses.

[0023] Communications interface 154 may include one or more transceivers operating according to one or more radio frequency protocols. In one embodiment, the transceivers may consist of a cellular transceiver and a Bluetooth transceiver. Communications interface 154 may be configured to receive and transmit signals containing information usable by user app 160.

[0024] The user device 105 may further include one or more output devices 155, for example, speakers or vibration transducers, for providing various signals to the user.

[0025] Drug Delivery Devices In various exemplary embodiments, the drug delivery device 102 may include a reservoir 124 and a drive mechanism 125 controllable by a controller 121 executing a drug delivery algorithm (MDA) 129 stored in a memory 123 onboard the drug delivery device (and in exemplary embodiments, a wearable drug delivery device). Alternatively, the controller 121 may operate to control the reservoir 124 and the drive mechanism 125 based on signals received from a user app 160 executing on the user device 105, communicated to the drug delivery device 102 via a communication link 194. The drive mechanism 125 operates to translate a plunger longitudinally through the reservoir to force the liquid medicament through the outlet fluid port and into the needle / cannula 186.

[0026] In another embodiment, the drug delivery device 102 may also include an optional second reservoir 124-2 and second drive mechanism 125-2 to enable independent delivery of two different liquid drugs. By way of example, the reservoir 124 may be filled with insulin, and the reservoir 124-2 may be filled with pramlintide or GLP-1. In some embodiments, each of the reservoirs 124, 124-2 may be configured with a separate drive mechanism 125, 125-2, respectively, which may be separately controllable by the controller 121 under the direction of the MDA 129. Both reservoirs 124, 124-2 may be connected to a common needle / cannula 186.

[0027] The drug delivery device 102 may optionally be configured with a user interface 127 that provides means for receiving input from a user and for outputting information to a user. The user interface 127 may include, for example, a light emitting diode, a button on the housing of the drug delivery device 102, a sound transducer, a microdisplay, a microphone, an accelerometer for detecting device movement or a user gesture (e.g., tapping on the housing of the device), or any other type of interface device configured to allow a user to input information and / or the drug delivery device 102 to output information for presentation to a user (e.g., an alarm signal, etc.).

[0028] The drug delivery device 102 includes a patient interface 186 for interfacing with a user to deliver a liquid drug. The patient interface 186 may be, for example, a needle or cannula for delivering the drug into the user's body (which may be done subcutaneously, intraperitoneally, or intravenously). The drug delivery device 102 may further include a mechanism for inserting the needle / cannula 186 into the user's body, which may be integral with the drug delivery device 102 or may be attachable to the drug delivery device 102. The insertion mechanism, in one embodiment, may consist of an actuator that inserts the needle / cannula 186 under the user's skin and then retracts the needle, leaving the cannula in place.

[0029] In one embodiment, the drug delivery device 102 includes a communication interface 126, which may be a transceiver operating according to one or more radio frequency protocols, such as Bluetooth, Wi-Fi, a near field communication standard, a cellular standard, etc. The controller 121 can communicate with the user device 105 and the analyte sensor 108, for example, via the communication interface 126.

[0030] In some embodiments, the drug delivery device 102 may include one or more sensors 184. The sensors 184 may be communicatively coupled to the controller 121 and may include one or more of a pressure sensor, a power sensor, etc. that provide various signals. For example, the pressure sensor may be configured to provide an indication of fluid pressure detected in the fluid path between the patient interface 186 and the reservoir 124. The pressure sensor may be coupled to or integral with an actuator for inserting the patient interface 186 into the user. In one example, the controller 121 may be operable to determine a drug infusion rate based on the indication of fluid pressure. The drug infusion rate may be compared to an infusion rate threshold, and the comparison result may be usable in determining an amount of remaining insulin (IOB) or a total daily insulin (TDI). In one embodiment, the analyte sensor 108 may be integral with the drug delivery device 102.

[0031] The drug delivery device 102 further includes a power source 128 , such as a battery, a piezoelectric device, an energy harvesting device, or the like, for powering the controller 121 , the memory 123 , the drive mechanism 125 and / or other components of the drug delivery device 102 .

[0032] The drug delivery device 102 may be configured to implement and execute the processes necessary to deliver a dose of medication to a user without input from the user device 105 or optional accessory device 106. As will be described in more detail, the MDA 129 may be operable to determine, for example, the amount of insulin to be delivered, the IOB, the amount of insulin remaining, etc., and cause the controller 121 to actuate the drive mechanism 125 to deliver the medication from the reservoir 124. The MDA 129 may take as input data received from the analyte sensor 108 or the user app 160.

[0033] Reservoirs 124, 124-2 can be configured to store drugs, medications, or therapeutic agents suitable for automated delivery, such as insulin, pramlintide, GLP-1, co-formulations of insulin and GLP-1, morphine, blood pressure medications, chemotherapy medications, and infertility medications.

[0034] The drug delivery device 102 may be a wearable device and may be attached to the user's body at an attachment location and may deliver any therapeutic agent, including any drug or pharmaceutical agent such as insulin, to the user at or near the attachment location. The surface of the drug delivery device 102 may include an adhesive to facilitate attachment to the user's skin.

[0035] When configured to communicate with an external device, such as a user device 105 or an analyte sensor 108, the drug delivery device 102 can receive signals from the user device 105 or the analyte sensor 108 via a wired or wireless link 194. A controller 121 of the drug delivery device 102 can receive and process the signals from the respective external device and effectuate the delivery of drugs to the user in accordance with a diabetes treatment plan or other drug delivery plan.

[0036] Attached Devices The optional accessory device 107 may be a wearable smart device, such as a smart watch (e.g., an Apple Watch®), smart glasses, smart jewelry, a global positioning system-enabled wearable, a wearable fitness device, or smart clothing. Like the user device 105, the accessory device 107 may be configured to perform various functions, including control of the drug delivery device 102. For example, the accessory device 107 may include a communications interface 174, a processor 171, a user interface 178, and memory 173. The user interface 178 may be a graphical user interface displayed on a touchscreen display of the smart accessory device 107. The memory 173 may store programming code for operating various functions of the smart accessory device 107, as well as instances of user apps 160 or simplified versions of the user apps 160 with reduced functionality. In some examples, the accessory device 107 may also include various types of sensors.

[0037] Analyte Sensor The analyte sensor 108 may include a controller 131, a memory 132, a sensing / measuring device 133, an optional user interface 137, a power / energy harvesting circuit 134, and a communication interface 135. The analyte sensor 108 may be communicatively coupled to a processor 151 of the management device 105 or a controller 121 of the drug delivery device 102. The memory 132 may be configured to store information and programming code 136.

[0038] The analyte sensor 108 may be configured to detect multiple different analytes, such as glucose, lactate, ketone, uric acid, sodium, potassium, alcohol level, etc., and output a detection result, such as a measurement value. In an exemplary embodiment, the analyte sensor 108 may be a continuous glucose monitor (CGM) configured to measure blood glucose levels at predetermined time intervals, such as every 5 minutes or every minute. The communication interface 135 of the analyte sensor 108 may have circuitry operating as a transceiver for communicating the measured blood glucose levels to the user device 105 via wireless link 195 or to the drug delivery device 102 via wireless communication link 108. Although referred to herein as the analyte sensor 108, the sensing / measuring device 133 of the analyte sensor 108 may include one or more additional sensing elements, such as a glucose measuring element, a heart rate monitor, a pressure sensor, etc. The controller 131 may include separate dedicated logic and / or components, an application-specific integrated circuit, a microcontroller or processor executing software instructions, firmware, programming instructions stored in a memory (such as memory 132), or any combination thereof.

[0039] Similar to the controller 121 of the drug delivery device 102, the controller 131 of the analyte sensor 108 may be operable to perform many functions. For example, the controller 131 may be configured by programming code 136 to manage the collection and analysis of data detected by the sensing and measuring device 133.

[0040] 1 is depicted as being separate from the drug delivery device 102, in various embodiments, the analyte sensor 108 and the drug delivery device 102 may be incorporated into the same unit. That is, in various examples, the analyte sensor 108 is part of the drug delivery device 102, may be integral with the drug delivery device 102, may be included in the same housing as the drug delivery device 102, or in a housing attachable to or otherwise adjacent to the housing of the drug delivery device 102. In such example configurations, the controller 121 can perform the functions necessary for proper delivery of medication alone, without any external input from the user device 105, the cloud-based service 111, another sensor (not shown), the optional accessory device 106, etc.

[0041] Cloud-based services The drug delivery system 100 can communicate with or receive services from a cloud-based service 111. The services provided by the cloud-based service 111 can include data storage for storing personal or anonymized data, such as blood glucose measurements, past IOB or TDI, previous carbohydrate compensation doses, and other forms of data. Additionally, the cloud-based service 111 can process anonymized data from multiple users to provide generalized information related to TDI, insulin sensitivity, IOB, etc. The communication link 115 coupling the cloud-based service 111 to each device 102, 105, 106, 108 of the system 100 can be a cellular link, a Wi-Fi link, a Bluetooth link, or a combination thereof.

[0042] communication links Wireless communication links 115 and 191-196 may be any type of wireless link operating using a known or proprietary wireless communication standard. By way of example, wireless communication links 191-196 may provide a communication link based on Bluetooth, Zigbee, Wi-Fi, a short-range wireless communication standard, a cellular standard, or any other wireless protocol via respective communication interfaces 126, 135, 154, 174.

[0043] Example of operation In an operational example, the user application 160 implements a graphical user interface that is the primary interface with the user and is used to start and stop the drug delivery device 102, program basal and bolus calculator settings for manual mode, and program settings specific to automatic mode (hybrid closed-loop or closed-loop).

[0044] The user app 160 provides a graphical user interface 158 that can use large text, graphics, and on-screen instructions to guide the user through the setup process and use of the system 100. It is also used to program a user-customized basal insulin delivery profile, check the status of the drug delivery device 102, initiate an insulin bolus, make changes to the patient's insulin delivery profile, handle system warnings and alarms, and allow the user to switch between automatic and manual modes.

[0045] The user app 160 can be configured to operate in a manual mode in which the user app 160 delivers insulin at programmed basal rates and bolus amounts, with the option to set basal or temporary basal profiles for different times of day. The controller 121 also has the ability to function as a sensor-equipped pump in manual mode, using sensor glucose data provided by the analyte sensor 108 to input into a bolus calculator.

[0046] The user app 160 can be configured to operate in an automatic mode that supports the use of multiple target blood glucose levels. For example, in one embodiment, the target blood glucose level can be in the range of 110-150 mg / dL, with increments of 10 mg / dL, 5 mg / dL, or other increments, preferably 10 mg / dL. The user experience mirrors the current setup flow, in which a healthcare provider assists the user in programming basal rate, glucose target, and bolus calculator settings, which then inform the user app 160 of insulin delivery parameters. The insulin delivery parameters are adapted over time based on the total daily insulin (TDI) delivered during each use of the drug delivery device 102. A temporary hypoglycemic protection mode can be implemented by the user in automatic mode for various periods of time. In hypoglycemic protection mode, the algorithm reduces insulin delivery and is intended for use during temporary periods when insulin sensitivity is expected to be increased, such as during exercise.

[0047] The user app 160 (or MDA 129) can provide periodic insulin microboluses based on past glucose measurements and / or predicted glucose over a prediction interval (e.g., 60 minutes). Optimal post-prandial control may require the user to deliver meal boluses in the same manner as current pump therapy, but the normal operation of the user app 160 compensates for missed meal boluses and reduces persistent hyperglycemia. The user app 160 employs a goal-directed control strategy that attempts to achieve and maintain set target glucose values, thereby reducing the duration of persistent hyperglycemia and hypoglycemia.

[0048] In some embodiments, the user device 105 and the analyte sensor 108 may not communicate directly with each other. Instead, data from the analyte sensor (e.g., a blood glucose measurement) may be communicated to the drug delivery device 102 via link 196 and then conveyed to the user device 105 via link 194. In some embodiments, the analyte sensor's serial number must be entered into the user app 160 to enable communication between the analyte sensor 108 and the user device 105.

[0049] The user app 160 may provide the ability to calculate a suggested bolus dose through the use of a bolus calculator. The bolus calculator is provided for the user's convenience to assist in determining a suggested bolus dose based on ingested carbohydrates, the most recent blood glucose reading (or blood glucose reading if using a fingerstick), a programmable correction factor, the insulin-to-carbohydrate ratio, the target glucose value, and active insulin (IOB). The IOB is estimated by the user app 160, taking into account any manual boluses and algorithmically delivered insulin, and the IOB may be divided between basal and bolus IOB, where the basal IOB takes into account algorithmically delivered insulin and the bolus IOB takes into account any bolus delivery.

[0050] Description of the embodiment The main embodiment of the present invention relates to a method for simplifying the optimization algorithm used in calculating the amount of liquid medication, e.g., insulin, to be periodically delivered to a user by a wearable drug delivery device 102. In the main embodiment, rather than running a computationally expensive optimization algorithm, a simple stepwise search over all possible search spaces can be performed, reducing the total computational cost and allowing the calculation to be performed in an embedded manner.

[0051] A typical control algorithm may utilize a model of the system being controlled to predict a proposed drug dosage output. This model can be used to determine the optimal dosage to administer to the user. For example, in an insulin delivery system, the user's glucose may be modeled as a recursive model of past glucose and insulin delivery values. While one skilled in the art will appreciate that any such model may be used, one such model may be represented by equation (1) as follows:

number

[0052] The glucose model can be run recursively to predict glucose levels for future cycles. For each series of proposed insulin doses I(k+N) for the next N cycles, a different glucose trajectory G(k+N) over those cycles can be calculated. The total value of this planned insulin dose and glucose trajectory can then be calculated by calculating a standard cost function, such as the exemplary cost represented by Equation (2) below:

number

[0053] Again, as will be appreciated by those skilled in the art, any cost function may be used, and exemplary cost functions are described in detail in U.S. Patent Application No. 16 / 789,051 (U.S. Published Patent Application No. 2021 / 0244881).

[0054] In certain embodiments, glucose deviation and insulin delivery cost can be performed in various ways, such as calculating deviations relative to particular targets (such as glucose control set points and basal insulin delivery), and deviations can be calculated for various orders, such as second or higher powers.

[0055] In a typical application, calculating the cost of each possible glucose and insulin trajectory and determining the trajectory with the lowest cost can be performed by a computationally expensive optimization algorithm, but as outlined above, many medical applications do not require a computationally expensive optimization algorithm.

[0056] This is especially true for insulin delivery applications, as the typical minimum resolution of insulin pumps (e.g., 0.05 U) and short-term variations in the timing of when each dose is delivered do not result in a significant change in the user's overall glucose results.

[0057] A stepwise algorithm that runs through all possible search spaces is now described and is illustrated in flowchart form in FIG. 3. In step 302 of the stepwise algorithm, a range of possible glucose delivery amounts for the current cycle is determined. Typically, the MDA 129 has the option to prevent or recommend that no dose be delivered during any cycle, so the lower limit of the range is 0.0. The upper range or limit of possible glucose delivery amounts may be constrained by safety constraints built into the MDA 129 to prevent excessive insulin from being delivered, or for any other reason. For example, in one embodiment, the delivery range may be set between 0.0 U and 0.6 U per cycle. This exemplary range is used for the purposes of describing the disclosed method.

[0058] In step 304, the range is defined into a coarse search space of coarsely discrete quantities. For example, for an exemplary range of 0U to 0.6U, the search space may be partitioned into coarsely discrete quantities of 0.1U. Thus, the coarsely partitioned search space would be partitioned as follows: 0.0U, 0.1U, 0.2U, 0.3U, 0.4U, 0.5U, 0.6U, etc. As will be appreciated by those skilled in the art, any coarsely discrete quantity that is coarse relative to the minimum delivery resolution of the drug delivery device can be used to define the range.

[0059] In step 306, a coarse search space is optimized. The search space can be first narrowed by determining a coarse optimal insulin delivery. Specifically, all insulin delivery rates in the following N cycles can be set at fixed values ​​in a coarse, discrete amount within the search space, and the corresponding glucose trajectory can be calculated.

[0060] In an illustrative example, the recursive model expressed in equation (1) may be calculated seven times, where I(k+1)...I(k+N) are calculated for each quantity in the coarse search space. The corresponding cost can then be calculated using a cost function, an example of which is expressed in equation (2). In an illustrative example, this results in the following exemplary calculation: [Table 1]

[0061] Thus, according to the exemplary model and cost function, administering 0.3 U of glucose in the next cycle will be the least costly.

[0062] In step 308, the coarse partition of the search space is refined. The search space is narrowed to a smaller region around the coarse discrete quantity with the lowest cost (0.3U in the illustrated example). In the second refined partition of the search space, the initial insulin delivery (i.e., I(k+1)) search space can be divided into smaller increments centered around the coarse discrete quantity with the lowest cost to provide finer resolution of the actual insulin delivery to be provided to the user. The search space of the remaining dataset (i.e., I(k+2)...I(k=N)) can be evaluated using the coarse discrete quantities.

[0063] In the illustrative example, the search space is narrowed and centered around 0.3 U. Thus, the first projected insulin delivery can be defined to be between 0.2 U and 0.4 U in smaller increments (e.g., 0.025 U), and the remaining projected insulin deliveries can be defined to be between 0.2 U and 0.4 U in the same or coarser increments (e.g., 0.05 U), preferably in coarser increments to reduce computational requirements. As will be appreciated by those skilled in the art, the smaller increments can be any desired amount, while the range can be any range, preferably centered around the coarse solution with the smallest cost. In step 310, the refined search space can be optimized. In the illustrative example, the calculations yield the following results: [Table 2]

[0064] Thus, in the illustrated example, only 52 total calculations were required, including 7 calculations in the coarse search space optimization step 306 and 45 calculations in the refined search space optimization step 310 (9[0.2, 0.225, 0.25, 0.275, 0.3, 0.325, 0.35, 0.375, 0.4] × 5[0.2, 0.25, 0.3, 0.35, 0.4]). In other embodiments in which different intervals and ranges are selected, the total number of calculations required in both steps 306 and 310 may vary. However, incremental exploration of the coarse and refined search spaces is much more computationally efficient than running traditional, computationally expensive optimization algorithms.

[0065] Finally, at step 312, the recommended insulin dosage is finalized. The solution that provides the lowest cost in the second refined search space may be provided as the recommended insulin delivery trajectory. In the illustrated example, as seen above, the minimum calculation of the cost function occurs in the 0.375 interval, and therefore the recommended insulin delivery dosage for the current cycle would be 0.375 U. At 314, the recommended dosage is delivered to the user by the drug delivery device 102, as described above.

[0066] It should be noted that if the recommended dose does not fall within the resolution of the drug delivery device 102, the actual delivered dose may be rounded up or down to the nearest discrete increment deliverable by the drug delivery device 102 based on the resolution. The remainder, which falls between the increment / discrete resolution of the drug delivery device 102, is added to or subtracted from the recommended dose for the next cycle.

[0067] The above process is repeated for each cycle of the MDA 129. In one embodiment, the MDA 129 may perform a cycle every five minutes, although other intervals may be selected.

[0068] As mentioned above, it is important to note that the various search space and resolution parameters used by the disclosed method are widely adjustable. For example, the coarse spacing calculations can be reduced to 0.2U resolution rather than the 0.1U resolution used in the illustrative example.

[0069] Overall, the disclosed method significantly reduces the number of iterations of glucose model and cost function calculations. The 52 calculations provided in the illustrative example above are significantly less computationally intensive than typical optimization algorithms with large, multivariate search spaces. The resolution of drug dosing recommendations is still within the resolution of the drug delivery mechanism.

[0070] The following examples relate to various embodiments of the systems and methods disclosed herein for providing a method for determining optimal delivery of a drug by providing a step-by-step exploration of a search space of possible amounts of the drug to be delivered to a user.

[0071] Example 1 is a method comprising the steps of defining a coarse search space, evaluating a model and a cost function for each discrete quantity in the coarse search space, defining a refined search space, evaluating the model and cost function across the refined search space, and selecting the refined discrete quantity with the lowest cost as the recommended dosage of the drug.

[0072] Example 2 is an extension of Example 1 or any other example disclosed herein, where the base uses a discrete quantity that is smaller than the discrete quantity used in the coarse search space.

[0073] Example 3 is an extension of Example 1 or any other example disclosed herein, where the refined search space is smaller than the coarse search space.

[0074] Example 4 is an extension of Example 1, or any other example disclosed herein, where the range of possible drug doses is from zero to a maximum amount determined by the drug delivery algorithm.

[0075] Example 5 is an extension of Example 4, or any other example disclosed herein, where the maximum amount depends on safety constraints built into the drug delivery algorithm.

[0076] Example 6 is an extension of Example 1 or any other example disclosed herein, where the drug is insulin delivered by a wearable drug delivery device.

[0077] Example 7 is an extension of Example 6, or any other example disclosed herein, where the model was a recursive glucose model used to predict a user's glucose levels for a predetermined number of future cycles based on a specific amount of insulin delivered.

[0078] Example 8 is an extension of Example 7 or any other example disclosed herein, where the cost function calculates the cost of each possible glucose and insulin trajectory and determines the trajectory with the lowest cost.

[0079] Example 9 is an extension of Example 8 or any other example disclosed herein, where the cost function is based on glucose deviation and insulin delivery, where the glucose deviation is calculated relative to a particular target.

[0080] Example 10 is a system comprising a processor and software implementing a drug delivery algorithm executed by the processor, wherein the software determines a recommended dosage of a drug for a current cycle of the drug delivery algorithm by performing functions of defining a coarse search space, evaluating a model and a cost function for each discrete quantity in the core search space, determining a refined search space, evaluating the model and the cost function over the refined search space, and selecting the refined discrete quantity with the lowest cost as the recommended dosage of the drug.

[0081] Example 11 is an extension of Example 10, or any other example disclosed herein, further including a drug delivery device for delivering a recommended dosage of a drug to a user.

[0082] Example 12 is an extension of Example 11 or any other example disclosed herein, where the processor and software are integral with the drug delivery device.

[0083] Example 13 is an extension of Example 10 or any other example disclosed herein, where the refined discrete quantity is smaller than the coarse discrete quantity.

[0084] Example 14 is an extension of Example 10, or any other example disclosed herein, where the refined search space is smaller than the coarse search space.

[0085] Example 15 is an extension of Example 10, or any other example disclosed herein, where the range of possible drug doses is from zero to a maximum amount determined by the drug delivery algorithm.

[0086] Example 16 is an extension of Example 15, or any other example disclosed herein, where the maximum amount depends on safety constraints built into the drug delivery algorithm.

[0087] Example 17 is an extension of Example 11, or any other example disclosed herein, where the drug is insulin delivered by a wearable drug delivery device.

[0088] Example 18 is an extension of Example 17, or any other example disclosed herein, where the model is a recursive glucose model used to predict a user's glucose levels for a predetermined number of future cycles based on a specific amount of insulin delivered.

[0089] Example 19 is an extension of Example 18, or any other example disclosed herein, where the cost function calculates the cost of each possible glucose and insulin trajectory and determines the trajectory with the lowest cost.

[0090] Example 20 is an extension of Example 19 or any other example disclosed herein, where the cost function is based on glucose deviation and insulin delivery deviation, where the glucose deviation is calculated relative to a specific target.

[0091] Software-related implementations of the techniques described herein may include, but are not limited to, firmware, application-specific software, or any other type of computer-readable instructions that may be executed by one or more processors. The computer-readable instructions may be provided via a non-transitory computer-readable medium. Hardware-related implementations of the techniques described herein include, but are not limited to, integrated circuits (ICs), application-specific ICs (ASICs), field programmable arrays (FPGAs), and / or programmable logic devices (PLDs). In some examples, the techniques described herein and / or any system or configuration components described herein may be implemented using a processor executing computer-readable instructions stored on one or more memory components.

[0092] Those skilled in the art to which the present invention pertains may realize many modifications and adaptations of the present invention. The implementations provided herein, including the values ​​of adjustable parameters, should be considered exemplary and are not intended to limit the present invention in any way. As those skilled in the art will appreciate, many variations of the implementations described herein are possible and fall within the scope of the present invention. Furthermore, it should be understood that the features of the various embodiments described herein are not mutually exclusive and may exist in various combinations and permutations without departing from the spirit and scope of the present invention, even if such combinations or permutations are not expressly stated herein. Accordingly, the methods and apparatus disclosed herein should be considered as illustrative of the present invention, and not as limiting thereof. [Configuration 1] determining a search space of possible drug delivery amounts in the current cycle of drug delivery in an automated drug delivery system; dividing the search space into ranges of solutions; narrowing the search space by determining a selection range within the range of solutions that minimizes a cost function; dividing the selected range into a list of refined values ​​for a first drug delivery, the list of refined values ​​being separated by a first small increment; evaluating a set of subsequent deliveries following the first drug delivery in the list of refined values, the set of subsequent deliveries being separated by a second coarse increment; and selecting a solution based on the set of subsequent deliveries that minimizes the cost function; A method comprising: [Configuration 2] 2. The method of claim 1, wherein the first fine increment is smaller than the second coarse increment. [Configuration 3] 2. The method of claim 1, wherein dividing the search space into solution ranges is performed using third coarse increments. [Configuration 4] 2. The method of claim 1, wherein the search space of possible drug delivery amounts ranges from zero to a maximum amount subject to one or more safety constraints. [Configuration 5] 2. The method of configuration 1, wherein the drug is insulin and the automatic drug delivery system is part of a wearable drug delivery device. [Configuration 6] 2. The method of claim 1, wherein determining the selection range includes evaluating a recursive glucose model configured to predict a glucose level of a user of the automated drug delivery system for a predetermined number of future cycles based on delivery of a specific amount of insulin. [Configuration 7] 7. The method of claim 6, wherein the cost function calculates the cost of each possible glucose and insulin trajectory and determines the trajectory with the lowest cost. [Configuration 8] A non-transitory computer-readable medium, comprising: determining a search space of possible drug delivery amounts for a current cycle of drug delivery in an automated drug delivery system; Dividing the search space into solution ranges; narrowing the search space by determining a selection range within a range of solutions that minimizes a cost function; dividing the selected range into a list of refined values ​​for a first drug delivery, the list of refined values ​​being separated by a first small increment; evaluating a set of subsequent deliveries following the first drug delivery in the list of refined values, the set of subsequent deliveries being separated by a second coarse increment; and selecting a solution based on the set of subsequent deliveries that minimizes the cost function; A non-transitory computer-readable medium that stores instructions configured to cause a processor to execute the following: [Configuration 9] 9. The medium of claim 8, wherein the first fine increment is smaller than the second coarse increment. [Configuration 10] 9. The medium of claim 8, wherein dividing the search space into solution ranges is performed using third coarse increments. [Configuration 11] 9. The medium of claim 8, wherein the search space of possible drug delivery amounts ranges from zero to a maximum amount subject to one or more safety constraints. [Configuration 12] 9. The medium of claim 8, wherein the drug is insulin and the automatic drug delivery system is part of a wearable drug delivery device. [Configuration 13] The medium of configuration 8, wherein determining the selection range includes evaluating a recursive glucose model configured to predict a glucose level of a user of the automated drug delivery system for a predetermined number of future cycles based on the delivery of a specific amount of insulin. [Configuration 14] 15. The medium of claim 14, wherein the cost function calculates the cost of each possible glucose and insulin trajectory and determines the trajectory with the lowest cost. [Configuration 15] a reservoir configured to store a drug; a drug delivery device configured to administer the drug to a user; processor and; A non-transitory computer-readable medium, comprising: determining a search space of possible drug delivery doses for the current cycle of drug delivery; Dividing the search space into solution ranges; narrowing the search space by determining a selection range within a range of solutions that minimizes a cost function; dividing the selected range into a list of refined values ​​for a first drug delivery, the list of refined values ​​being separated by a first small increment; evaluating a set of subsequent deliveries following the first drug delivery in the list of refined values, the set of subsequent deliveries being separated by a second coarse increment; and selecting a solution based on the set of subsequent deliveries that minimizes the cost function; a non-transitory computer-readable medium storing instructions configured to cause the processor to execute the A system for automated drug delivery comprising: [Configuration 16] 16. The system of claim 15, wherein the first fine increment is smaller than the second coarse increment. [Configuration 17] 16. The system of claim 15, wherein dividing the search space into solution ranges is performed using third coarse increments. [Configuration 18] 16. The system of claim 15, wherein the search space of possible drug delivery amounts ranges from zero to a maximum amount subject to one or more safety constraints. [Configuration 19] 16. The system of claim 15, wherein the medication is insulin and the automatic medication delivery system is part of a wearable medication delivery device. [Configuration 20] 16. The system of claim 15, wherein determining the selection range includes evaluating a recursive glucose model configured to predict the user's glucose levels for a predetermined number of future cycles based on delivery of a specific amount of insulin.

Claims

1. determining a search space of possible drug delivery amounts for a current cycle of drug delivery to a user of the automated drug delivery system; modeling the user's analyte state as a recursive model of past analytes and drug delivery amounts; exploring the search space by calculating predicted analyte trajectories of the possible drug delivery amounts using the recursive model; and finalizing a next drug delivery dose by refining the search space; rounding the possible drug delivery amount in the search space to the nearest discrete amount deliverable by the automated drug delivery system to generate a rounded difference; adding or subtracting the rounded difference to or from the calculated drug delivery amount for the next cycle of drug delivery to the user. A computer-implemented method comprising:

2. The computer-implemented method of claim 1 , wherein the search space of possible drug delivery amounts ranges from zero to a maximum amount subject to one or more safety constraints.

3. 10. The computer-implemented method of claim 1, wherein the medication is insulin and the automatic medication delivery system is part of a wearable medication delivery device.

4. The computer-implemented method of claim 1, wherein the analyte state is a user's glucose level.

5. The computer-implemented method of claim 4, wherein the recursive model is a recursive glucose model configured to predict the glucose levels of a user of the automated drug delivery system for a predetermined number of future cycles based on the delivery of a specific amount of insulin.

6. The computer-implemented method of claim 5, wherein refining the search space includes applying a cost function that calculates the cost of each possible glucose and insulin trajectory and determines the trajectory with the lowest cost.

7. Searching the search space comprises: dividing the search space into solution ranges; narrowing the search space by determining a selection range within a range of solutions that minimize a cost function; dividing the selection range into a list of refined values ​​for a first drug delivery, the list of refined values ​​being separated by a first small increment; evaluating a set of subsequent deliveries following the first drug delivery in the list of refined values, the set of subsequent deliveries being separated by a second coarse increment; and selecting a solution based on the set of subsequent deliveries that minimizes the cost function; The computer-implemented method of claim 1 further comprising:

8. A non-transitory computer-readable medium, comprising: determining a search space of possible drug delivery amounts for a current cycle of drug delivery to a user of the automated drug delivery system; modeling the user's analyte state as a recursive model of past analytes and drug delivery amounts; exploring the search space by calculating predicted analyte trajectories of the possible drug delivery amounts using the recursive model; finalizing a next drug delivery dose by refining the search space; rounding the possible drug delivery amount in the search space to the nearest discrete amount deliverable by the automated drug delivery system to generate a rounded difference; adding or subtracting the rounded difference to or from the calculated drug delivery amount for the next cycle of drug delivery to the user; A non-transitory computer-readable medium that stores instructions configured to cause a processor to execute the following:

9. The medium of claim 8 , wherein the search space of possible drug delivery amounts ranges from zero to a maximum amount subject to one or more safety constraints.

10. 9. The medium of claim 8, wherein the drug is insulin and the automatic drug delivery system is part of a wearable drug delivery device.

11. The medium described in claim 8, wherein the analyte state is the user's glucose level.

12. The medium described in claim 11, wherein the recursive model is a recursive glucose model configured to predict the glucose levels of a user of the automated drug delivery system for a predetermined number of future cycles based on the delivery of a specific amount of insulin.

13. The medium described in claim 12, wherein refining the search space includes calculating the cost of each possible glucose and insulin trajectory and applying a cost function to determine the trajectory with the lowest cost.

14. Searching the search space comprises: dividing the search space into solution ranges; narrowing the search space by determining a selection range within a range of solutions that minimize a cost function; dividing the selection range into a list of refined values ​​for a first drug delivery, the list of refined values ​​being separated by a first small increment; evaluating a set of subsequent deliveries following the first drug delivery in the list of refined values, the set of subsequent deliveries being separated by a second coarse increment; and selecting a solution based on the set of subsequent deliveries that minimizes the cost function; The medium of claim 8 further comprising:

15. An automated drug delivery system comprising: a reservoir configured to store a drug; a drug delivery device configured to administer the drug to a user; a processor; A non-transitory computer-readable medium, comprising: determining a search space of possible drug delivery amounts for a current cycle of drug delivery to a user of the automated drug delivery system; modeling the user's analyte state as a recursive model of past analytes and drug delivery amounts; exploring the search space by calculating predicted analyte trajectories of the possible drug delivery amounts using the recursive model; finalizing a next drug delivery dose by refining the search space; rounding the possible drug delivery amount in the search space to the nearest discrete amount deliverable by the automated drug delivery system to generate a rounded difference; adding or subtracting the rounded difference to or from the calculated drug delivery amount for the next cycle of drug delivery to the user; a non-transitory computer-readable medium storing instructions configured to cause the processor to execute the An automated drug delivery system comprising:

16. 16. The automated drug delivery system of claim 15, wherein the search space of possible drug delivery amounts ranges from zero to a maximum amount subject to one or more safety constraints.

17. 16. The automatic drug delivery system of claim 15, wherein the drug is insulin and the automatic drug delivery system is part of a wearable drug delivery device.

18. The automatic drug delivery system of claim 15, wherein the analyte state is a user's glucose level.

19. The automatic drug delivery system of claim 18, wherein the recursive model is a recursive glucose model configured to predict the glucose levels of a user of the automatic drug delivery system for a predetermined number of future cycles based on the delivery of a specific amount of insulin.

20. The automated drug delivery system of claim 19, wherein refining the search space includes calculating the cost of each possible glucose and insulin trajectory and applying a cost function to determine the trajectory with the lowest cost.

Images