Use of logarithmic transformation / filtering to improve the operation of drug delivery devices

Abstract

Exemplary embodiments may apply a transformation or filtering to the user's analyte level values ​​to fit a normal distribution that is symmetric about a mean. The transformed or filtered analyte level values ​​may be used by a control system of a drug delivery device in determining drug delivery dosages. In some embodiments, the drug is insulin and the analyte level is the user's glucose value. In such cases, logarithmic filtering or transformation may be applied to the user's glucose measurements.

Description

[Technical Field]

[0001] Related Applications This application claims the benefit of U.S. Provisional Patent Application No. 63 / 374,042, filed August 31, 2022, the entire contents of which are incorporated herein by reference. [Background technology]

[0002] Many conventional drug delivery devices employ control systems that attempt to maintain a user's analyte levels at target analyte levels. For example, many conventional insulin delivery devices attempt to maintain a user's glucose levels at a target glucose value. These conventional control systems continuously collect the user's glucose measurements and compare the glucose measurements to a target glucose value. Based at least in part on the difference between the glucose measurements and the target glucose value, the conventional insulin delivery device makes a decision regarding the amount of insulin to deliver to the user. Other factors may also play a role.

[0003] Conventional insulin delivery devices can employ a glucose cost function that assigns a cost to glucose excursions that would exist in the future if a given dose of insulin were delivered to the user currently. A glucose excursion is a process in which the user's glucose value deviates from a target glucose value for a non-negligible period of time. The objective of the control system is to reduce such glucose excursions. The control system of such conventional insulin delivery devices can apply an optimization algorithm to select an insulin dose with the lowest glucose cost and deliver the selected insulin dose with the lowest glucose cost to the user. Summary of the Invention [Problem to be solved by the invention]

[0004] One challenge with such conventional analyte cost functions (e.g., glucose cost functions) is that they may assume that the user's analyte levels used by the control system to determine the user's drug dosage follow a Gaussian or normal distribution and are symmetric about a mean value. However, the actual distribution of the user's analyte levels may be non-Gaussian and asymmetric. As a result, the control system of such conventional drug delivery devices may not perform as desired. [Means for solving the problem]

[0005] According to an aspect of the present invention, an insulin delivery device for delivering insulin to a user includes a non-transitory computer-readable storage medium storing computer program instructions and a processor for executing the computer program instructions. By executing the computer program instructions, the processor determines an insulin dose having a minimum glucose cost to be delivered to the user by the insulin delivery device. The glucose cost is determined based on the user's effective glucose value, and the effective glucose value is determined by applying a logarithmic transformation to predicted glucose values ​​for a time horizon. By executing the computer program instructions, the insulin dose is delivered to the user by the insulin delivery device.

[0006] The predicted glucose values ​​may be predicted from the user's previous glucose values ​​for previous times and / or from the user's predicted glucose values ​​for previous times when previous glucose values ​​are not yet available. The non-transitory computer-readable storage medium may store the user's glucose level history, and the user's glucose values ​​for times prior to the earliest time in the time horizon may be part of the glucose value history. The insulin delivery device may further include an insulin reservoir for holding insulin for delivering insulin to the user. The computer program instructions, when executed by the processor, may cause the processor to initiate delivery of an insulin dose from the reservoir to the user. The processor may apply a logarithmic function to the predicted glucose values ​​when applying a logarithmic transformation to the predicted glucose values ​​for the time horizon. Applying the logarithmic transformation may involve, for each predicted glucose value, applying the logarithmic function to a product of a scaling factor and a ratio of the predicted glucose value to a setpoint glucose value to produce a logarithmic value. Applying the logarithmic transformation may further involve multiplying the logarithmic value by a setpoint glucose value to generate one of the predicted logarithmic glucose values. The glucose cost may be the product of a coefficient and the sum of the squares of the differences between each of the predicted glucose values ​​and the setpoint glucose value over the time horizon.

[0007] According to another aspect of the present invention, an insulin delivery system for delivering insulin to a user includes a non-transitory computer-readable storage medium storing computer program instructions and a processor for executing the computer program instructions. The computer program instructions cause the processor to select, for delivery to the user by an insulin delivery device, a selected one of the candidate insulin doses having a best cost value when applying a cost function to the candidate insulin doses. The cost function includes an insulin cost component and a glucose cost component. The glucose cost component is based on a cumulative difference between predicted effective glucose values ​​and a setpoint glucose value over a time horizon. The predicted effective glucose value is determined by applying a logarithmic function to the user's predicted glucose values ​​over the time horizon. The computer program instructions include computer program instructions for delivering the selected one of the candidate insulin doses to the user.

[0008] At least one of the predicted glucose values ​​may be determined by the processor from the user's glucose values ​​at a time point prior to the time horizon. The insulin delivery system may further include an insulin tank for holding insulin. The computer program instructions, when executed by the processor, may cause the processor to initiate delivery of a selected one of the candidate insulin doses from the tank to the user. The predicted glucose values ​​may be determined by the processor from glucose values ​​prior to the time horizon and / or from predicted glucose values ​​when previous glucose values ​​are not yet available. Each of the predicted valid glucose values ​​may be determined by applying a logarithmic function to the product of a scaling factor and a ratio of the corresponding predicted glucose value to one of the setpoint glucose values ​​to generate a logarithmic value. The computer program instructions, when executed by the processor, may further cause the processor to multiply the logarithmic value by the setpoint glucose value to generate one of the predicted logarithmic glucose values. The glucose cost value may be the product of a factor and the sum of the squares of the differences between each of the predicted glucose values ​​over the time horizon and the setpoint glucose value. The insulin delivery system, in some embodiments, may include other components, such as a smartphone or one or more sensors.

[0009] According to another aspect of the present invention, an insulin delivery device for delivering insulin to a user includes a non-transitory computer-readable storage medium storing computer program instructions and a processor for executing the computer program instructions for controlling automatic insulin delivery (AID) by the insulin delivery device to the user by the processor, the control including converting the user's glucose value to a logarithmic glucose value, using the logarithmic glucose value in determining an insulin dose to be delivered to the user by the insulin delivery device as part of the AID, and delivering the determined insulin dose from the insulin delivery device to the user.

[0010] The determined insulin dose may be for basal delivery to the user. The converting may include applying a logarithmic function to the glucose value. [Brief explanation of the drawings]

[0011] [Figure 1] FIG. 1 illustrates an exemplary embodiment of a drug delivery system for delivering drugs to a user.

[0012] [Figure 2] FIG. 2 illustrates an exemplary control loop of the drug delivery system of an exemplary embodiment.

[0013] [Figure 3] FIG. 3 shows a flowchart of exemplary steps that may be performed by a control loop in an exemplary embodiment.

[0014] [Figure 4A] FIG. 4A shows a histogram of the density of the population's estimated glucose values ​​(EGVs) and a normal fit overlay.

[0015] [Figure 4B] Figure 4B shows the curve of the cumulative probability of EGV in the population superimposed with the corresponding curve of the cumulative probability of the Gaussian distribution.

[0016] [Figure 5A] FIG. 5A shows the inputs and outputs of a conventional control system for an insulin delivery device.

[0017] [Figure 5B] FIG. 5B shows the inputs and outputs of the control system of an insulin delivery device according to an exemplary embodiment.

[0018] [Figure 6A] Figure 6A shows a histogram of the density of EGVs in the population and an overlay of the log-normal fit.

[0019] [Figure 6B] Figure 6B shows the curve for the cumulative probability of EGV in the population superimposed with the corresponding curve for the cumulative probability of EGV in a log-normal distribution.

[0020] [Figure 7A] FIG. 7A shows a flowchart of exemplary steps that may be performed in an exemplary embodiment to determine the insulin dosage for a cycle.

[0021] [Figure 7B] FIG. 7B shows a histogram of the density of the population's valid glucose values ​​Geff(i) and a normal fit overlay.

[0022] [Figure 7C] FIG. 7C shows the curve of the cumulative probability of the value of Geff(i) for the population superimposed with the corresponding curve of the cumulative probability of the value of Geff(i) for a normal distribution.

[0023] [Figure 8] FIG. 8 shows a flowchart of exemplary steps that may be performed in an exemplary embodiment to calculate Geff(i).

[0024] [Figure 9] FIG. 9 shows a flowchart of exemplary steps that may be performed in an exemplary embodiment to determine the total cost of a candidate insulin dose using Geff(i).

[0025] [Figure 10] FIG. 10 shows a flowchart of exemplary steps that may be performed in an exemplary embodiment to determine glucose costs.

[0026] [Figure 11] FIG. 11 shows a flowchart of exemplary steps that may be performed in an exemplary embodiment to determine Geff(i) for future cycles. DETAILED DESCRIPTION OF THE INVENTION

[0027] Exemplary embodiments recognize that the analyte levels of a drug delivery device user, used by the drug delivery device's control system to affect drug delivery, may not fit a Gaussian distribution and may not be symmetric about a mean value. For example, exemplary embodiments described herein recognize that the distribution of glucose measurements across a population fits a log-normal distribution rather than a Gaussian ("normal") distribution. Exemplary embodiments may apply a transformation (such as a logarithmic function) or filtering to the user's analyte level values ​​to fit the analyte level values ​​to a normal distribution that is symmetric about the mean. The transformed or filtered analyte level values ​​may be used by the drug delivery device's control system in determining the drug delivery dosage. In some embodiments, the drug is insulin and the analyte level is the user's glucose value. In such cases, logarithmic filtering or transformation may be applied to the user's glucose measurements.

[0028] In some exemplary embodiments, the drug delivery device may be an insulin delivery device, such as a patch insulin pump worn by a user. In such embodiments, the control system of the insulin delivery device may use a glucose cost function to determine the cost of a glucose excursion of a candidate insulin dose. In exemplary embodiments, the glucose cost function may calculate the cost based on the logarithm of the user's current glucose value. The logarithm of the current glucose value shows a Gaussian distribution of values ​​across a population of users. Using logarithmic values ​​rather than raw glucose measurements may improve the performance of the insulin delivery device. Simulation results show a 1% to 2% improvement in time in the desired range for users when the described approach is used.

[0029] 1 illustrates an exemplary drug delivery system 100 suitable for delivering a drug to a user 108 according to an exemplary embodiment. The drug delivery system 100 may include a drug delivery device 102. The drug delivery device 102 may be attached to the body of the user 108 or may be a wearable device carried by the user 108. The drug delivery device 102 may be directly coupled to the user (e.g., attached directly to a body part and / or skin of the user via adhesive, etc.) or may be carried by the user (e.g., carried on a belt or in a pocket) with the drug delivery device 102 connected to an injection site where the drug is injected using a needle and / or cannula. The surface of the drug delivery device 102 may include an adhesive to facilitate attachment to the user 108.

[0030] The drug delivery device 102 may include a processor 110. The processor 110 may be, for example, a microprocessor, logic circuit, field programmable gate array (FPGA), application specific integrated circuit (ASIC), or microcontroller. The processor 110 may maintain date and time as well as other functions (e.g., calculations, etc.). The processor 110 may be operable to execute a control application 116 encoded with computer program instructions stored in the memory device 114, which allows the processor 110 to direct the operation of the drug delivery device 102. The control application 116 may be a single program, multiple programs, modules, libraries, etc. The control application may be responsible for executing a control loop that provides feedback and adjustment to drug dosage (i.e., when delivery occurs and how much drug dosage is delivered). The processor 110 may execute computer program instructions stored in the memory device 114 for a user interface 117, which may include one or more display screens displayed on the display 109. The display 109 may display information to the user 108 and, in some cases, may receive input from the user 108, such as when the display 109 is a touchscreen.

[0031] The control application 116 may control the delivery of drugs to the user 108 according to the control approach as described herein. The storage device 114 may maintain user history 111, such as basal delivery history, bolus delivery history, and / or other history, such as meal event history, exercise event history, glucose value history, other analyte level history, etc. The storage device 114 may include one or more basal profiles 115 used when the drug delivery device is operating in open-loop mode. Additionally, the processor 110 may be operable to receive data or information. The storage device 114 may include both primary and secondary storage. The storage device 114 may include random access memory (RAM), read-only memory (ROM), optical storage, magnetic storage, removable storage media, solid-state storage, etc.

[0032] The drug delivery device 102 may have a tray or cradle and / or one or more housings that house various components, including a pump 113, a power source (not shown), and a reservoir 112 that stores the drug for delivery to the user 108. A fluid pathway to the user 108 may be provided, and the drug delivery device 102 may use the pump 113 to eject medication from the reservoir 112 for delivery to the user 108 via the fluid pathway. The fluid pathway may include, for example, tubing connecting the drug delivery device 102 to the user 108 (e.g., tubing connecting a cannula to the reservoir 112) or may include a conduit to a separate infusion site.

[0033] For example, there may be one or more communication links with one or more devices physically separate from the drug delivery device 102, including the user's and / or the user's caregiver's management device 104, sensors 106, smart watches 130, fitness monitors 132, and / or various other wearable devices 134. The communication links may include any wired or wireless communication link operating according to any known communication protocol or standard, such as Bluetooth, Wi-Fi, a near field communication standard, a cellular standard, or any other wireless protocol.

[0034] The drug delivery device 102 may communicate with the network 122 via a wired or wireless communication link. The network 122 may include a local area network (LAN), a wide area network (WAN), or a combination thereof. A computing device 126 may communicate with the network, and the computing device may communicate with the drug delivery device 102 or the management device 104.

[0035] The drug delivery system 100 may include one or more sensors 106 that detect the level of one or more analytes. The sensor(s) 106 may be coupled to the user 108, for example, by adhesive or the like, and may provide information or data regarding one or more medical conditions, physical attributes, or analyte levels of the user 108. The sensor(s) 106 may be physically separate from the drug delivery device 102 or may be an integrated component. In some embodiments, the sensor(s) 106 may include a glucose sensor, such as a continuous glucose monitor (CGM). The sensor(s) 106 may also detect other analyte levels, such as, for example, heart rate, respiration rate, temperature, altitude, movement, sweat, ketone levels, blood oxygen, alcohol levels, blood drug levels, etc.

[0036] The drug delivery system 100 may include a management device 104. In some embodiments, a management device is not required because the drug delivery device 102 can manage itself. The management device 104 may be a dedicated device, such as a dedicated personal diabetes manager (PDM) device. The management device 104 may be a programmed general-purpose device, such as any portable electronic device including a dedicated controller, such as a processor, microcontroller, or the like. The management device 104 may be used to program or coordinate the operation of the drug delivery device 102 and / or the sensor(s) 106. The management device 104 may be any portable electronic device, including, for example, a dedicated device, a smartphone, a smartwatch, or a tablet. In the depicted example, the management device 104 may include a processor 119 and a memory device 118. The processor 119 may execute processes for managing the user's glucose levels and controlling the delivery of medication to the user 108. The drug delivery device 102 may provide data from the sensor(s) 106 and other data to the management device 104. Data may be stored in the storage device 118. The processor 119 may be operable to execute program code stored in the storage device 118. For example, the storage device 118 may be operable to store one or more control applications 120 for execution by the processor 119. The control applications 120 may be responsible for controlling the drug delivery device 102, such as controlling the automatic drug delivery (ADD) of a drug to the user 108. The storage device 118 may store the control applications 120, a history 121, one or more base histories 135, as described above for the drug delivery device 102, and other data and / or programs.

[0037] A display 127, such as a touchscreen, may be provided for displaying information. The display 127 may display a user interface (UI) 123. The display 127 may also be used to receive input, such as when it is a touchscreen. The management device 104 may further include input elements 125, such as a keyboard, buttons, knobs, etc., for receiving input from the user 108.

[0038] The management device 104 may interact with a network 124, such as a LAN or a WAN, or a combination of such networks, via a wired or wireless communication link. The management device 104 may communicate with one or more servers or cloud services 128 via the network 124. Data, such as sensor values, may be transmitted directly from the drug delivery device 102 to one or more cloud services / servers 128 or from the management device 104 to one or more cloud services / servers 128 for storage and processing in some embodiments. The cloud service / server(s) 128 may provide output from the model 115 to the management device 104 and / or drug delivery device 102 during operation, as needed.

[0039] Other devices, such as a smart watch 130, a fitness monitor 132, and a wearable device 134, may be part of the drug delivery system 100. These devices 130, 132, and 134 may communicate with the drug delivery device 102 and / or the management device 104 to receive information and / or issue commands to the drug delivery device 102. These devices 130, 132, and 134 may execute computer program instructions to perform some of the control functions performed by the processor 110 or the processor 119, for example, via the control applications 116 and 120. These devices 130, 132, and 134 may have a display for displaying information. The display may display a user interface for providing input by the user 108, such as a request to change or pause a dosage, or to request, start, or confirm delivery of a bolus of drug, or may display a user interface for displaying output, such as a change in dosage (e.g., a basal delivery rate), determined by the processor 110 or the management device 104. These devices 130, 132, and 134 may have a wireless communication connection with the sensor 106 to directly receive the analyte measurement data.

[0040] A wide variety of drugs may be delivered by the drug delivery device 102. The drug may be insulin for treating diabetes. The drug may be glucagon for increasing a user's glucose levels. The drug may be a glucagon-like peptide (GLP)-1 receptor agonist for lowering glucose or slowing gastric emptying, thereby slowing post-prandial glucose spikes. Alternatively, the drug delivered by the drug delivery device 102 may be one of a pain reliever, a chemotherapy agent, an antibiotic, a blood thinner, a hormone, a blood pressure lowering agent, an antidepressant, an antipsychotic, a statin, an anticoagulant, an anticonvulsant, an antihistamine, an anti-inflammatory agent, a steroid, an immunosuppressant, an anti-anxiety agent, an antiviral agent, a nutritional supplement, or a vitamin.

[0041] The functionality described below for the exemplary embodiments may be under the control of or performed by the control application 116 of the drug delivery device 102 or the control application 120 of the management device 104. In some embodiments, the functionality may be under the control of or performed by a cloud service or server 128, a computing device 126, or other enumerated devices, including a smart watch 130, a fitness monitor 132, or another wearable device 134.

[0042] The drug delivery device 102 may operate in an open-loop mode or a closed-loop mode. In the open-loop mode, the user 108 manually inputs the amount of drug to be delivered for a portion of the day (such as per hour). The input may be stored in the user's 108 basal profile 115, 135. In other embodiments, a basal profile may not be used. In the open-loop mode, the control application 116, 120 uses the input information from the basal profile 115, 135 to control basal drug delivery. In contrast, in the closed-loop mode, the control application 116, 120 continuously determines the drug delivery amount for the user 108 based on a feedback loop. For an insulin delivery device, the objective of the closed-loop mode is to bring the user's glucose level to a target glucose level. The basal dose may be delivered at regular intervals, designated as a cycle, such as every 5 minutes. In some embodiments, a cycle may represent a period of time between about 1 minute and about 30 minutes, more particularly between about 2 minutes and about 15 minutes, and particularly between about 3 minutes and about 10 minutes.

[0043] As described above, a control loop may be provided to adjust a basal delivery dose of a drug based on a current analyte level measurement, such as a glucose measurement. FIG. 2 shows a simplified block diagram of an example of such a control loop 200 suitable for implementing exemplary embodiments. The exemplary control loop 200 may include a controller 202, a pump mechanism or other fluid extraction mechanism 204 (hereinafter, “pump 204”), and a sensor 208. The controller 202 may be part of the control application 116 or 120. The controller 202, pump 204, and sensor 208 may be communicatively coupled to each other via wired or wireless communication paths. In some exemplary embodiments, the sensor 208 may be a glucose monitor, such as a CGM. The sensor 208 may be operable to measure a user's glucose level, for example, to generate a measured analyte level 212. In some embodiments, the sensor 208 may apply a logarithmic filter or transformation to the analyte level measurement and transmit the resulting value to the control application 116 or 120 via a wireless connection, etc.

[0044] In the illustrated example, the controller 202 may receive a desired analyte level 210 indicating a desired analyte level or range for the user. The desired analyte level 210 may be received from a user interface to the controller 202 or another device, or by an algorithm that automatically determines the desired analyte level 210 for the user. The sensor 208 may be coupled to the user and operable to measure an approximation of the user's actual analyte level. Note that if the analyte level is a glucose value, the measured glucose value is only an approximation of the user's glucose value. There may be error in the measured glucose level. The error may be due to factors such as the age of the sensor 208, the location of the sensor 208 on the user's body, environmental factors (e.g., altitude, humidity, barometric pressure), etc. These measured glucose values ​​are sometimes referred to as "estimated glucose values" (EGVs). In response to the measured analyte level or value, the sensor 208 may generate a signal indicative of the measured analyte level 212. The controller 202 may receive the measured analyte level signal 212 from the sensor 208 via a communication path.

[0045] Based on the desired analyte level signal 210 and the measured analyte level signal 212, the controller 202 may calculate a glucose cost using a glucose cost function, as described below. In an exemplary embodiment, a logarithmic transformation may be applied to analyte level values, such as glucose values, and the transformed values ​​may be used to determine the analyte cost. The lowest cost drug dose may be selected for delivery by the controller 202. The controller 202 may then generate one or more control signals 214 that direct the operation of the pump 204. For example, one of the control signals 214 may cause the pump 204 to deliver a dose of medication 216 to the user via the output 206. The dose of medication 216 may be determined as the appropriate amount of medication to drive the user's actual analyte level toward the desired analyte level. Based on the operation of the pump 204 determined by the control signal 214, the user may receive a dose of medication 216 from the pump 204.

[0046] 3 shows a flowchart 300 of steps that may be performed by an exemplary embodiment in determining how much medication to deliver to a user as part of a closed-loop control system. These steps may be performed (at least in part) by processor 110, processor 119, or other components, such as smart watch 130, fitness monitor, or wearable device 134. That said, for simplicity, the following will only refer to processor 110. First, at 302, an analyte level measurement is obtained by sensor 208, as described above with respect to FIG. 2. At 304, the analyte level measurement is transmitted to controller 202 via signal 212.

[0047] The control system attempts to minimize the total penalty of the cost function over the range of possible doses constrained by the control system. At 306, the dose with the best cost function value (e.g., lowest cost) is selected. Depending on how the cost function is configured, the best value may be the smallest or largest value. The cost functions used in exemplary embodiments are described in more detail below. A candidate dose is a dose that meets the constraints imposed by the control system within the search space of all available doses. For example, the minimum dose size may be zero or the smallest positive amount. Another constraint may be, for example, a maximum dose size. The control system may apply an optimization strategy to find the lowest-cost candidate dose within the space. A regression strategy may also be used. The term "regression strategy" as used herein may refer to various statistical procedures for estimating the relationship between a dependent variable and one or more independent variables, particularly to identify local minima or maxima of the dependent variable, e.g., glucose cost or insulin cost.

[0048] At 308, a control signal 214 may be generated by the controller 202 and sent to the pump 204 to cause the pump to deliver a desired drug dose 216 to the user.

[0049] Before delving into the details of logarithmic transformation or filtering and cost functions, it is useful to review a problem encountered with some conventional drug delivery devices. FIG. 4A shows a plot 400 including a histogram 402 of the density of various EGVs for a population of users of a particular type of insulin delivery device. Also shown overlaid on the plot 400 is an ideal normal distribution curve 404. As shown, there are meaningful differences between the histogram 402 and the ideal normal distribution curve 404. The histogram 402 indicates that the users' actual EGVs reflected by the histogram 402 do not conform to a symmetric distribution. There are more EGVs above the peak value of the histogram than below the peak value. In other words, users typically experience more hyperglycemia than hypoglycemia (e.g., more blood glucose levels above 150 mg / dL than below 70 mg / dL). Furthermore, the peaks of the histogram 402 do not coincide with the peaks of the ideal normal distribution curve 404.

[0050] The difference between raw EGV and a Gaussian distribution can also be seen in FIG. 4B, which shows a plot of cumulative probability across all raw EGVs in a population. Curve 412 is the cumulative probability curve for the raw EGVs of the population. Curve 414 represents the cumulative probability curve for the raw EGVs of a Gaussian distribution (a "normal fit"). As shown, there is a discrepancy between the two curves 412 and 414, indicating that the actual EGVs do not fit a normal Gaussian distribution.

[0051] 5A shows an example block diagram 500 of a conventional control system 504 for an insulin delivery device. As described above, a user's glucose values ​​from a sensor(s) 106, such as a CGM, are input to the control system 504. The control system 504 uses the glucose values ​​504 to determine and output an insulin dose 506. Often, this process is repeated every operating cycle, where a cycle is a fixed period of time, such as five minutes. The insulin dose 506 is the insulin delivery for the current cycle.

[0052] 5B shows an example block diagram 510 of an example embodiment of a control system 514. In this example, the control system 514 uses logarithmic glucose values ​​512 rather than glucose values ​​to determine insulin doses 516. As discussed above, the use of logarithmic glucose values ​​improves a user's time in range.

[0053] 6A shows a plot of the density of EGVs. Histogram 602 shows the density of EGVs in a population. Curve 604 is a log-normal distribution curve. As shown, the distribution of EGVs appears to fit a log-normal fit (i.e., a log-Gaussian distribution). Thus, the log-normal fit appears to fit the EGVs better than the normal fit.

[0054] The good fit of the lognormal fit to EGV is also reflected in the plot 610 of cumulative probability for EGV shown in Figure 6B. Curve 612 is the cumulative probability of EGV from the population data, and curve 614 is the cumulative probability of the lognormal fit. The two curves 612 and 614 correspond closely.

[0055] 7A shows a flowchart 700 of exemplary steps that may be performed in an exemplary embodiment to determine the dose of insulin to be delivered for cycle i. At 702, the effective glucose value G eff (i) is determined from the glucose value G(i) for cycle i. The glucose value G(i) is then converted to the effective glucose value G(i), as described in more detail below. eff (i) is converted or filtered to a logarithmic representation designated as (i). The conversion / filtering modifies the glucose values ​​to logarithmic values ​​that fit a log-normal distribution. At 704, a cost function is used to find the lowest cost G eff (i) is determined. Therefore, G eff (i) is used in the cost function rather than the traditional approach of using G(i) in the cost function. eff An insulin value I(i) associated with (i) is delivered to the user by the insulin delivery device.

[0056] Figure 7B shows the G eff A density plot 710 for (i) is shown. A histogram 712 shows the density of the various Gs of the user population. eff A normal distribution curve 714 is overlaid on histogram 712. As shown, histogram 712 closely matches the normal distribution of curve 714. Figure 7C shows the distribution of G eff 7 shows a plot 720 of the cumulative probability curve for (i). A cumulative probability curve 724 for a normal distribution is overlaid. As shown, the two curves 722 and 724 are very similar, and G eff (i) provides further evidence for a fit to a Gaussian normal distribution.

[0057] FIG. 8 shows an example embodiment of the effective glucose value G eff 8 shows a flowchart 800 of exemplary steps that may be performed to determine (see 702) (i). eff One suitable formula for calculating (i) is as follows:

number

[0058] It should be understood that other transformations / filtering may be applied to fit the glucose values ​​to a normal fit or other desired type of distribution.

[0059] As discussed above, a cost function may be used to determine the best insulin dose for a user in a cycle. Figure 9 shows a flowchart 900 of exemplary steps that may be performed in an exemplary embodiment to determine the total cost of candidate insulin doses for a cycle. At 902, the effective glucose value G eff (i) to determine the glucose cost for a candidate insulin dose. G eff An example of how to calculate glucose cost using (i) is described in more detail below: Glucose cost is the cost of the glucose excursion. At 904, determine the insulin cost for the candidate insulin dose. Insulin cost is the cost of the insulin excursion relative to the basal insulin standard. At 906, determine the total cost for the candidate insulin dose by summing the glucose cost and the insulin cost. A cost equation suitable for an exemplary embodiment is as follows:

number

number

number

[0060] 10 shows a flowchart 1000 of exemplary steps that may be performed in an exemplary embodiment to determine the glucose cost of a candidate insulin dose for cycle i. The appropriate formula is:

number

[0061] For the value of future time horizon, G effTo determine (i), an exemplary embodiment may perform the steps shown in flowchart 1100 of Figure 11 for each future cycle in the time horizon. First, at 1102, the user's glucose value for cycle i is predicted. One suitable approach to predicting the glucose value G(i) in the current or future cycle is to base the prediction on the glucose value of the immediately preceding cycle. The following formula may be used:

number

[0062] The present disclosure also relates to computer programs having instructions (also referred to as computer program instructions) that perform functions under the circumstances. The instructions may be executed by a processor. The instructions may be executed by multiple processors, for example, in a distributed computer system. The computer programs of the present disclosure may be pre-installed or downloaded to, for example, a drug delivery device, a management device, or a fluid delivery device, for example, in its storage device.

[0063] Although exemplary embodiments have been described herein, it should be understood that various changes in form and detail may be made therein without departing from the intended scope of the claims appended hereto.

[0064] While the invention is defined in the appended claims, it should be understood that the invention may (alternatively) be defined according to the following embodiments. 1. An insulin delivery device for delivering insulin to a user, comprising: a non-transitory computer-readable storage medium storing computer program instructions; a processor for executing the computer program instructions, determining an insulin dose with a minimum glucose cost to be delivered to a user by the insulin delivery device, the glucose cost being determined based on an effective glucose value of the user, the effective glucose value being determined by applying a logarithmic transformation to predicted glucose values ​​for a time horizon; delivering the insulin dose to a user by the insulin delivery device; a processor that executes An insulin delivery device comprising: 2. The insulin delivery device of claim 1, wherein the predicted glucose value is predicted from the user's previous glucose value for a previous time and / or from the user's predicted glucose value for a previous time when the previous glucose value is not yet available. 3. The insulin delivery device of claim 2, wherein the non-transitory computer-readable storage medium stores a user's glucose value history, and the user's glucose values ​​for times before the earliest time in the time horizon are part of the glucose value history. 4. The insulin delivery device of claim 1, further comprising an insulin reservoir for holding insulin for delivering insulin to a user. 5. The insulin delivery device of claim 4, wherein the computer program instructions, when executed by the processor, cause the processor to initiate delivery of an insulin dose from the reservoir to a user. 6. The insulin delivery device of claim 1, wherein the processor, when applying a logarithmic transformation to the predicted glucose values ​​for the time horizon, applies a logarithmic function to the predicted glucose values. 7. The insulin delivery device of claim 6, wherein applying the logarithmic transformation comprises applying, for each of the predicted glucose values, a logarithmic function to the product of a scaling factor and a ratio of the predicted glucose value to a setpoint glucose value to produce a logarithmic value. 8. The insulin delivery device of claim 7, wherein applying the logarithmic transformation further comprises multiplying the logarithmic value by the setpoint glucose value to generate one of the predicted logarithmic glucose values. 9. The insulin delivery device of claim 8, wherein the glucose cost is the product of a coefficient and the sum of the squares of the differences between each of the predicted glucose values ​​over the time horizon and the setpoint glucose value. 10. An insulin delivery system for delivering insulin to a user, comprising: a non-transitory computer-readable storage medium storing computer program instructions; a processor for executing the computer program instructions, selecting a selected one of the candidate insulin doses having a best cost value when applying a cost function to the candidate insulin doses for delivery to the user by the insulin delivery device; the cost function has an insulin cost component and a glucose cost component; the glucose cost component is based on a cumulative difference between a predicted valid glucose value and a setpoint glucose value over a time horizon; determining the predicted valid glucose value by applying a logarithmic function to the user's predicted glucose values ​​over the time horizon; delivering a selected one of the candidate insulin doses to a user; a processor that executes An insulin delivery system comprising: 11. The insulin delivery system of claim 10, wherein at least one of the predicted glucose values ​​is determined by the processor from the user's glucose values ​​at a time point prior to the time horizon. 12. The insulin delivery system of claim 10, further comprising an insulin reservoir for holding insulin. 13. The insulin delivery system of claim 12, wherein the computer program instructions, when executed by the processor, cause the processor to initiate delivery of a selected one of the candidate insulin doses from the reservoir to a user. 14. The insulin delivery system of claim 13, wherein the predicted glucose value is determined by the processor from glucose values ​​prior to the time horizon and / or from the predicted glucose value. 15. The insulin delivery system of claim 14, wherein each of the predicted valid glucose values ​​is determined by applying a logarithmic function to the product of a scaling factor and the ratio of the corresponding predicted glucose value to one of the setpoint glucose values ​​to produce a logarithmic value. 16. The insulin delivery system of claim 15, wherein the computer program instructions, when executed by the processor, further cause the processor to multiply the logarithmic value by the setpoint glucose value to generate one of the predicted logarithmic glucose values. 17. The insulin delivery system of claim 16, wherein the glucose cost value is the product of a coefficient and the sum of the squares of the differences between each of the predicted glucose values ​​and the setpoint glucose value over the time horizon. 18. An insulin delivery device for delivering insulin to a user, comprising: a non-transitory computer-readable storage medium storing computer program instructions; a processor for executing the computer program instructions, controlling insulin delivery to a user by the insulin delivery device, said control comprising: converting the user's glucose value to a logarithmic glucose value; using the logarithmic glucose value in determining an insulin dose to be delivered to a user by the insulin delivery device; delivering the determined insulin dose from the insulin delivery device to a user; a processor that executes the An insulin delivery device comprising: 19. The insulin delivery device of claim 18, wherein the determined insulin dosage is for basal delivery to a user. 20. The insulin delivery device of claim 18 or 19, wherein the converting comprises applying a logarithmic function to the glucose value. 21. A computer-implemented method for determining an insulin dose having a lowest glucose cost to be delivered to a user by an insulin delivery device, the method comprising determining the glucose cost based on the user's effective glucose value, the effective glucose value being determined by applying a logarithmic transformation to predicted glucose values ​​for a time horizon. 22. Selecting a selected one of the candidate insulin doses having a best cost value when applying a cost function to the candidate insulin doses for delivery to a user by an insulin delivery device, comprising: the cost function has an insulin cost component and a glucose cost component; the glucose cost component is based on a cumulative difference between a predicted valid glucose value and a setpoint glucose value over a time horizon; A computer-implemented method comprising determining the predicted valid glucose values ​​by applying a logarithmic function to the user's predicted glucose values ​​over the time horizon.

Claims

1. 1. An insulin delivery device for delivering insulin to a user, comprising: a non-transitory computer-readable storage medium storing computer program instructions; a processor for executing the computer program instructions, determining an insulin dose with a minimum glucose cost to be delivered to a user by the insulin delivery device, the glucose cost being determined based on an effective glucose value of the user, the effective glucose value being determined by applying a logarithmic transformation to predicted glucose values ​​for a time horizon; delivering the insulin dose to a user by the insulin delivery device; a processor that executes An insulin delivery device comprising:

2. 10. The insulin delivery device of claim 1, wherein the predicted glucose value is predicted from the user's previous glucose value for a previous time and / or from the user's predicted glucose value for a previous time when the previous glucose value is not yet available.

3. 3. The insulin delivery device of claim 1, wherein the non-transitory computer-readable storage medium stores a user's glucose value history, and the user's glucose values ​​for times before the earliest time in the time horizon are part of the glucose value history.

4. The insulin delivery device of any one of claims 1 to 3, further comprising an insulin reservoir for holding insulin for delivering insulin to a user.

5. 5. The insulin delivery device of claim 4, wherein the computer program instructions, when executed by the processor, cause the processor to initiate delivery of an insulin dose from the reservoir to a user.

6. The insulin delivery device of claim 1 , wherein the processor applies a logarithmic function to the predicted glucose values ​​when applying a logarithmic transformation to the predicted glucose values ​​for the time horizon.

7. 7. The insulin delivery device of claim 6, wherein applying the logarithmic transformation comprises, for each of the predicted glucose values, applying a logarithmic function to the product of a scaling factor and a ratio of the predicted glucose value to a setpoint glucose value to produce a logarithmic value.

8. 8. The insulin delivery device of claim 7, wherein applying the logarithmic transformation further comprises multiplying the logarithmic value by the setpoint glucose value to generate one of the predicted logarithmic glucose values.

9. 9. The insulin delivery device of claim 1, wherein the glucose cost is the product of a coefficient and the sum of the squares of the differences between each of the predicted glucose values ​​and the setpoint glucose value over the time horizon.

10. 1. An insulin delivery system for delivering insulin to a user, comprising: a non-transitory computer-readable storage medium storing computer program instructions; a processor for executing the computer program instructions, selecting a selected one of the candidate insulin doses having a best cost value when applying a cost function to the candidate insulin doses for delivery to the user by the insulin delivery device; the cost function has an insulin cost component and a glucose cost component; the glucose cost component is based on a cumulative difference between a predicted valid glucose value and a setpoint glucose value over a time horizon; determining the predicted valid glucose value by applying a logarithmic function to the user's predicted glucose values ​​over the time horizon; delivering a selected one of the candidate insulin doses to a user; a processor that executes An insulin delivery system comprising:

11. 11. The insulin delivery system of claim 10, wherein at least one of the predicted glucose values ​​is determined by the processor from the user's glucose values ​​at a time point prior to the time horizon.

12. 12. The insulin delivery system of claim 10 or 11, further comprising an insulin reservoir for holding insulin.

13. 13. The insulin delivery system of claim 12, wherein the computer program instructions, when executed by the processor, cause the processor to initiate delivery of a selected one of the candidate insulin doses from the reservoir to a user.

14. 14. The insulin delivery system of claim 10, wherein the predicted glucose value is determined by the processor from glucose values ​​prior to the time horizon and / or from the predicted glucose value.

15. 15. The insulin delivery system of claim 10, wherein each of the predicted valid glucose values ​​is determined by applying a logarithmic function to the product of a scaling factor and the ratio of the corresponding predicted glucose value to one of the setpoint glucose values ​​to produce a logarithmic value.

16. 16. The insulin delivery system of claim 15, wherein the computer program instructions, when executed by the processor, further cause the processor to multiply the logarithmic value by the setpoint glucose value to generate one of the predicted logarithmic glucose values.

17. 17. The insulin delivery system of claim 10, wherein the glucose cost value is the product of a coefficient and the sum of the squares of the differences between each of the predicted glucose values ​​and the setpoint glucose value over the time horizon.

18. 1. An insulin delivery device for delivering insulin to a user, comprising: a non-transitory computer-readable storage medium storing computer program instructions; a processor for executing the computer program instructions, controlling insulin delivery to a user by the insulin delivery device, said control comprising: converting the user's glucose value to a logarithmic glucose value; using the logarithmic glucose value in determining an insulin dose to be delivered to a user by the insulin delivery device; delivering the determined insulin dose from the insulin delivery device to a user; a processor that executes the An insulin delivery device comprising:

19. 20. The insulin delivery device of claim 18, wherein the determined insulin dose is for basal delivery to a user.

20. 20. The insulin delivery device of claim 18 or 19, wherein the converting comprises applying a logarithmic function to the glucose value.

Images