Transdermal analyte sensors and monitors, their calibration, and associated methods

Abstract

The present disclosure provides systems and methods for calibrating an analyte concentration sensor using only signals from the analyte concentration sensor within a biological system. For example, at steady state, the analyte concentration value within the biological system is known and that value can provide a calibration source. Similar techniques can be employed with a slow moving average. Several variations are disclosed.

Description

[0001] This application is a divisional application of a Chinese patent application filed on September 8, 2016, with application number 201680036303.6 and invention title "Transcutaneous Analyte Sensors and Monitors, Their Calibration and Associated Methods". Technical Field

[0002] Systems and methods for processing sensor data from a continuous analyte sensor and for calibration of the sensor. Background Art

[0003] Diabetes is a disorder in which the pancreas fails to produce sufficient insulin (type I or insulin-dependent) and / or the insulin is not effective (type 2 or non-insulin-dependent). In a diabetic state, a patient experiences hyperglycemia, which can cause a number of physiological disorders associated with the deterioration of small blood vessels, such as kidney failure, skin ulcers, or vitreous hemorrhage in the eye. Hypoglycemic reactions (hypoglycemia) can be caused by an inadvertent overdose of insulin, or after a normal dose of insulin or glucose-lowering agent accompanied by excessive exercise or insufficient food intake.

[0004] Conventionally, people with diabetes carry self-monitoring blood glucose (SMBG) monitors, which typically require an uncomfortable finger-sticking method. Due to the lack of comfort and convenience, people with diabetes usually measure their glucose levels only two to four times a day. Unfortunately, these time intervals are spread too far apart, such that a person with diabetes may discover hyperglycemic or hypoglycemic conditions too late, sometimes incurring dangerous side effects. Alternatively, glucose levels can be continuously monitored by a sensor system that includes a sensor assembly on the skin. The sensor system can have a wireless transmitter that transmits measurement data to a receiver, which can process and display information based on the measurement.

[0005] To date, a variety of glucose sensors have been developed for continuous measurement of glucose values. Many implantable glucose sensors suffer from problems within the body and only provide short-term and less accurate sensing of blood glucose. Similarly, transcutaneous sensors have encountered problems with continuously and accurately sensing and reporting back glucose values over an extended time period. Some efforts have been made to obtain blood glucose data from implantable devices and retrospectively determine blood glucose trends for analysis; however, these efforts have not helped diabetic patients determine real-time blood glucose information. Some efforts have also been made to obtain blood glucose data from transcutaneous devices for prospective data analysis, yet similar problems have occurred.

[0006] In a continuous glucose monitor (CGM), after implanting the sensor, it is calibrated, after which it provides substantially continuous sensor data to the sensor electronics. The sensor electronics transform the sensor data such that an estimated analyte value can be continuously provided to the user. As used herein, the terms "substantially continuous," "continuously," etc. may refer to a data stream of individual measurements taken at time-spaced intervals, the range of which may be from fractions of a second up to, for example, 1, 2, or 5 minutes or more. While the sensor electronics continuously receive sensor data, the sensor may be occasionally recalibrated to account for possible changes (drifts) in sensor sensitivity and / or baseline. Sensor sensitivity may refer to the amount of electrical current generated in the sensor by a predetermined amount of the measured analyte.

[0007] The sensor baseline refers to the signal output by the sensor when no analyte is detected. Over time, the sensitivity and baseline change due to various factors, including cellular attack or migration of cells to the sensor, which can affect the ability of the analyte to reach the sensor.

[0008] This background is provided to introduce a brief context for the subsequent detailed description and the invention. This background is not intended to help determine the scope of the claimed subject matter, nor is it considered to limit the claimed subject matter to embodiments that solve any or all of the disadvantages or problems presented above. Summary of the Invention

[0009] Without limiting the scope of embodiments of the invention as expressed by the appended claims, the prominent features of systems and methods in accordance with the principles of the invention will be briefly discussed. After considering this discussion, and particularly after reading the section entitled "Detailed Description," the features of embodiments of the invention will be understood to provide the advantages described herein.

[0010] In a first aspect, a method is provided for calibrating an analyte concentration sensor using only signals from an analyte concentration sensor within a biological system, where the analyte concentration value within the biological system is known upon the occurrence of a repeatable event, the method comprising: on a monitoring device, detecting when an analyte concentration value measured by an analyte concentration sensor disposed within the biological system constitutes a first repeatable event; and on the monitoring device or on a device or server operatively coupled to the monitoring device, correlating a measurement of the analyte concentration value when the biological system is in the detected first repeatable event with the known analyte concentration value.

[0011] Embodiments of the examples and aspects may include one or more of the following. The correlating may include determining a functional relationship between a sensor reading and the known analyte concentration value. The functional relationship may include a multiplicative constant. The detecting may include waiting a predetermined time after an event such as a meal or exercise has occurred on the monitoring device. The method may further include: after the correlating, detecting the occurrence of a second repeatable event different from the first repeatable event; and recalibrating the analyte concentration sensor by correlating the sensor reading when the biological system is in the detected second repeatable event with the known analyte concentration value. The sensor reading may have a first raw value at initial calibration and a second raw value at recalibration, where the first and second raw values are different. The method may further include: after the correlating, displaying a graph or table indicating current measured and historical values of the analyte concentration calibrated at least in part based on the correlating; and, after the recalibrating, updating the display of the graph or table indicating current measured and historical values of the analyte concentration according to the recalibration. The updating may change the display of the historical values of the analyte concentration. The method may further include: determining the difference between the first and second raw values; comparing the amount based on the difference with a predetermined criterion, and determining whether the sensor calibration has drifted based on the comparison. The method may further include determining a quantified amount by which the sensor calibration has drifted. The method may further include adjusting the sensor calibration based on the determined quantified amount. The amount may be the slope between the first and second raw values. The method may further include if the slope exceeds a predetermined threshold, then prohibiting further calibration based on steady state until the slope no longer exceeds the predetermined threshold. The method may further include prompting the user to enter a measured value. The sensor may be a glucose sensor. The method may further include: after the correlating, receiving a signal from the sensor; and displaying a value corresponding to the received signal, the displayed value being based on the received signal and the known analyte concentration value. The method may further include the step of determining the known analyte concentration value by prompting the user to enter a measured value. The method may further include the step of determining the known analyte concentration value by accessing a population average. The recalibration may be configured to occur at a time when the sensor reading is substantially stable, or within a predetermined range of readings over a threshold time period, thereby reducing the occurrence of unexpected jumps in the readings, and these recalibrations at these times may be caused or configured to occur in any of the described examples or aspects.

[0012] In a second aspect, a method is provided for compensating for drift in an analyte concentration sensor within a biological system using only signals from the analyte concentration sensor, the method comprising: measuring an analyte value using an implanted analyte concentration sensor; determining a first slow moving average of the measured values of the analyte over a first time period and calibrating the sensor at least in part based on the first slow moving average; after the first determination, determining a second slow moving average of the measured values of the analyte over a second time period; and adjusting the calibration of the sensor at least in part based on a difference between the first slow moving average and the second slow moving average.

[0013] Embodiments of the aspect and examples can include one or more of the following. The duration of the first time period can be greater than about 12 hours or greater than about 24 hours. The duration of the first time period can be the same as the duration of the second time period. The method can further comprise: after calibrating the sensor at least in part based on the first slow moving average, displaying a graph or table that indicates at least historical values of the analyte concentration calibrated at least in part based on the first slow moving average; and after the adjustment, updating the display of the graph or table indicating at least historical values of the analyte concentration according to the adjusted calibration. The update can change the display of the historical values of the analyte concentration. The displayed graph or table can further indicate the current measured value of the analyte concentration. Calibrating the sensor at least in part based on the first slow moving average can further comprise calibrating based on a seed value, such as a population average or a seed value received from a previous session. The method can further comprise after the adjustment, changing the seed value at least in part based on the adjustment. The method can further comprise changing the seed value based on the difference between the first slow moving average and the second slow moving average.

[0014] In a third aspect, a method is provided for compensating for drift in an analyte concentration sensor within a biological system using only signals from the analyte concentration sensor, the method comprising: measuring an analyte value using an implanted analyte concentration sensor; determining a first slow moving average of the measured value of the analyte over a first time period and making a first apparent sensitivity of the sensor at least partially based on the first slow moving average; after the first determination, determining a second slow moving average of the measured value of the analyte over a second time period and making a second apparent sensitivity of the sensor at least partially based on the second slow moving average; determining whether a change in the apparent sensitivity of the sensor between the first apparent sensitivity and the second apparent sensitivity matches a predetermined criterion; if the change in apparent sensitivity matches the predetermined criterion, then adjusting an actual sensitivity of the sensor to an adjusted value based on a difference between the first and second apparent sensitivities; if the change in apparent sensitivity fails to match the predetermined criterion, then prompting a user for input data, whereby a cause of the change in apparent sensitivity can be determined.

[0015] Implementations of the aspect and embodiments can include one or more of the following. The determination can include determining whether the change in apparent sensitivity is due to sensitivity drift or a change in the slow moving average. If the change in apparent sensitivity is due to a change in the slow moving average, then the method can further include prompting the user for data regarding the change. The predetermined criterion can include a known performance of the sensitivity change of the sensor over time. The known performance of the sensitivity change can constitute an envelope of acceptable sensitivity changes relative to time. The adjusted value can be at least partially based on the second slow moving average. The predetermined criterion can further include a known value of a physiologically viable analyte change. The prompting of the user can include prompting the user for a calibration value. The prompting of the user can include prompting the user for meal or exercise information. After receiving the calibration value or the meal or exercise information from the user, the method can further include determining whether the change in apparent sensitivity is due to sensitivity drift or a change in the slow moving average. The method can further include adjusting the actual sensitivity of the sensor based on the received calibration value or meal or exercise information.

[0016] In a fourth aspect, there is provided a method of checking calibration of an analyte concentration sensor system within a biological system using only signals from the analyte concentration sensor, comprising: after an initial calibration, measuring values of the analyte over time using the implanted analyte concentration sensor; calculating a clinical value of the analyte concentration based on the measured values and the initial calibration; adjusting the initial calibration to an updated calibration, the adjustment being based only on the measured values of the analyte over time or a subset thereof; and calculating a clinical value of the analyte concentration based on the measured values and the updated calibration.

[0017] Embodiments of the aspects and examples may include one or more of the following. The initial calibration may be based on a population average or data entered by a user. The adjustment may be based on a moving average of the measured values of the analyte over time. The adjustment may be based on a steady state value of the analyte. The initial calibration may be based on data determined prior to a session associated with the implanted analyte concentration sensor. The data may be determined a priori, on a bench, or in vitro.

[0018] In a fifth aspect, there is provided a method of calibrating an analyte concentration sensor using signals from an analyte concentration sensor within a biological system, wherein, at steady state, the analyte concentration value within the biological system is known, the method comprising: receiving, on a monitoring device, a seed value of a calibration parameter; detecting, on the monitoring device, when an analyte concentration value measured by an analyte concentration sensor implanted in the biological system is at steady state; and correlating, on the monitoring device or on a device or server operatively coupled to the monitoring device, the measurement of the analyte concentration value when the biological system is at the detected steady state with the known analyte concentration value; after the correlation, receiving signals from the sensor; and calculating and displaying a value corresponding to the received signals, the calculated value being based on the received signals, the known analyte concentration value, and the seed value.

[0019] Embodiments of aspects and examples may include one or more of the following. The received seed value may be received from a source containing factory calibration information. The method may further include: detecting performance of the received signal outside of pre-specified parameters; and prompting the user to input external calibration information. The displayed value may further be based on the external calibration information. The external calibration information may be received from SMBG or finger stick calibration. The method may further include resetting the known calibration value to a new known calibration value, the resetting being at least partially based on the external calibration information. The method may further include resetting the seed value to a new seed value, the resetting being at least partially based on the external calibration information. The method may further include changing the display based on the determined accuracy of the value. The changing of the display may include displaying a range instead of a value, or displaying a value instead of a range. The received seed value of the calibration parameter may be a user input characterization of a disease state. The user input characterization of the disease state may include an indication of type I diabetes, type II diabetes, non-diabetes, or pre-diabetes. The received seed value of the calibration parameter may be a value based on one or more user input blood glucose values. The display of the value corresponding to the received signal may include displaying a graph or table indicating current measured and historical values of the analyte concentration, and the method further includes: detecting that a calibration change has occurred; adjusting one or more calibration parameters of the analyte concentration sensor according to the calibration change; and after the adjustment, updating the display of the graph or table indicating current measured and historical values of the analyte concentration according to the adjusted calibration parameters. The detecting that a calibration change has occurred may include: detecting a change in a slow moving average; or detecting a change in the steady state value.

[0020] In a sixth aspect, a method of calibrating an analyte concentration sensor is provided, where after insertion of the sensor within a patient, only using a sensor signal or a parameter derivable from the sensor signal, the method includes: receiving at least an initial value of an analyte concentration and an initial value or initial value distribution of sensor sensitivity; after insertion of the analyte concentration sensor, monitoring a signal from the sensor over a duration; during the duration, calculating a plurality of analyte concentration values based on the monitored sensor signal and the initial value or value distribution of the sensor sensitivity; determining a value distribution of the monitored signal during the duration; optimizing the initial value or the value distribution of the sensor sensitivity and the plurality of analyte concentration values to match the value distribution of the monitored signal; and determining an updated sensitivity based on the optimization.

[0021] Embodiments of the aspects and examples may include one or more of the following. The receiving may be receiving an initial value distribution of sensor sensitivity, and the calculating of the plurality of analyte concentration values may be based on the monitored sensor signal and a representative value from the initial value distribution of the sensor sensitivity. The representative value may be selected from an average value or a midpoint or a median. The determining of the updated sensitivity may further include: dividing the representative value by the initial value of the analyte concentration; and updating the value of the sensitivity to be equal to the result of the division. The initial value of the analyte population may be a population average, may be input by a user, or may be transmitted from a previous session. The optimization may include optimizing the product of the initial value or value distribution of the sensor sensitivity and the plurality of analyte concentration values. The optimizing of the product may include optimizing the product to match the value distribution of the monitored signal, while adjusting the parameters of the value distribution of the sensor sensitivity and the plurality of analyte concentration values to most closely match the respective population averages. The receiving may further include receiving an initial value distribution of a baseline, and the optimization may further include optimizing the value distribution of the baseline together with the value distribution of the sensor sensitivity and the plurality of analyte concentration values to match the value distribution of the monitored signal. The initial value distribution of the baseline may follow a normal distribution. At least the initial value of the analyte concentration may be used as part of a seed value input to a slow moving average filter. The initial value distribution of the sensor sensitivity may be bounded by a normal distribution. The determined value distribution of the monitored signal may follow a lognormal distribution. The duration may be one day. The method may further include continuing to determine an updated sensitivity based on a previously updated sensitivity and the received analyte concentration values. The method may further include detecting a slow moving average of the monitored analyte concentration values. If the absolute value of the change in the slow moving average over a predetermined time unit is greater than a predetermined threshold, then the method may include prompting the user to input data. If the absolute value of the change in the slow moving average over a predetermined time unit is greater than a predetermined threshold, then the method may include determining whether the change is due to a systematic error or due to an actual sensitivity change of the sensor. The determining of whether the change is due to a systematic error or due to an actual sensitivity change of the sensor may include determining whether the subsequent behavior of the sensitivity is consistent with a known sensitivity distribution, including being consistent with the envelope of the sensitivity curve. If the absolute value of the change in the slow moving average is determined to be due to an actual sensitivity change of the sensor, then the method may include updating the sensitivity at least in part based on the value of the change in the slow moving average. The determining of whether the change is due to a systematic error or due to an actual sensitivity change of the sensor may include determining whether the subsequent behavior of the analyte concentration values is consistent with a known envelope of physiological feasibility.If the absolute value of the change in the slow moving average is determined to be due to a systematic error, then the method may include prompting the user to enter data.

[0022] In a seventh aspect, there is provided a method of calibrating an analyte concentration sensor using signals from an analyte concentration sensor within a biological system, the method comprising: receiving or determining a seed value for a calibration parameter associated with the analyte concentration sensor; using the seed value to at least partially determine the calibration of the analyte concentration sensor; and using the analyte concentration sensor, measuring a value of the analyte concentration; and displaying the measured value as at least partially calibrated using the seed value.

[0023] Implementations of the aspect and embodiments may include one or more of the following. The receiving or determining may be performed on a monitoring device in operative signal communication with the analyte concentration sensor. The displaying may be performed on the monitoring device or on a mobile device in signal communication with the monitoring device. The displaying the measured value may include displaying a graph or table indicating at least historical values of the analyte concentration, and the method may further include: detecting that a calibration change has occurred; adjusting one or more calibration parameters of the analyte concentration sensor in accordance with the detected calibration change; and after the adjustment, updating the display of the graph or table indicating at least historical values of the analyte concentration in accordance with the adjusted calibration parameters. The updating may change the display of the historical values of the analyte concentration. The seed value may be at least partially based on a code. The code may be input by the user into the monitoring device. The monitoring device may be configured to receive the code without substantial participation of the user. The seed value may be at least partially based on an impedance measurement. The seed value may be at least partially based on information associated with a manufacturing lot of the sensor. The seed value may be at least partially based on a population average. The seed value may be at least partially based on the user's just past analyte value.

[0024] In an eighth aspect, there is provided a method of calibrating and compensating for drift in an analyte concentration sensor using only signals from an implanted analyte concentration sensor within a biological system, where the analyte concentration value within the biological system is known at steady state. The method includes: on a monitoring device, detecting when an analyte concentration value measured by an analyte concentration sensor implanted within the biological system is at steady state; on the monitoring device or on a device or server operatively coupled to the monitoring device, correlating the measurement of the analyte concentration value when the biological system is at the detected steady state to the known analyte concentration value; determining a first slow moving average of the measured values of the analyte over a first time period, and calibrating the sensor at least in part based on the first slow moving average and based on the known analyte concentration value; after the first determination, determining a second slow moving average of the measured values of the analyte over a second time period; and adjusting the calibration of the sensor at least in part based on a difference between the first slow moving average and the second slow moving average.

[0025] In a ninth aspect, there is provided a method of calibrating a first portion of a batch of sensors, where a second portion has been in use. The method includes: receiving calibration data from some of the sensors in the second portion; and updating one or more calibration parameters of the first portion based on the received data.

[0026] Embodiments of the aspects and examples can include one or more of the following. The update can be performed before the first portion is implanted in a user's body. The update can be performed after the first portion has been implanted in a user's body. The update can be performed by transmitting new or updated calibration parameters via a network to a monitoring device or a sensor electronics module associated with the sensor. The sensors in the second portion can be configured to be calibrated using a priori calibration. The sensors in the second portion can be configured to be calibrated using user data. The sensors in the second portion can be configured to be calibrated using in vitro bench calibration. The sensors in the second portion can be configured to be calibrated using blood measurements.

[0027] In a tenth aspect, there is provided a method of compensating for drift in an analyte concentration sensor within a biological system using only signals from the analyte concentration sensor, which includes: measuring time-varying values of an analyte using an implanted analyte concentration sensor; filtering the measured values using a double exponential smoothing filter; and after the filtering, displaying the filtered measured values versus time.

[0028] Embodiments of the aspects and examples can include one or more of the following. The double exponential smoothing filter can be governed by the equations described herein. The subsequent glucose signal varying over time can be provided by the equations described herein.

[0029] In an eleventh aspect, a method of calibrating an analyte concentration sensor using only signals from an analyte concentration sensor within a biological system is provided, where at or during a repeatable event, the analyte concentration value within the biological system is known, the method comprising: on a monitoring device, detecting when a set of analyte concentration values measured by an analyte concentration sensor disposed within the biological system constitutes a repeatable event; and on the monitoring device or on a device or server operatively coupled to the monitoring device, correlating the set of analyte concentration values at the repeatable event to the known analyte concentration value.

[0030] Embodiments can include that the repeatable event is selected from the group consisting of: steady state, postprandial rise, daily high-low glucose excursion, decay rate, or rate of change.

[0031] In a twelfth aspect, a method of compensating for drift in an analyte concentration sensor within a biological system using only signals from the analyte concentration sensor is provided, which includes: measuring an analyte value using a disposed analyte concentration sensor; determining a first slow moving average of the measured values of the analyte over a first set of time periods, where the first set includes event-based time periods, and calibrating the sensor at least in part based on the first slow moving average; after the first determination, determining a second slow moving average of the measured values of the analyte over a second set of time periods, where the second set includes event-based time periods; and adjusting the calibration of the sensor at least in part based on the difference between the first slow moving average and the second slow moving average.

[0032] Embodiments can include one or more of the following. The first and second event-based time periods can be selected from the group consisting of: postprandial time periods, sleep time periods, and post-breakfast time periods.

[0033] In a thirteenth aspect, a method of calibrating an analyte concentration sensor within a biological system is provided, which includes: for a set of sensors of a type, determining the sensitivity distribution over time; for an individual sensor of the type, measuring the sensitivity distribution; measuring the electrical characteristics of a transmitter; and reading the identifier of the sensor and receiving data corresponding to the sensitivity of the sensor, and storing the identifier and the received data on the transmitter.

[0034] Embodiments may include one or more of the following. One type of the set of sensors may correspond to a set of sensors within a batch. The method may further include packaging individual sensors of the transmitter as a kit. The method may further include coupling the transmitter to a mobile device running a monitoring application. The method may further include using the monitoring application to calibrate the transmitter and the sensors. The calibration may be with respect to the measured electrical characteristics of the transmitter. The monitoring application may be configured to start a sensor session after a signal from the transmitter that detects the coupling of the transmitter to a sensor. The transmitter may be configured to start a sensor session when the transmitter detects a coupling to a sensor. The method may further include coupling the transmitter to a mobile device running a monitoring application. The method may further include receiving a set of representative measured analyte values. The method may further include using the received set of representative measured analyte values or a subset thereof to determine seed parameters for a forward filter, a reverse filter, or both. The seed value may be determined using a median signal value, a drift value, or both. Both a forward filter and a reverse filter may be employed, and the method may further include optimizing the seed value to minimize the mean squared error between the two signal filters. The method may further include adjusting the sensitivity and baseline of the sensors according to a signal-based calibration algorithm that uses the average of the signals from the forward and reverse filters together with the raw sensor signal. The method may further include adjusting the sensitivity and the baseline based on one or more criteria. The criteria may include that the average glucose value should be consistent with an expected diabetes mean. The criteria may include that the CGM glucose variability should be consistent with the average glucose level. The method may further include: detecting a sensor change amount; determining that the sensor change amount is above a threshold criterion; and preventing the display of a reading so that a potentially inaccurate reading is not displayed to the user.

[0035] In a fourteenth aspect, there is provided a method of compensating for drift in an analyte concentration sensor within a biological system using only signals from the analyte concentration sensor, the method comprising: measuring a value of an analyte using an implanted analyte concentration sensor; determining a first slow moving average of the measured values of the analyte over a first time period and calibrating the sensor at least in part based on the first slow moving average; after the first determination, determining a second slow moving average of the measured values of the analyte over a second time period; and adjusting the calibration of the sensor at least in part based on a seed value and based on a difference between the first slow moving average and the second slow moving average.

[0036] Embodiments may include one or more of the following. The seed value may be determined using a median signal value, a drift value, or both. Both a forward filter and a reverse filter may be employed, and the method may further include optimizing the seed value to minimize the mean square error between the two signal filters. The method may further include adjusting the sensitivity and baseline of the sensor according to a signal-based calibration algorithm that uses the average of the signals from the forward and reverse filters along with the original sensor signal. The method may further include adjusting the sensitivity and the baseline based on one or more criteria. The criteria may include that the average glucose value should be consistent with an expected diabetes mean. The criteria may include that the CGM glucose variability should be consistent with the average glucose level.

[0037] In further aspects and embodiments, the above method features of the various aspects are formulated in terms of a system configured to implement the method features in the various aspects. Any feature of an embodiment of any aspect, including but not limited to any embodiment of any of the first through fourteenth aspects mentioned above, applies to all other aspects and embodiments pointed out herein, including but not limited to any embodiment of any of the first through fourteenth aspects mentioned above. Moreover, any feature of an embodiment of the various aspects, including but not limited to any embodiment of any of the first through fourteenth aspects mentioned above, may be combined with other embodiments described herein in part or in whole in any manner, such as one, two, or three or more embodiments may be combined in whole or in part. Additionally, any feature of an embodiment of the various aspects, including but not limited to any embodiment of any of the first through fourteenth aspects mentioned above, may be optional with respect to other aspects or embodiments. Any aspect or embodiment of a method may be performed by the system or device of another aspect or embodiment, and any aspect or embodiment of a system or device may be configured to perform the method of another aspect or embodiment, including but not limited to any embodiment of any of the first through fourteenth aspects mentioned above.

[0038] This summary is provided to introduce a series of concepts in a simplified form. These concepts are further described in the detailed description section. Elements or steps other than those described in this summary are possible, and no element or step is necessarily required. This summary is not intended to identify key features or essential features of the claimed subject matter, nor is it intended to be used to assist in determining the scope of the claimed subject matter. The claimed subject matter is not limited to embodiments that solve any or all of the disadvantages mentioned in any part of this disclosure. BRIEF DESCRIPTION OF THE DRAWINGS

[0039] Embodiments of the present invention will now be described in detail with an emphasis on highlighting advantageous features. These embodiments depict novel and non - obvious sensor signal processing and calibration systems and methods shown in the accompanying drawings, which are for illustrative purposes only and not drawn to scale, and actually emphasize the principles of the present disclosure. These drawings include the following figures, where like reference numerals indicate like parts:

[0040] Figure 1 is a schematic diagram of a continuous analyte sensor system attached to a body and communicating with a plurality of exemplary devices;

[0041] Figure 2 illustrates in connection with Figure 1 the electronic devices associated with the sensor system of;

[0042] Figure 3 depicts a graph illustrating the linear relationship between measured sensor counts and analyte concentration.

[0043] Figure 4 illustrates an exemplary change in sensitivity over time.

[0044] Figure 5 illustrates various ways of providing factory information about the sensor to the transmitter electronics.

[0045] Figure 6 illustrates a "code - less" option for providing factory information about the sensor to the sensor electronics.

[0046] Figure 7 illustrates an option for providing factory information about the sensor to the sensor electronics without requiring explicit user input.

[0047] Figure 8 illustrates an option for providing factory information about the sensor to the sensor electronics with user input.

[0048] Figure 9 and 10 illustrate the steps of using a portion of a batch of sensors that have obtained field data to calibrate another portion of the batch of sensors.

[0049] Figure 11 is a flowchart illustrating an exemplary method in accordance with the principles of the present invention and particularly for performing the method in accordance with Figure 9 and 10 of;

[0050] Figure 12 is a schematic depiction of sensors and a transmitter within a body communicating with a receiver and / or a smart phone.

[0051] Figure 13is a flowchart illustrating another exemplary method in accordance with the principles of the present invention.

[0052] Figure 14 and 15 is a graph depicting analyte concentration over time before (14) and after (15) a change in sensitivity.

[0053] Figure 16 is a modular depiction of an analyte concentration measurement system in accordance with the principles of the present invention.

[0054] Figure 17 is a flowchart illustrating another exemplary method in accordance with the principles of the present invention.

[0055] Figure 18 is a flowchart illustrating another exemplary method of performing calibration using a steady state in accordance with the principles of the present invention.

[0056] Figure 19 is a graph illustrating two calibration lines before and after drift has occurred.

[0057] Figure 20 and 21 illustrates a slow moving average of sensor counts over time.

[0058] Figure 22 is a flowchart illustrating another exemplary method using a slow moving average in accordance with the principles of the present invention.

[0059] Figure 23A is a flowchart illustrating another exemplary method of updating historical values in accordance with the principles of the present invention.

[0060] Figure 23B is a chart showing sensitivity data in an extended sensor session, showing characteristic drift.

[0061] Figure 23C is a chart showing sensitivity data in an extended sensor session, showing characteristic drift along with failure modes.

[0062] Figure 24 is a flowchart illustrating another exemplary method in accordance with the principles of the present invention.

[0063] Figure 25 is a flowchart illustrating another exemplary method in accordance with the principles of the present invention.

[0064] Figure 26 is a flowchart illustrating another exemplary method in accordance with the principles of the present invention.

[0065] Figure 27 is a flowchart illustrating another exemplary method in accordance with the principles of the present invention.

[0066] Figure 28 is a graph showing the sensitivity distribution.

[0067] Figure 29 is a graph showing the baseline distribution.

[0068] Figure 30 is a graph showing the glucose value distribution (e.g., long-term glucose value).

[0069] Figure 31 is a graph showing the most likely sensitivity (slope) values, giving exemplary parameters.

[0070] Figure 32 is a graph showing the most likely baseline values, giving exemplary parameters.

[0071] Figure 33 is a graph showing the most likely glucose values, giving exemplary parameters.

[0072] Figure 34 and 35 illustrates an exemplary glucose trajectory. Figure 35 also illustrates the effect of a double-exponential filter operating on the glucose trajectory.

[0073] Figure 36 illustrates in the determination of Figure 34 and 35 the estimated drift curve of the sensor employed.

[0074] Figures 37 to 39 is an additional graph in which the drift correction according to the above principle is illustrated.

[0075] Figure 40 illustrates the measured relationship between the coefficient of signal variation and the standard deviation of glucose concentration values.

[0076] Figure 41 illustrates the measured glucose concentration signal over a time period.

[0077] Figure 42 illustrates the linear relationship between the coefficient of signal variation and the standard deviation of glucose concentration, with exemplary values marked.

[0078] Figure 43 shows the distribution of the difference between the measured standard deviation and the expected standard deviation.

[0079] Figure 44 shows the relationship between the average glucose and the glucose standard deviation.

[0080] Figure 45Shows the difference between glucose standard deviations among patient groups (i.e., non - diabetic, type I diabetic, and type II diabetic).

[0081] Figure 46 Shows data points separated by a time lag, where Δ represents the individual rate of change between two adjacent points.

[0082] Figure 47 Is a flowchart illustrating another embodiment of a method in accordance with the principles of the present invention. Detailed Description

[0083] The following description and examples detail some example embodiments of the disclosed invention. Those skilled in the art will recognize that the present invention has numerous variations and modifications covered by its scope. Thus, the description of a particular example embodiment should not be regarded as limiting the scope of the present invention.

[0084] Definition

[0085] To facilitate understanding of the preferred embodiments, several terms are defined below.

[0086] As used herein, the term "analyte" is a broad term and will give its ordinary and customary meaning to a person of ordinary skill in the art (and is not limited to a special or customized meaning), and further refers to (without limitation) substances or chemical constituents in an analyzable biological fluid (e.g., blood, interstitial fluid, cerebrospinal fluid, lymph fluid, or urine). Analytes can include naturally occurring substances, artificial substances, metabolites, and / or reaction products. In some embodiments, the analyte measured by the sensor head, device, and method is glucose. However, other analytes are also contemplated, including (but not limited to): carboxyprothrombin; acylcarnitine; adenine phosphoribosyltransferase; adenosine deaminase; albumin; alpha-fetoprotein; amino acid profiles (arginine (Krebs cycle), histidine / urocanic acid, homocysteine, phenylalanine / tyrosine, tryptophan); androstenedione; antipyrine; arabinitol enantiomers; arginase; biotinidase; biopterin; c-reactive protein; carnitine; carnosinase; CD4; ceruloplasmin; chenodeoxycholic acid; chloroquine; cholesterol; cholinesterase; conjugated 1-beta-hydroxy bile acid; cortisol; creatine kinase; creatine kinase MM isoenzyme; cyclosporine A; d-penicillamine; de-ethyl chloroquine; dehydroepiandrosterone sulfate; DNA (acetylation polymorphism, alcohol dehydrogenase, alpha1-antitrypsin, cystic fibrosis, Duchenne / Becker muscular dystrophy, analyte-6-phosphate dehydrogenase, hemoglobin A, hemoglobin S, hemoglobin C, hemoglobin D, hemoglobin E, hemoglobin F, D-Punjab, beta-thalassemia, hepatitis B virus, HCMV, HIV-1, HTLV-1, Leber hereditary optic neuropathy, MCAD, RNA, PKU, Plasmodium vivax, sex differentiation, 21-deoxycortisol); desbutyl halofantrine; dihydropteridine reductase; diphtheria / tetanus antitoxin; erythrocyte arginase; erythrocyte protoporphyrin; esterase D; fatty acid / acylglycine; free beta-human chorionic gonadotropin; free erythrocyte protoporphyrin; free thyroxine (FT4); free triiodothyronine (FT3); fumaroylacetoacetase; galactose / gal-1-phosphate; galactose-1-phosphate uridyltransferase; gentamicin; analyte-6-phosphate dehydrogenase; glutathione; glutathione peroxidase; glycocholic acid; glycosylated hemoglobin; halofantrine; hemoglobin variants; hexosaminidase A; human erythrocyte carbonic anhydrase I; 17-alpha-hydroxyprogesterone; hypoxanthine phosphoribosyltransferase; immunoreactive trypsin; lactate; lead; lipoproteins ((a), B / A-1, beta); lysozyme; mefloquine; netilmicin; phenobarbital; phenytoin; phytanic / pristanic acid; progesterone; prolactin; proline peptidase; purine nucleoside phosphorylase; quinine; reverse triiodothyronine (rT3); selenium; serum pancreatic lipase; sisomicin; somatomedin C;Specific antibodies (adenovirus, antinuclear antibody, anti-zeta antibody, arbovirus, pseudorabies virus, dengue virus, Dracunculus medinensis, Echinococcus granulosus, Entamoeba histolytica, enterovirus, Giardia duodenalis, Helicobacter pylori, hepatitis B virus, herpes virus, HIV-1, IgE (atopic diseases), influenza virus, Leishmania donovani, Leptospira interrogans, measles / mumps / rubella, Mycobacterium leprae, Mycoplasma pneumoniae, myoglobin, Onchocerca volvulus, parainfluenza virus, Plasmodium falciparum, poliovirus, Pseudomonas aeruginosa, respiratory syncytial virus, Rickettsia (scrub typhus), Schistosoma mansoni, Toxoplasma gondii, Treponema pallidum, Trypanosoma cruzi / Trypanosoma rangeli, vesicular stomatitis virus, Wuchereria bancrofti, yellow fever virus); specific antigens (hepatitis B virus, HIV-1); succinylacetone; sulfadoxine; theophylline; thyroid stimulating hormone (TSH); thyroxine (T4); thyroxine binding globulin; trace elements; transferrin; UDP-galactose-4-epimerase; urea; uroporphyrinogen I synthase; vitamin A; white blood cells; and zinc protoporphyrin. Salts, sugars, proteins, fats, vitamins, and hormones that occur naturally in blood or interstitial fluid may also constitute analytes in certain embodiments. Analytes may occur naturally in biological fluids such as metabolites, hormones, antigens, antibodies, and the like. Alternatively, analytes may be introduced into the body such as contrast agents for imaging, radioisotopes, chemical reagents, fluorocarbon-based artificial blood, or pharmaceutical or drug compositions, including (but are not limited to): insulin; ethanol; inhalants (nitrous oxide, amyl nitrite, butyl nitrite, chlorinated hydrocarbons, hydrocarbons); stimulants (amphetamine, methylphenidate, pimozide, phentermine, benzphetamine hydrochloride, PreState, clortermine hydrochloride, Sandrex, phendimetrazine); sedatives (barbiturates, methaqualone, tranquilizers such as diazepam tablets, chlordiazepoxide, meprobamate, oxazepam, carbutamide, Sanofi); anesthetics (codeine, morphine, pethidine, acetaminophen, hydrocodone combination, hydrocodone cough suppressant, fentanyl, Darvon, pentazocine, diphenoxylate tablets); chemical hallucinogens (fentanyl analogs, pethidine, amphetamine); anabolic steroids; and nicotine. Metabolites of pharmaceutical and drug compositions are also expected analytes. Analytes such as neurochemicals and other chemicals produced within the body may also be analyzed, such as ascorbic acid, uric acid, dopamine, norepinephrine, 3-methoxytyramine (3MT), 3,4-dihydroxyphenylacetic acid (DOPAC), homovanillic acid (HVA), 5-hydroxytryptamine (5HT), and 5-hydroxyindoleacetic acid (FHIAA).;

[0087] As used herein, the terms "microprocessor" and "processor" are broad terms and will give their ordinary and customary meaning to one of ordinary skill in the art (and are not limited to a special or customized meaning), and furthermore refer (without limitation) to computer systems, state machines, and the like that use logic circuits to perform arithmetic and logical operations, the logic circuits responding to and processing the basic instructions that drive the computer.

[0088] As used herein, the terms "raw data stream" and "data stream" are broad terms and will give their ordinary and customary meaning to one of ordinary skill in the art (and are not limited to a special or customized meaning), and furthermore refer (without limitation) to analog or digital signals directly related to measured glucose from a glucose sensor. In one example, the raw data stream is digital data in "counts" converted by an A / D converter from an analog signal (e.g., voltage or current), and includes one or more data points representing glucose concentration. The term broadly encompasses data points from multiple time intervals of a substantially continuous glucose sensor, which includes individual measurements taken over time intervals ranging from fractions of a second up to, for example, 1, 2, or 5 minutes or longer. In another example, the raw data stream includes integrated digital values, where the data includes one or more data points representing the glucose sensor signal averaged over a time period.

[0089] As used herein, the term "calibration" is a broad term and will give its ordinary and customary meaning to one of ordinary skill in the art (and are not limited to a special or customized meaning), and furthermore refer (without limitation) to the process of determining the relationship between sensor data and corresponding reference data, which can be used to transform the sensor data into a meaningful value that is substantially equivalent to the reference data, with or without real-time use of the reference data. In some embodiments, namely in continuous analyte sensors, calibration can be updated or recalibrated over time (at the factory, in real-time, and / or retrospectively) because the relationship between sensor data and reference data changes, for example due to changes in sensitivity, baseline, transportation, metabolism, and the like. Calibration can also be achieved by automatically estimating sensor signal parameters (auto-calibration) through the analysis of one or more signal characteristics or features.

[0090] As used herein, the terms "calibrated data" and "calibrated data stream" are broad terms and will give their ordinary and customary meaning to one of ordinary skill in the art (and are not limited to a special or customized meaning), and furthermore refer (without limitation) to data that has been transformed from its original state to another state using a function (e.g., a conversion function, including use of sensitivity) to provide a meaningful value to the user.

[0091] As used herein, the terms "smoothed data" and "filtered data" are broad terms and will give their ordinary and customary meaning to one of ordinary skill in the art (and are not limited to a special or customized meaning), and furthermore refer (without limitation) to data that has been modified to make it smoother and more continuous and / or to remove or reduce outliers, for example, by performing a moving average of the original data stream, including a slow moving average. Examples of data filters include FIR (finite impulse response), IIR (infinite impulse response), moving average filters, and like filters.

[0092] As used herein, the terms "smooth" and "filter" are broad terms and will give their ordinary and customary meaning to one of ordinary skill in the art (and are not limited to a special or customized meaning), and furthermore refer (without limitation) to modifying a data set to make it smoother and more continuous or to remove or reduce outliers, for example, by performing a moving average of the original data stream.

[0093] As used herein, the term "algorithm" is a broad term and will give its ordinary and customary meaning to one of ordinary skill in the art (and are not limited to a special or customized meaning), and furthermore refers (without limitation) to a computational process (e.g., a program) involved in transforming information from one state to another, for example, by using computer processing.

[0094] As used herein, the term "count" is a broad term and will give its ordinary and customary meaning to one of ordinary skill in the art (and are not limited to a special or customized meaning), and furthermore refers (without limitation) to a unit of measurement of a data signal. In one example, the original data stream measured in counts is directly related to a voltage (e.g., converted by an A / D converter), and the voltage is directly related to the current from a working electrode.

[0095] As used herein, the term "sensor" is a broad term and will give its ordinary and customary meaning to one of ordinary skill in the art (and are not limited to a special or customized meaning), and furthermore refers (without limitation) to a component or region of a device that can be used to quantify an analyte.

[0096] As used herein, the terms "glucose sensor" and "component for determining the amount of glucose in a biological sample" are broad terms and will give their ordinary and customary meaning to one of ordinary skill in the art (and are not limited to a special or customized meaning), and furthermore refer (without limitation) to any mechanism (e.g., enzymatic or non-enzymatic) that can be used to quantify glucose. For example, some embodiments utilize a thin film containing glucose oxidase, which catalyzes the conversion of oxygen and glucose to hydrogen peroxide and gluconate, as illustrated by the following chemical reaction:

[0097] Glucose + O2 → Gluconate + H2O2

[0098] Because for each glucose molecule metabolized, there is a proportional change in the co-reactant O2 and the product H2O2, an electrode can be used to monitor the change in current in the co-reactant or product to determine the glucose concentration.

[0099] As used herein, the terms "operatively connected" and "operatively linked" are broad terms and will give their ordinary and customary meaning to one of ordinary skill in the art (and are not limited to a special or customized meaning), and further refer (without limitation) to the linking of one or more components to another (or other) component in such a way as to allow the emission of signals between the components. For example, one or more electrodes can be used to detect the amount of glucose in a sample and convert the information into a signal, such as an electrical or electromagnetic signal; the signal can then be transmitted to an electronic circuit. In this case, the electrode is "operatively linked" to the electronic circuit. These terms are broad enough to encompass wireless connectivity.

[0100] The term "determine" encompasses a wide variety of actions. For example, "determine" can include operations, calculations, processing, derivations, investigations, lookups (e.g., looking up in a table, database, or other data structure), confirmations, and similar actions. Moreover, "determine" can include receiving (e.g., receiving information), accessing (e.g., accessing data in a memory), and similar actions. Moreover, "determine" can include parsing, selecting, picking, calculating, deriving, establishing, and / or similar actions. Determining can also include confirming that a parameter matches a predetermined criterion, including having been met, passed, exceeded a threshold, and the like.

[0101] As used herein, the term "substantially" is a broad term and will give its ordinary and customary meaning to one of ordinary skill in the art (and are not limited to a special or customized meaning), and further refer (without limitation) to a greater degree but not necessarily all of the specified content.

[0102] As used herein, the term "subject" is a broad term and will give its ordinary and customary meaning to one of ordinary skill in the art (and are not limited to a special or customized meaning), and further refer (without limitation) to a mammal, particularly a human.

[0103] As used herein, the term "continuous analyte (or glucose) sensor" is a broad term and will give its ordinary and customary meaning to one of ordinary skill in the art (and are not limited to a special or customized meaning), and further refer (without limitation) to a device that continuously or continuously measures the concentration of an analyte, for example, at time intervals ranging from fractions of a second up to, for example, 1, 2, or 5 minutes or longer. In one exemplary embodiment, the continuous analyte sensor is a glucose sensor such as that described in U.S. Patent No. 6,001,067, which is incorporated herein by reference in its entirety.

[0104] As used herein, the term "continuous analyte (or glucose) sensing" is a broad term and will give its ordinary and customary meaning to one of ordinary skill in the art (and is not limited to a special or customized meaning), and further refers (without limitation) to a period in which monitoring of an analyte is performed continuously or persistently, for example, at time intervals ranging from fractions of a second up to, for example, 1, 2, or 5 minutes or longer.

[0105] As used herein, the terms "reference analyte monitor", "reference analyte meter", and "reference analyte sensor" are broad terms and will give their ordinary and customary meaning to one of ordinary skill in the art (and is not limited to a special or customized meaning), and further refers (without limitation) to a device that measures the concentration of an analyte and can be used as a reference for a continuous analyte sensor, for example, a self - monitoring blood glucose meter (SMBG) can be used as a reference for a continuous glucose sensor for comparison, calibration, and similar operations.

[0106] As used herein, the term "sensing membrane" is a broad term and will give its ordinary and customary meaning to one of ordinary skill in the art (and is not limited to a special or customized meaning), and further refers (without limitation) to a permeable or semi - permeable membrane, which can include two or more domains and is typically constructed of a material having a thickness of several micrometers or more, the material being permeable to oxygen and optionally permeable to glucose. In one example, the sensing membrane includes immobilized glucose oxidase, which enables an electrochemical reaction to occur to measure the concentration of glucose.

[0107] As used herein, the term "physiologically viable" is a broad term and will give its ordinary and customary meaning to one of ordinary skill in the art (and is not limited to a special or customized meaning), and further refers (without limitation) to physiological parameters obtained from continuous studies of glucose data in humans and / or animals. By way of example, a maximum sustained rate of glucose change in humans of about 4 to 5 mg / dL / min and a maximum acceleration of the rate of change of about 0.1 to 0.2 mg / dL / min / min are considered the limits of physiological viability. Values outside these limits will be considered non - physiological and may be, for example, the result of a signal error. As another example, the rate of glucose change is lowest at the maximum and minimum values of the daily glucose range, which are the areas of greatest risk in patient treatment, and thus the physiologically viable rate of change can be set at the maximum and minimum values based on continuous studies of glucose data. As a further example, it has been observed that the optimal solution for the shape of the curve at any point along the glucose signal data stream over a certain time period (e.g., about 20 to 30 minutes) is a straight line, which can be used to set physiological limits.

[0108] As used herein, the term "frequency content" is a broad term and will give its ordinary and customary meaning to a person of ordinary skill in the art (and is not limited to a special or customized meaning), and furthermore refers to (without limitation) spectral density, including the frequencies contained in a signal and its power.

[0109] As used herein, the term "linear regression" is a broad term and will give its ordinary and customary meaning to a person of ordinary skill in the art (and is not limited to a special or customized meaning), and furthermore refers to (without limitation) finding a line such that a data set has a minimum measurement deviation or separation from the line. By-products of this algorithm include the slope, the y-intercept, and the R-squared value, which determine how well the measured data fits the line. In some cases, robust regression techniques can also be employed to handle outliers in the regression.

[0110] As used herein, the term "nonlinear regression" is a broad term and will give its ordinary and customary meaning to a person of ordinary skill in the art (and is not limited to a special or customized meaning), and furthermore refers to (without limitation) fitting a data set in a non-linear form to describe the relationship between a response variable and one or more explanatory variables.

[0111] As used herein, the term "mean" is a broad term and will give its ordinary and customary meaning to a person of ordinary skill in the art (and is not limited to a special or customized meaning), and furthermore refers to (without limitation) the sum of the observed values divided by the number of observed values.

[0112] As used herein, the term "non-recursive filter" is a broad term and will give its ordinary and customary meaning to a person of ordinary skill in the art (and is not limited to a special or customized meaning), and furthermore refers to (without limitation) an equation that uses a moving average as an input and an output.

[0113] As used herein, the terms "recursive filter" and "autoregressive algorithm" are broad terms and will give their ordinary and customary meaning to a person of ordinary skill in the art (and are not limited to a special or customized meaning), and furthermore refer to (without limitation) an equation in which a previous average is part of the next filtered output. More specifically, a series of observations, and thus the generation of the value of each observation, depends in part on the values of those observations immediately preceding it. An example is a regression structure in which lagged response values serve as independent variables.

[0114] As used herein, the term "variation" is a broad term and will give its ordinary and customary meaning to a person of ordinary skill in the art (and is not limited to a special or customized meaning), and furthermore refers to (without limitation) the divergence or amount of change from a point, line, or data set. In one embodiment, an estimated analyte value can have a variation that includes a range of values outside the estimated analyte value that represent a range of possibilities based on, for example, known physiological patterns.

[0115] As used herein, the terms "physiological parameter" and "physiological boundary" are broad terms and will give their ordinary and customary meaning to one of ordinary skill in the art (and are not limited to a special or customized meaning), and further refer (without limitation) to parameters obtained from continuous studies of physiological data of humans and / or animals. By way of example, the maximum sustainable rate of change of glucose in humans of about 6 to 8 mg / dL / min and the maximum acceleration of the rate of change of about 0.1 to 0.2 mg / dL / min 2 are considered the physiologically viable limits; values outside these limits will be considered non-physiological. As another example, the rate of change of glucose is lowest at the maximum and minimum values of the daily glucose range, which are the areas of greatest risk in patient treatment, and thus physiologically viable rates of change can be set at the maximum and minimum values based on continuous studies of glucose data. As a further example, it has been observed that the optimal solution for the shape of the curve at any point along the glucose signal data stream over a certain time period (e.g., about 20 to 30 minutes) is a straight line, which can be used to set physiological limits. These terms are broad enough to encompass physiological parameters for any analyte.

[0116] As used herein, the term "measured analyte value" is a broad term and will give its ordinary and customary meaning to one of ordinary skill in the art (and are not limited to a special or customized meaning), and further refer (without limitation) to the analyte value or set of analyte values during the time period in which the analyte sensor has measured analyte data. The term is broad enough to encompass data from the analyte sensor before or after data processing (e.g., data smoothing, calibration, and similar processes) in the sensor and / or receiver.

[0117] As used herein, the term "estimated analyte value" is a broad term and will give its ordinary and customary meaning to one of ordinary skill in the art (and are not limited to a special or customized meaning), and further refer (without limitation) to the analyte value or set of analyte values that have been extrapolated algorithmically from the measured analyte values.

[0118] As used herein, the term "sensor data" is a broad term and will give its ordinary and customary meaning to a person of ordinary skill in the art (and is not limited to a special or customized meaning), and further refers to (without limitation) any data associated with a sensor such as a continuous analyte sensor. Sensor data includes the raw data stream or simply the data stream of analog or digital signals (or other signals received from another sensor) directly related to the measured analyte from an analyte sensor, as well as the calibrated and / or filtered raw data. In one example, the sensor data includes digital data in "counts" converted by an A / D converter from an analog signal (e.g., voltage or current), and includes one or more data points representing glucose concentration. Thus, the terms "sensor data point" and "data point" generally refer to the digital representation of sensor data at a particular time. The terms broadly encompass data points from a sensor, such as from a substantially continuous glucose sensor, over a plurality of time intervals, which include individual measurements taken over time intervals ranging from fractions of a second up to, for example, 1, 2, or 5 minutes or longer. In another example, the sensor data includes integrated digital values that represent one or more data points averaged over a time period. Sensor data can include calibrated data, smoothed data, filtered data, transformed data, and / or any other data associated with the sensor.

[0119] As used herein, the term "matched data pair" or "data pair" is a broad term and will give its ordinary and customary meaning to a person of ordinary skill in the art (and is not limited to a special or customized meaning), and further refers to (without limitation) reference data (e.g., one or more reference analyte data points) that matches sensor data (e.g., one or more sensor data points) that is substantially time-corresponding.

[0120] As used herein, the term "sensor electronics" is a broad term and will give its ordinary and customary meaning to a person of ordinary skill in the art (and is not limited to a special or customized meaning), and refers to (without limitation) components of a device (e.g., hardware and / or software) configured to process data.

[0121] As used herein, the term "calibration set" is a broad term and will give its ordinary and customary meaning to a person of ordinary skill in the art (and is not limited to a special or customized meaning), and refers to (without limitation) a data set that includes information for calibration. In some embodiments, the calibration set is formed from one or more matched data pairs that are used to determine the relationship between reference data and sensor data; however, other pre-implanted data derived externally or internally can also be used. As another example, data can also be adopted from a previous sensor session of the subject user.

[0122] As used herein, the term "sensitivity" or "sensor sensitivity" is a broad term and will give its ordinary and customary meaning to one of ordinary skill in the art (and is not limited to a special or customized meaning), and refers to (without limitation) the amount of signal generated by measuring a certain concentration of an analyte or a measured substance (e.g., H2O2) associated with the measured analyte (e.g., glucose). For example, in one embodiment, the sensor has a sensitivity of from about 1 to about 300 picoamperes of current for each 1 mg / dL of glucose analyte.

[0123] As used herein, the term "sensitivity profile" or "sensitivity curve" is a broad term and will give its ordinary and customary meaning to one of ordinary skill in the art (and is not limited to a special or customized meaning), and refers to (without limitation) a representation of the change in sensitivity over time.

[0124] Other definitions will be provided within the following description and in some cases from the context in which the term is used.

[0125] As used herein, the following abbreviations apply: Eq and Eqs (equivalent); mEq (milliequivalent); M (mole); mM (millimole); μM (micromole); N (normal); mol (gram-molecule); mmol (milligram-molecule); μmol (microgram-molecule); nmol (nanogram-molecule); g (gram); mg (milligram); μg (microgram); Kg (kilogram); L (liter); mL (milliliter); dL (deciliter); μL (microliter); cm (centimeter); mm (millimeter); μm (micrometer); nm (nanometer); h and hr (hour); min. (minute); s and sec. (second); °C (degree Celsius).

[0126] Overview / General Description of the System

[0127] Conventional in vivo continuous analyte sensing techniques typically rely on reference measurements performed during a sensor session for calibration of a continuous analyte sensor. The reference measurements are matched with sensor data that is generally time-corresponding to create matched data pairs. A regression is then performed on the matched data pairs (e.g., by using least squares regression) to produce a conversion function that defines the relationship between the sensor signal and the estimated glucose concentration.

[0128] In an intensive care unit, calibration of a continuous analyte sensor is often performed by using a calibration solution with a known concentration of the analyte as a reference. This calibration procedure can be cumbersome because a calibration bag that is separate from (and added to) an IV (intravenous) bag is typically used. In an outpatient ward, calibration of a continuous analyte sensor has traditionally been performed by capillary blood glucose measurement (e.g., finger stick glucose testing), whereby reference data is obtained and input into the continuous analyte sensor system. This calibration procedure typically involves frequent finger stick measurements, which can be inconvenient and painful.

[0129] To date, systems and methods for in vitro calibration (e.g., factory calibration) of continuous analyte sensors by manufacturers without relying on periodic recalibration have mostly been insufficient for the high levels of sensor accuracy required for blood glucose management. This can be attributed in part to changes in sensor properties (e.g., sensor sensitivity) that occur during sensor use. Thus, calibration of continuous analyte sensors has typically involved periodic input of reference data, whether associated with a calibration solution or finger stick measurements. This can be very cumbersome for users in daily life and for patients in outpatient wards or medical staff in intensive care units.

[0130] The following description and examples describe embodiments of the present invention with reference to the accompanying drawings. In the drawings, reference numerals label elements of embodiments of the present invention. These reference numerals are reproduced below in connection with the discussion of the corresponding drawing features.

[0131] Described herein are systems and methods for calibrating a continuous analyte sensor that can achieve a high level of accuracy without relying (or with reduced reliance) on reference data from a reference analyte monitor (e.g., from a blood glucose meter).

[0132] Sensor System

[0133] Figure 1 Depicted is an example system 100 according to some example embodiments. System 100 includes a continuous analyte sensor system 8, which includes sensor electronics 12 and a continuous analyte sensor 10. System 100 can include other devices and / or sensors, such as a drug delivery pump 2 and a blood glucose meter 4. The continuous analyte sensor 10 can be physically connected to the sensor electronics 12 and can be integral with (e.g., non - releasably attached to) or releasably attached to the continuous analyte sensor 10. The sensor electronics 12, the drug delivery pump 2, and / or the blood glucose meter 4 can be coupled to one or more devices such as display devices 14, 16, 18, and / or 20.

[0134] In some example embodiments, system 100 may include a cloud-based analyte processor 490 configured to analyze analyte data (and / or other patient-related data) associated with a subject (also referred to as a patient) provided via network 406 (e.g., via wired, wireless, or a combination thereof) from sensor system 8 and other devices (such as display devices 14 to 20 and the like), and generate a report providing high-level information (e.g., statistics) regarding the measured analyte over a certain time period. A complete discussion of using a cloud-based analyte processing system can be found in U.S. Patent Application No. 13 / 788,375, filed on March 7, 2013, titled "Cloud-Based Processing of Analyte Data," which is incorporated herein by reference in its entirety.

[0135] In some example embodiments, sensor electronics 12 may include electronic circuitry associated with measuring and processing data generated by continuous analyte sensor 10. This generated continuous analyte sensor data may also include algorithms that can be used to process and calibrate the continuous analyte sensor data, although these algorithms may also be provided in other ways. Sensor electronics 12 may include hardware, firmware, software, or a combination thereof to provide a measurement of the level of an analyte via a continuous analyte sensor (such as a continuous glucose sensor). Example embodiments of sensor electronics 12 are further described below with respect to Figure 2 Further description of sensor electronics 12.

[0136] Sensor electronics 12 as described may be coupled (e.g., wirelessly and the like) to one or more devices such as display devices 14, 16, 18, and / or 20. Display devices 14, 16, 18, and / or 20 may be configured to present information (and / or alerts) such as sensor information transmitted by sensor electronics 12 for display at display devices 14, 16, 18, and / or 20.

[0137] Display devices may include a relatively small keychain-like display device 14, a relatively large handheld display device 16, a cellular phone 18 (e.g., a smart phone, a tablet computer, and the like), a computer 20, and / or any other user device configured to present at least information (such as drug delivery information, discrete self-monitoring glucose readings, heart rate monitors, calorie intake monitors, and the like).

[0138] In some example embodiments, a relatively small keychain-like display device 14 can include a wristwatch, a belt, a necklace, a charm, a piece of jewelry, a patch, a pager, a keychain, a plastic card (e.g., a credit card), an identification (ID) card, and / or the like. This small display device 14 can include a relatively small display (e.g., smaller than the large display device 16), and can be configured to display certain types of displayable sensor information, such as numerical values and arrows or color codes.

[0139] In some example embodiments, a relatively large handheld display device 16 can include a handheld receiver device, a personal digital assistant, and / or the like. This large display device can include a relatively large display (e.g., larger than the small display device 14), and can be configured to display information, such as a graphical representation of continuous sensor data including current and historical sensor data output by the sensor system 8.

[0140] In some example embodiments, the continuous analyte sensor 10 includes a sensor for detecting and / or measuring an analyte, and the continuous analyte sensor 10 can be configured to continuously detect and / or measure the analyte as a non-invasive device, a subcutaneous device, a transdermal device, and / or an intravascular device. In some example embodiments, the continuous analyte sensor 10 can analyze multiple intermittent blood samples, but other analytes can also be used.

[0141] In some example embodiments, the continuous analyte sensor 10 can include a glucose sensor configured to measure glucose in blood or interstitial fluid using one or more measurement techniques, such as enzymatic, chemical, physical, electrochemical, spectrophotometric, polarization, calorimetric, iontophoresis, radiometric, immunochemical, and similar techniques. In embodiments where the continuous analyte sensor 10 includes a glucose sensor, the glucose sensor includes any device capable of measuring glucose concentration, and multiple techniques can be used to measure glucose, including invasive, minimally invasive, and non-invasive sensing techniques (e.g., fluorescence monitoring), to provide data indicative of the glucose concentration in the subject, such as a data stream. The data stream can be sensor data (raw and / or filtered), which can be converted to a calibrated data stream for providing glucose values to the subject, such as a user, a patient, or a caregiver (e.g., a parent, a relative, a guardian, a teacher, a doctor, a nurse, or any other individual concerned with the health of the subject). Additionally, the continuous analyte sensor 10 can be implanted as at least one of the following types of sensors: an implantable glucose sensor, a transdermal glucose sensor, implanted in a subject's blood vessel or outside the body, a subcutaneous sensor, a refillable subcutaneous sensor, an intravascular sensor.

[0142] Although the disclosure herein relates to some embodiments of a continuous analyte sensor 10 that includes a glucose sensor, the continuous analyte sensor 10 can also include other types of analyte sensors. Moreover, although some embodiments relate to a glucose sensor as an implantable glucose sensor, other types of devices that are capable of detecting glucose concentration and providing an output signal indicative of the glucose concentration can also be used. Additionally, although the description herein relates to glucose as the analyte being measured, processed, and otherwise operated on, other analytes can also be used, including, for example, ketone bodies (e.g., acetone, acetoacetate, and β-hydroxybutyrate, lactate, etc.), glucagon, acetyl coenzyme A, triglycerides, fatty acids, interstitial in the citric acid cycle, choline, insulin, cortisol, testosterone, and the like.

[0143] Figure 2 An example of sensor electronics 12 is depicted in accordance with some example embodiments. The sensor electronics 12 can include sensor electronics configured to process sensor information such as sensor data and generate transformed sensor data and displayable sensor information, for example, via a processor module. For example, the processor module can transform the sensor data into one or more of the following: filtered sensor data (e.g., one or more filtered analyte concentration values), raw sensor data, calibrated sensor data (e.g., one or more calibrated analyte concentration values), rate of change information, trend information, acceleration / deceleration rate information, sensor diagnostic information, location information, alert / warning information, calibration information, smoothing and / or filtering algorithms for sensor data, and / or similar data.

[0144] In some embodiments, the processor module 214 is configured to implement most, if not all, of the data processing. The processor module 214 can be integral with the sensor electronics 12 and / or can be remotely located, for example, in one or more of the devices 14, 16, 18, and / or 20 and / or the cloud 490. In some embodiments, the processor module 214 can include a plurality of smaller sub-components or sub-modules. For example, the processor module 214 can include an alert module (not shown) or a prediction module (not shown), or any other suitable module that can be used to effectively process the data. When the processor module 214 is composed of multiple sub-modules, the sub-modules can be located within the processor module 214, including within the sensor electronics 12 or other associated devices (e.g., 14, 16, 18, 20, and / or 490). For example, in some embodiments, the processor module 214 can be at least partially located within the cloud-based analyte processor 490 or elsewhere in the network 406.

[0145] In some example embodiments, the processor module 214 may be configured to calibrate sensor data, and the data storage memory 220 may store the calibrated sensor data points as transformed sensor data. Also, in some example embodiments, the processor module 214 may be configured to wirelessly receive calibration information from display devices such as devices 14, 16, 18, and / or 20 to effect calibration of the sensor data from the sensor 12. Additionally, the processor module 214 may be configured to perform additional algorithmic processing on the sensor data (e.g., calibrated and / or filtered data and / or other sensor information), and the data storage memory 220 may be configured to store the transformed sensor data associated with the algorithm and / or sensor diagnostic information.

[0146] In some example embodiments, the sensor electronics 12 may include an application specific integrated circuit (ASIC) 205 coupled to the user interface 222. The ASIC 205 may further include a potentiostat 210, a telemetry module 232 for transmitting data from the sensor electronics 12 to one or more devices (such as devices 14, 16, 18, and / or 20), and / or other components for signal processing and data storage (e.g., the processor module 214 and the data storage memory 220). Although Figure 2 the ASIC 205 is depicted, other types of circuitry may be used, including field programmable gate arrays (FPGAs), one or more microprocessors configured to provide some (if not all) of the processing performed by the sensor electronics 12, analog circuitry, digital circuitry, or combinations thereof.

[0147] In Figure 2 the example depicted, the potentiostat 210 is coupled to a continuous analyte sensor 10, such as a glucose sensor, to generate sensor data from the analyte. The potentiostat 210 may also provide a voltage via the data line 212 to the continuous analyte sensor 10 to bias the sensor for measurement of values (such as current and the like) indicative of the analyte concentration in the subject (also referred to as the analog portion of the sensor). Depending on the number of working electrodes at the continuous analyte sensor 10, the potentiostat 210 may have one or more channels.

[0148] In some example embodiments, potentiostat 210 may include a resistor that translates the current value from sensor 10 into a voltage value, and in some example embodiments, a current-to-frequency converter (not shown) may also be configured to continuously integrate the measured current value from sensor 10 using, for example, a charge counting device. In some example embodiments, an analog-to-digital converter (not shown) may digitize the analog signal from sensor 10 into so-called "counts" to allow processing by processor module 214. The resulting counts may be directly related to the current measured by potentiostat 210, which may be directly related to the analyte level in the subject, such as glucose level.

[0149] Telemetry module 232 may be operably connected to processor module 214 and may provide the hardware, firmware, and / or software to enable wireless communication between sensor electronics 12 and one or more other devices, such as display devices, processors, network access devices, and the like. A variety of wireless radio technologies that may be implemented in telemetry module 232 include Bluetooth, Bluetooth Low Energy, ANT, ANT+, ZigBee, IEEE 802.11, IEEE 802.16, cellular radio access technologies, radio frequency (RF), infrared (IR), paging network communication, magnetic induction, satellite data communication, spread spectrum communication, frequency hopping communication, near field communication, and / or similar technologies. In some example embodiments, telemetry module 232 includes a Bluetooth chip, but Bluetooth technology may also be implemented in a combination of telemetry module 232 and processor module 214.

[0150] Processor module 214 may control the processing performed by sensor electronics 12. For example, processor module 214 may be configured to process data (e.g., counts) from the sensor, filter the data, calibrate the data, perform self-diagnostic checks, and / or similar operations.

[0151] In some example embodiments, the processor module 214 may include a digital filter, such as an infinite impulse response (IIR) or finite impulse response (FIR) filter. This digital filter may smooth the raw data stream received from the sensor 10. Generally, the digital filter is programmed to filter data sampled at a predetermined time interval (also referred to as the sample rate). In some example embodiments, such as when the potentiostat 210 is configured to measure an analyte (e.g., glucose and / or the like) at discrete time intervals, these time intervals determine the sampling rate of the digital filter. In some example embodiments, the potentiostat 210 may be configured to continuously measure an analyte, for example, using a current-to-frequency converter. In these current-to-frequency converter embodiments, the processor module 214 may be programmed to request digital values from the integrator of the current-to-frequency converter at a predetermined time interval (acquisition time). These digital values obtained by the processor module 214 from the integrator may be averaged over the acquisition time due to the continuity of the current measurement. Thus, the acquisition time may be determined by the sampling rate of the digital filter. Other uses of the FIR filter are described in more detail below.

[0152] The processor module 214 may further include a data generator (not shown) configured to generate data packets for transmission to devices such as display devices 14, 16, 18, and / or 20. Additionally, the processor module 214 may generate data packets for transmission to these external sources via the telemetry module 232. In some example embodiments, the data packets may be customized for each display device as described, and / or may include any available data, such as timestamps, displayable sensor information, transformed sensor data, identifier codes for the sensor and / or sensor electronics 12, raw data, filtered data, calibrated data, rate-of-change information, trend information, error detection or correction, and / or similar data.

[0153] The processor module 214 may also include a program memory 216 and other memory 218. The processor module 214 may be coupled to a communication interface such as communication port 238 and a power source such as battery 234. Moreover, the battery 234 may be further coupled to a battery charger and / or regulator 236 to provide power to the sensor electronics 12 and / or charge the battery 234.

[0154] The program memory 216 may be implemented as a semi-static memory for storing data such as identifiers (e.g., sensor identifier (ID)) of the coupled sensors 10 and for storing code (also referred to as program code) to configure the ASIC 205 to perform one or more of the operations / functions described herein. For example, the program code may configure the processor module 214 to process data streams or counts, perform filtering, execute the calibration methods described below, perform self-checks and the like.

[0155] The memory 218 may also be used to store information. For example, the processor module 214 that includes the memory 218 may be used as a cache memory for the system, where temporary storage is provided for the most recent sensor data received from the sensors. In some example implementations, the memory may include memory storage components such as read only memory (ROM), random access memory (RAM), dynamic RAM, static RAM, non-static RAM, electrically erasable programmable read only memory (EEPROM), rewritable ROM, flash memory, and the like.

[0156] The data storage memory 220 may be coupled to the processor module 214 and may be configured to store a variety of sensor information. In some example implementations, the data storage memory 220 stores continuous analyte sensor data for one or more days. For example, the data storage memory may store 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 20, and / or 30 (or more days) of continuous analyte sensor data received from the sensors 10. The stored sensor information may include one or more of the following: timestamps, raw sensor data (one or more raw analyte concentration values), calibrated data, filtered data, transformed sensor data, and / or any other displayable sensor information, calibration information (e.g., reference BG values and / or previous calibration information), sensor diagnostic information, and the like.

[0157] The user interface 222 may include a variety of interfaces, such as one or more buttons 224, a liquid crystal display (LCD) 226, a vibrator 228, an audio transducer (e.g., a speaker) 230, a backlight (not shown), and / or the like. Components including the user interface 222 may provide controls for interacting with a user (e.g., a subject). One or more buttons 224 may allow, for example, bistable triggering, menu selection, option selection, status selection, yes / no responses to on-screen questions, a "disconnect" function (e.g., for an alarm), a "confirm" function (e.g., for an alarm), reset, and / or similar functions. The LCD 226 may provide, for example, visual data output to the user. The audio transducer 230 (e.g., a speaker) may provide an audible signal in response to the triggering of certain alerts, such as current and / or predicted hyperglycemic and hypoglycemic conditions. In some example embodiments, the audible signal may be differentiated by tone, volume, duty cycle, pattern, duration, and / or the like. In some example embodiments, the audible signal may be configured to be silenced (e.g., confirmed or disconnected) by pressing one or more buttons 224 on the sensor electronics 12 and / or by using buttons on a display device (e.g., a key fob, a cell phone, and / or the like) or selecting to signal the sensor electronics 12.

[0158] Although audio and vibration alarms are described, other alarm mechanisms may also be used. For example, in some example embodiments, a tactile alarm is provided, including a poking mechanism configured to "poke" or physically contact the patient in response to one or more alarm conditions. Figure 2

[0159] The battery 234 may be operatively connected to the processor module 214 (and possibly other components of the sensor electronics 12) and provide the necessary power for the sensor electronics 12. In some example embodiments, the battery is a lithium manganese dioxide battery, although any suitable size and power battery may be used (e.g., AAA, nickel cadmium, zinc carbon, alkaline, lithium, nickel metal hydride, lithium ion, zinc air, zinc mercury oxide, silver zinc, or hermetically sealed batteries). In some example embodiments, the battery is rechargeable. In some example embodiments, multiple batteries may be used to power the system. In still other embodiments, the receiver may be powered transcutaneously, for example, via inductive coupling.

[0160] The battery charger and / or regulator 236 can be configured to receive energy from an internal and / or external charger. In some example embodiments, the battery regulator (or balancer) 236 regulates the recharge process by discharging excess charging current to allow all cells or batteries in the sensor electronics 12 to be fully charged without overcharging other cells or batteries. In some example embodiments, the battery 234 (or batteries) is configured to be charged via an inductive and / or wireless charging pad, but any other charging and / or power mechanism may also be used.

[0161] One or more communication ports 238, also referred to as external connectors, may be provided to allow communication with other devices. For example, a PC communication (com) port may be provided to enable communication with a system separate from or integral with the sensor electronics 12. The communication port may include, for example, a serial (e.g., Universal Serial Bus or “USB”) communication port and allow communication with another computer system (e.g., a PC, personal digital assistant or “PDA”, server, or the like). In some example embodiments, the sensor electronics 12 is capable of transmitting historical data to a PC or other computing device (e.g., an analyte processor as disclosed herein) for retrospective analysis by a patient and / or physician. As another example of data transmission, factory information may also be sent from the sensor or from a cloud data source to an algorithm.

[0162] In some continuous analyte sensor systems, the on-skin portion of the sensor electronics may be simplified to minimize the complexity and / or size of the on-skin electronics, e.g., by providing only raw, calibrated, and / or filtered data to a display device configured to run calibration and other algorithms required to display the sensor data. However, the sensor electronics 12 (e.g., via the processor module 214) may be implemented to perform prospective algorithms for generating transformed sensor data and / or displayable sensor information, including, for example, the following algorithms: evaluating the clinical acceptability of reference and / or sensor data, evaluating calibration data for optimal calibration based on inclusion criteria, evaluating the quality of calibration, comparing estimated analyte values with measured analyte values corresponding in time, analyzing changes in estimated analyte values, evaluating the stability of the sensor and / or sensor data, detecting signal artifacts (noise), replacing signal artifacts, determining the rate of change and / or trend of sensor data, performing dynamic and intelligent analyte value estimation, performing diagnostics on the sensor and / or sensor data, setting an operating mode, evaluating whether the data is abnormal, and / or similar algorithms.

[0163] Although Figure 2 separate data storage and program memories are shown, various configurations may also be used. For example, one or more memories may be used to provide storage space to support the data processing and storage requirements at the sensor electronics 12.

[0164] Calibration

[0165] While some continuous glucose sensors rely on (and use the accuracy of) BG values and / or factory-derived information for calibration, the disclosed embodiments utilize real-time information (e.g., in some embodiments, exactly the sensor data itself) to determine aspects of calibration and perform calibration based thereon.

[0166] In some cases, calibration of an analyte sensor can use prior calibration distribution information. For example, in some embodiments, prior calibration distribution information or code can be received as information from a previous calibration and / or sensor session (e.g., the same sensor system, stored internally), stored in memory, written at the factory (e.g., as part of the factory settings), on a package barcode, sent from a cloud or network of remote servers, written by a caregiver or user, received from another sensor system or electronic device, based on the results from a laboratory test, and / or the like.

[0167] As used herein, prior information includes information obtained prior to a particular calibration. By way of example, prior calibration from a particular sensor session (e.g., feedback from a prior calibration), information obtained prior to sensor insertion (e.g., factory information from in vitro testing or data obtained from a previously implanted analyte concentration sensor, such as sensors in the same manufacturing lot and / or sensors from one or more different lots), prior in vivo testing of similar sensors on the same subject, and / or prior in vivo testing of similar sensors on different subjects. Calibration information includes information for calibrating a continuous glucose sensor, such as but not limited to: sensitivity (m); sensitivity change (Δdm / dt), which may also relate to sensitivity drift; rate of change of sensitivity (ddm / ddt); baseline / intercept (b); baseline change (Δdb / dt); rate of change of baseline (ddb / ddt); baseline and / or sensitivity distribution associated with the sensor (i.e., change over a time period); linearity; response time; relationship between properties of the sensor (e.g., relationship between sensitivity and baseline); or relationship between a particular stimulus signal output (e.g., output indicative of impedance, capacitance, or other electrical or chemical property of the sensor) and sensor sensitivity or temperature (e.g., determined from prior in vivo and / or in vitro studies), such as described in U.S. Patent Publication 2012-0265035-A1, which is incorporated herein by reference in its entirety; sensor data obtained from a previously implanted analyte concentration sensor; calibration code associated with the sensor being calibrated; patient-specific relationship between the sensor and sensitivity, baseline, drift, impedance, impedance / temperature relationship (e.g., determined from prior studies of the patient or other patients sharing common characteristics with the patient); site of sensor implantation (abdomen, arm, etc.); and / or particular relationship (different sites may have different vascular densities). Distribution information includes ranges, distribution functions, distribution parameters (mean, standard deviation, skewness, etc.), generalized functions, statistical distributions, distributions, or the like, which represent multiple possible values of the calibration information. Prior calibration distribution information together includes the range or distribution of values provided prior to a particular calibration process for calibrating a sensor (e.g., sensor data) (e.g., describing its associated probability, probability density function, likelihood, or frequency of occurrence).

[0168] By way of example, in some embodiments, prior calibration distribution information includes a probability distribution based on, for example, sensitivity (m) or sensitivity-related information and baseline (b) or baseline-related information, based on the sensor type. As described above, the prior distribution of sensitivity and / or baseline may be factory-derived (e.g., from in vitro or in vivo testing of representative sensors) or derived from a prior calibration.

[0169] As described above, an analyte sensor generally includes electrodes to monitor changes in current in a co-reactant or product to determine analyte concentration, such as glucose concentration. In one example, sensor data includes digital data in "counts" converted from an analog signal (e.g., voltage or current) by an A / D converter. Calibration is the process of determining the relationship between the measured sensor signal in counts and the analyte concentration in clinical units. For example, calibration allows a given sensor measurement in counts to be associated with a measured analyte concentration value, e.g., in mg / dL. See Figure 3 of the curve Figure 10 , this relationship is generally a linear relationship of the form y = mx + b, where 'y' is the sensor signal in counts (y-axis 12), 'x' is the clinical value of the analyte concentration (axis 14), and'm' is the sensor sensitivity with units of [counts / (mg / dL)]. Line 16 is illustrated, and its slope is referred to as the sensor sensitivity. 'b' (see line segment 15) is the baseline sensor signal, which may be considered, or for advanced sensors can generally be reduced to zero or near zero; in any event, in many cases, the baseline can be assumed to be small or capable of being compensated in a predictable manner. In some embodiments, a constant background signal is seen, and this is modeled by y = m(x + c), where c is the glucose offset between the sensor site and the blood glucose.

[0170] Once line 16 has been determined, the system can convert the measured number of counts (or amperes as described above, e.g., picoamperes) to the clinical value of the analyte concentration.

[0171] However, the values of m and b vary between sensors and need to be determined. Additionally, the slope value m is not always constant. For example, and see Figure 4 , it can be seen that over the course of time within a session, the value of m changes from an initial sensitivity value m0 to a final sensitivity value m F . It is seen that the rate of change is greatest in the first few days of use, and this rate of change is referred to as m R .

[0172] The slope changes over time in vivo for several reasons. Specifically for the initial changes in calibration, this is often due to the sensor membrane "settling in" and achieving equilibrium with the in vivo environment. Sensors are generally calibrated in vitro or on a bench, and efforts are made to make the in vitro environment as close as possible to the in vivo environment, but the differences are still significant, and the in vivo environment itself varies between users. Additionally, sensors can vary due to differences in sterilization or shelf-life / storage conditions. Calibration changes that occur later in a session are often due to changes in the tissue surrounding the sensor, such as the accumulation of biofilm on the sensor.

[0173] Regardless of the cause, certain effects of variability have been measured and determined. For example, it is known that the variability of the final sensitivity m F is the largest contributing factor to the overall sensor inaccuracy. Similarly, it is known that the variability of the initial sensitivity m0 and physiology are the largest contributing factors to the inaccuracy of the sensor on the first day.

[0174] Due to the above variability, the initial calibration step involves determining and using seed values for one or more calibration parameters until additional data is obtained to adjust the seed values to more accurate values.

[0175] Once calibration is achieved, the sensor and analyte concentration measurement system can be employed to accurately determine the clinical value of the analyte concentration in the user's body. This can then lead to the discrimination of other sensor performance, including the determination of errors and drifts in sensitivity, as described in more detail below.

[0176] The most common current calibration method is to use an external blood glucose meter. This is commonly referred to as "finger stick calibration" and is a well-known and common part of life for many diabetics. The advantage of this technique is that it does not require a large amount of factory information and further provides a low risk of outliers. The disadvantages are that it requires a large amount of user participation and the need to know certain other factory calibration information required for proper calibration. Since the measured values from these meters are trusted, once the meter itself is calibrated, the values from the meter can be used to calibrate the implanted analyte concentration meter. Even as Figure 4 seen in the sensor sensitivity changes over time, the changing sensitivity is also not important in cases where the user wishes to perform numerous external calibrations.

[0177] However, users generally do not wish to perform numerous such calibrations, and in many cases, such as patients with type II diabetes, pre-diabetes, or even non-diabetic patients, the additional accuracy provided by such calibrations is not strictly required. For example, it may be sufficient for the user to know the range they are in rather than the exact analyte concentration value. In another embodiment, associated confidence intervals can be provided with the data to let the user know how much confidence is placed in the displayed data.

[0178] Therefore, efforts have been made to reduce the number of calibrations. However, many current CGM systems still require blood glucose values for at least initial calibration and also often require blood glucose values when dosing quantitatively. The systems and methods of the present invention in accordance with the principles of the present invention are in part directed to ways of reducing or eliminating such required calibrations.

[0179] A simple and convenient way to provide some level of calibration is to use calibration information about the relevant sensor. Even if the calibration information is approximate, this may still be sufficient for use by certain groups of patients. For example, and referring to Figure 5Process Figure 18 If factory calibration of one or more sensors within a manufacturing batch is known, this information can be provided to other sensors within the manufacturing batch that have not yet been used in a patient (step 20). This step is often referred to as providing a "code" to the transmitter because the code is often used as part of the CGM system to identify the manufacturing batch (and thus details) of the sensor when the sensors are coupled together during insertion into the patient. The transmitter can then identify the manufacturing batch from the code and apply appropriate calibration parameters according to a look-up table or other technique. However, it should be understood that the code can be provided not only to the transmitter, but also to any device in which the count can be converted to clinical units, such as a dedicated receiver, an off-the-shelf device that can be used to receive and display analyte concentration, such as a smart phone, a tablet computer, or the like. Additionally, the code can not be a code in the typical sense, but can simply provide any identifier to any device that requires the code for calibration purposes.

[0180] Referring again to Figure 5 describes ways of implementing the step (step 20) of providing factory information about the sensor to the transmitter or other electronics. Some of these ways constitute options for providing calibration information without using a code (step 22). Others include steps of providing the code using user data input (step 28). In some cases, the code can be entered without user input (step 24). Still other ways of providing the code can be used (step 26), and these additional ways are also described below. The details of these methods are now described.

[0181] Referring to Figure 6 Process Figure 30 describes an example of providing factory calibration information without a code (step 22). The first way is to use a representative value (e.g., average or median) of the manufacturing batch or other measure (e.g., range) (step 34). That is, if the average of the manufacturing batch is known, or even if the averages of different manufacturing batches made using the same technology are known, it can be assumed that the sensor calibration parameters will be similar and can thus be used as part of the calibration of the new sensor. Alternatively, predictable relationships can also be used to interpolate the sensor calibration parameters, such as bracketing the sensor batch.

[0182] As another example, impedance measurements can be employed in the determination of calibration parameters (step 40).

[0183] Calibration can also occur using information about a previous calibration (step 38). For example, if the user just turns off a sensor that was calibrated and measures the user's glucose concentration at, for example, 120 mg / dL, then it can be assumed that the proper measurement of the newly inserted sensor should result in the user's glucose value being 120 mg / dL again. In some cases, if a predicted glucose value has been determined, then the predicted value can be used for the newly inserted sensor. Even if a short time period has elapsed between the last reading of the old sensor and the reading of the new sensor, the identification of physiologically plausible glucose changes will result in bounds on what the new measurement value can be, and thus bounds on what the calibration parameters of the new sensor can be.

[0184] In yet another variation, various self - calibration algorithms can be employed (step 36) to self - calibrate the CGM system. In this sense, the CGM system can be said to become “self - aware”. For example, the CGM system can seed with, for example, the average glucose value from a previous session (if known), including using the previous session steady - state value or the previous session slow - moving average, as will be described in more detail below. The CGM system can also seed with the A1C value (if available). Various assumptions can also be made if appropriate. The seeded average can be represented by a distribution, the techniques of which will also be described in more detail below.

[0185] Figure 7 Describe a system where code can be provided from the sensor to the transmitter without a user input step (step 24). Again, it should be noted that while language regarding providing code to the transmitter is used here, it will be understood that the code can also be provided to various devices in signal communication with the transmitter, including dedicated receivers, smart phones, tablet computers, follower devices, or other computing environments.

[0186] In Figure 7 the Figure 42 embodiment, code or the like is provided to the transmitter without substantial user involvement. For example, a degree of coding can be achieved by sending sensors of a manufacturing batch to different markets (step 46). In one embodiment, sensors with similar codes can be sent to different geographical locations (step 48). For example, sensors sent to a particular geographical region can be from a similar or the same manufacturing batch, and when the sensor is inserted and makes an initial contact with the network, the calibration parameters known for that batch can be provided to the transmitter, thus providing an immediate degree of calibration based on geographical location. Geographical location can be used to identify a location, and the location can then be used to identify or classify the sensor.

[0187] In a similar manner, sensors of similar batches can be grouped by product, so that different codes are associated with different products (step 54). For example, a first product can have a first code associated with it, and all sensors for that product can be manufactured in the same or a similar way, resulting in little manufacturing variability between sensors associated with a particular product. In this case, once the product is identified, the associated sensor calibration parameters can be uniquely determined, at least as an average.

[0188] In another embodiment, without specifying a geographical location or product, a group of sensors with a particular code can be shipped with a code specifically associated with the corresponding user transmitter (step 56). In this case, once the calibration parameters are known for one member of the transmitted group, and this can be known for a long time before the group is transmitted, then the calibration parameters for the rest of the group are also known.

[0189] RFID technology can also be used to identify the manufacturing batch of sensors (step 58). For example, a small RFID chip can be located on the substrate of the sensor and can be read by the transmitter when the sensor and the transmitter are coupled (step 60). In another embodiment, the RFID can be read by a receiver (step 62), or alternatively by a smart phone or other device. Alternatively, the RFID device can be located on the applicator, and the transmitter can read the identification information (and thus the calibration information) again when the sensor is installed in the patient using the applicator.

[0190] In yet another embodiment, near field communication (NFC) can be employed on the package or any other component of the system (step 66) to convey identification information.

[0191] Other types of communication schemes can be employed to convey information from the sensor to the transmitter. For example, a mechanical sensor on the transmitter can allow the conveyance of code information (step 72), such as bumps, vertical pins, a mechanical system for sensing the orientation of the transmitter to the substrate, or other mechanical elements that can be read by the transmitter. A magnetic sensor can be employed for the same purpose (step 78), and in the same way, an optical reader on the transmitter (step 74) can be used to read, for example, barcodes or QR codes, as well as other identification marks or colors. A resistive sensor can also be employed (step 76), or other sensors that detect the connection state. For example, sensors with different codes of corresponding different lengths can be provided. Multiple contact pads can be provided with the sensor, and the sensor is aligned to the contact pads. Sensors with different codes can have different resistances, and the measurement of the resistance can determine the code.

[0192] Figure 8Figure 80 is illustrated, which shows a way of code communication using user data input (step 28). Perhaps in the simplest way, the code can be provided to the user at the time of purchase, and the code is simply manually input (step 84) into a receiver, smart phone, or other device having a UI that permits data input. For example, the user can input text, numbers, colors, or the like. The sensor can also be equipped with a card, such as a SIM card, and the SIM card can be inserted into the receiver (step 90) to permit the communication of calibration information without the user inputting a manual code. The transmitter can be equipped with a switching system (step 88), and the user can adjust the position of the switch on the transmitter according to instructions on the receiving sensor. For example, the transmitter switch can be a four-position switch or a DIP switch, and by appropriate adjustment, the user can provide the code associated with the sensor to the transmitter. The receiver or smart phone can also be enabled to scan a label associated with the sensor via an integrated camera or barcode reader to permit the communication of information in that form. The scanning can be of a bar label, QR code, or the like.

[0193] The following is regarding Figures 9 to 11 Describe a variation. This embodiment uses human data from the field to improve or achieve factory calibration. More specifically, human data is often used to optimally identify factory calibration parameters (e.g., sensitivity and baseline over time). Although bench data is related to human data, the correlation is not perfect, and there are often offsets in the correlation. Accessing the human data generated by each batch of sensors produced during manufacturing will generally be the best data set for generating factory calibration information. The factory calibration parameters can vary between batches, and thus it can be advantageous to characterize each batch as improvements are made.

[0194] Figures 9 to 11 A method is shown of using a portion of a batch of sensors and using data collected in humans to generate or adjust factory calibration numbers for the remainder of the batch. There are several arrangements for this method.

[0195] In one arrangement, calibrated sensors are sent to the market for patient use. Since these sensors are calibrated in a connected system, such as via a blood glucose calibration technique or other calibration techniques, including those that use only the CGM signal itself, the calibration information can be returned to the manufacturer via the cloud or other Internet-based network. The information can be used to generate factory calibration settings for the remainder of the batch of sensors that have not been shipped to the market, and then the sensors can be shipped.

[0196] In another embodiment, there are initial factory calibration settings shipped with the product. Again, cloud or network information can be monitored and a determination can be made as to how closely the actual parameters match the initial factory calibration settings. Then, the factory calibration settings of the sensors not yet shipped can be adjusted based on this determined proximity. In this embodiment and in the previous embodiment, the release of the sensor product can be staggered such that subsequent shipments have improved accuracy. This embodiment can be further implemented even after all sensors have been shipped because the adjustment can be performed over the network or through the cloud.

[0197] More specifically and referring to Figure 9 , a factory 148 is illustrated that has a manufacturing batch or lot of sensors 150, the manufacturing batch or lot generally being produced in the same (or very similar) manufacturing process. The batch or lot 150 can be divided into a first portion 154 and a second portion 156. The first portion 154 can remain temporarily at the factory 148 while the second portion 156 can be sent to a group of users 158.

[0198] Next referring to Figure 10 , data from the second portion 156 can be used at the factory 148 to inform the factory calibration of the first portion 154, thereby transforming it into a calibrated first portion 154'. If the first portion 154 has been shipped, then within the user group 158, the first portion can be calibrated before or after insertion, indicated as first portion 154". The calibration of the first portion after shipment can occur as described above by accessing network or cloud resources regarding factory calibration information, especially in the case where the first portion has been updated in the field with data from the sensor.

[0199] Figure 11 is a flow chart 160 illustrating the above method. First, a manufacturing batch or lot of sensors is manufactured at the factory under known and reproducible conditions (step 162). The batch or lot is divided into at least two portions (step 164). Here, two portions are described for convenience, but it will be understood that a manufacturing batch can be divided into any number of portions for staggered release.

[0200] In this example, the second portion is sent to the users (step 166). Then the second portion is calibrated (step 168), and the calibration can occur in a known manner, such as using prior information, bench calibration values, user data, finger stick calibration, or similar. The calibration can also occur using the techniques disclosed herein.

[0201] Next, the calibration information from the second part can be sent to the factory (step 170). Next, the calibration of the first part of the sensor can be generated or adjusted based on the data from the second part (step 172). That is, if a factory calibration has been generated for the first part, then the factory calibration can be adjusted if needed. If a factory calibration has not been generated, then the received data from the second part can be used to inform the calibration of the first part, such as the average determined sensitivity from the field, etc. The adjustment or generation can occur at the factory (step 174), or it can occur after shipment, either before or after insertion into the patient (step 176), where the transmitter, receiver, or other monitoring device (e.g., a smart phone) is in network communication with a server or other network resource operated by the factory 148.

[0202] Once the sensor is inserted into the user and the initial calibration is complete, any calibration from that point on is referred to as an "in - progress" or "continuous" calibration. Figure 12 This situation is schematically illustrated in FIG. 102. The user, patient, or subject 112 has an indwelling sensor 114, which is connected to a transmitter 115. In many cases, the transmitter is used multiple times for different sensors. In other cases, the transmitter can be made disposable.

[0203] The transmitter 115 allows the signal measured by the sensor 114 to be communicated to a device such as a receiver or a dedicated device 104 or a smart phone 108. The receiver 104 is illustrated as having a display 106, and the smart phone 108 is illustrated as having a display 110. The display 106 or 110 can be used to indicate to the user the clinical value of the analyte concentration, such as the glucose concentration. In doing so, they rely on the relationship described above, where the measured current or count is related to the clinical value of the analyte concentration by a linear relationship with a slope representing sensitivity.

[0204] The systems and methods according to the principles of the present invention describe the formation or determination of this linear relationship that is based largely or uniquely on the characteristics of the sensor signal itself, and in some embodiments do not rely on external data as in prior systems. Additionally, these "self - sensing" systems that employ "self" or "automatic" calibration can be used not only to more accurately measure subsequent analyte concentrations, but also to retrospectively modify the results of previous measurements. In this way, when this is displayed on a display such as display 106 or display 110, the measured data can be communicated more accurately. In other words, retrospective processing can be employed to correct or modify a previous calibration and even update the data measured therefrom. In this way, if the display indicates historical data as well as current data, then at least the historical data will be updated, i.e., this display will change to reflect calibration parameters that are better known or known with a greater confidence than the previous calibration.

[0205] Figure 13The flowchart 116 of FIG. illustrates this method, where the first step is the receipt or determination of a seed value (step 118), which can be received or determined using the initial calibration procedure described above. Then, calibration can be determined using the seed value (step 120). For example, if the received calibration parameter is a specific value of the slope or sensitivity m, it can be used to correlate the measured count with the clinical value of the analyte concentration and can be used to immediately start notifying the user of the measured analyte concentration, such as a glucose measurement. That is, the analyte can be measured with the sensor (step 122), and the measured value can be displayed to the user at least partially based on the seed value received in step 118 (step 124).

[0206] In some cases, a change in calibration will occur (step 126), and it can be detected in various ways, including the ways described below. Then, the calibration can be adjusted, and specifically, the calibration parameters including sensitivity and baseline can be adjusted (step 128). After the update of the calibration parameters, the display can be updated (step 130).

[0207] As described above, the update of the display can involve not only adjusting the currently measured value of the analyte concentration but also the display of the historical values that are recalculated and thus changed based on the adjusted calibration. For example, the sensitivity can be "seeded" with an initial value, but after the receipt of data, it can be determined to be actually 10% lower than the initial seed value. In this case, not only will the analyte value being displayed be adjusted, but in one embodiment, the historical values can also be adjusted to reflect the updated sensitivity. This example illustrates the case where the seed value is updated with the measured value. In some cases, a previously determined value (determined by seed or measurement) can be updated with a later determined value. This can occur, for example, when the sensor calibration parameters "drift". For example, if the calibration of the sensor is determined to have drifted, then a change can be made to the calibration parameters so that the receiver, smart phone, or other monitoring device continues to display the accurate value of the analyte concentration. In one embodiment, if it can be determined when the drift occurs, then some of the historical values in the display can be updated, i.e., those measured after the drift, while other historical values, such as those measured before the drift, do not need to be updated.

[0208] In one embodiment, if the determined seed value is close to the initial seed value, for example, within 10%, then the initial seed value (or other calibration parameters) can be simply adjusted accordingly. However, if the values are further apart, then the user can be prompted to intervene, for example, by an optional finger prick.

[0209] Figure 14 and 15 are graphs depicting the analyte concentration over time before (14) and after (15) the change in sensitivity. Specifically, Figure 14Illustrated is a graph 132, which shows a plot 138 of analyte concentration over time. Axis 134 represents values of analyte concentration, and axis 136 represents time. After the sensitivity is changed, the graph becomes graph 140, with transformed historical analyte values 146. Depending on the implementation, the sensitivity change can be regarded as an updated sensitivity or an updated seed value. A variety of other ways of adjusting calibration using sensor signal characteristics can be employed, including (but not limited to) the mean sensor signal, standard deviation, or CV (coefficient of variation), or the interquartile range of the sensor signal, or other higher-order or ranked statistics.

[0210] More specifically, and in contrast to prior efforts, the preferred embodiment describes a system and method for substantially real-time graphical representation (e.g., a trend graph showing glucose concentration over the previous several minutes or hours) of glucose data that is periodically or substantially continuously post-processed (e.g., updated) with processed data, where the data has been processed in response to an update of calibration, such as as a result of sensor drift, system error, or the like.

[0211] Referring to Figure 16 the analyte concentration measurement system 135 depicted in, and specifically at block 137, a sensor data receiving module (also referred to as a sensor data module) or a processor module receives sensor data (e.g., a data stream), including sensor data points for one or more time intervals. In some embodiments, the data stream is stored in the sensor for additional processing; in some alternative embodiments, the sensor periodically transmits the data stream to a receiver or other monitoring device, such as a smart phone, which can communicate with the sensor either wired or wirelessly. In some embodiments, the raw and / or filtered data is stored in the sensor and / or transmitted and stored in the receiver.

[0212] At block 139, the processor module is configured to process the sensor data in various ways. The processor module can also be combined with a calibration module 143 to determine whether a calibration change has occurred, as described in more detail above and below. More specifically, at block 143, the calibration module uses the data in the data stream to detect calibration changes and more specifically sensitivity changes.

[0213] At block 141, the output module provides the output to the user via a user interface (not shown). The output represents an estimated glucose value determined by converting sensor data into a meaningful clinical glucose value. The user output can be in the form of, for example: a numerical estimated glucose value, an indication of the directional trend of the glucose concentration, and / or a graphical representation of the estimated glucose data over a time period. Other representations of the estimated glucose value are possible, such as audio and tactile. In some embodiments, the output module displays the "real-time" glucose value (e.g., a number representing the most recently measured glucose concentration) and a graphical representation of the processed and / or post-processed sensor data.

[0214] In one embodiment, the estimated glucose value is represented by a numerical value. In other exemplary embodiments, the user interface graphically represents the trend of the estimated glucose data over a predetermined time period (e.g., one, three, and nine hours respectively). In alternative embodiments, other time periods can be represented. In alternative embodiments, pictures, animations, charts, graphs, value ranges, and numerical data can be selectively displayed.

[0215] The processor module can be further configured to perform post-processing steps, such as being configured to periodically or substantially continuously post-process (e.g., update) the displayed graphical representation of the data corresponding to the time period based on the received data (e.g., the most recently received data). For example, the glucose trend information (e.g., the previous 1, 3, or 9 hour trend graph) can be updated to better represent the actual glucose value by taking into account newly determined calibration values. In some embodiments, the post-processing module post-processes data segments (e.g., 1, 3, or 9 hour trend graph data) every few seconds, minutes, hours, days, or any time in between and / or when the user requests (e.g., in response to a button activation such as a display request for a 3 hour trend graph).

[0216] Generally speaking, post-processing includes processing performed by a processor module (e.g., within a handheld receiver unit) on "recent" sensor data (e.g., data containing time points within the past few minutes to hours) after its initial display of the sensor data and before an analysis generally referred to in the prior art as "retrospective analysis" (e.g., an analysis that, in contrast to intermittent, periodic, or continuous, is generally performed retrospectively on an entire data set at one time, such as the display of sensor data for physician analysis). Post-processing can include programming that is executed to recalibrate the sensor data (e.g., to better match a reference value), fill data gaps (e.g., data eliminated due to noise or other problems), smooth (filter) the sensor data, compensate for time lags in the sensor data, and similar operations. Preferably, the post-processed data is displayed "in real time" (e.g., containing the most recent data within the past few minutes or hours) on a personal handheld unit (e.g., on 1-, 3-, and 9-hour trend graphs of a receiver or smartphone), and can be updated (post-processed) automatically (e.g., periodically, intermittently, or continuously) or selectively (e.g., in response to a request) when new or additional information becomes available (e.g., new reference data, new sensor data, etc.). In some alternative embodiments, post-processing can be triggered depending on the duration of a calibration period change; for example, data associated with a change in a calibration event that extends over approximately 30 minutes can be processed and / or displayed differently from the data during the initial 30 minutes of the calibration period change.

[0217] In one exemplary embodiment, the processor module filters a data stream to recalculate data for a previous time period, and periodically or substantially continuously displays a graphical representation (e.g., a trend graph) of the recalculated data for that time period. In another exemplary embodiment, the processor module adjusts the data for a time lag from data of a previous time period (e.g., removes the time lag caused by real-time filtering), and displays a graphical representation (e.g., a trend graph) of the time-lag-adjusted data for that time period. In another exemplary embodiment, the processor module algorithmically smooths one or more sensor data points (e.g., including time points before and after the one or more sensor data points) over a moving window for data of a previous time period, and displays a graphical representation (e.g., a trend graph) of the updated, averaged, or smoothed data for that time period.

[0218] In some embodiments, the processor module is configured to filter sensor data and display a graphical representation of the filtered sensor data in response to a determination of the start of a calibration event change. In some embodiments, the processor module is configured to display a graphical representation of unfiltered data (e.g., raw data) in response to a determination of the end of a calibration event change. In some embodiments, the processor module is configured to display a graphical representation of unfiltered data except when a change in the calibration event is determined. It has been found that adaptive filtering as described herein, including selective filtering during a change in the calibration event, increases the accuracy of the displayed data, reduces the display of noisy data, and / or reduces data gaps and / or early cutoffs compared to conventional sensors.

[0219] Calibration Routine

[0220] As described above, it is desirable to provide a more convenient calibration routine for the user, and especially for type II users or those who use the system for weight loss optimization and / or exercise and fitness optimization.

[0221] One way to reduce the need for user-based calibration is to employ a more enhanced factory calibration, and certain details regarding the method associated with factory calibration can be found in USSN 13 / 827,119, filed on March 14, 2013 and published as US2014-0278189-A1; and USSN 62 / 053,733, filed on September 22, 2014, both of which are owned by the assignee of the present application and incorporated herein by reference in their entirety.

[0222] Other techniques can also be employed to facilitate calibration requirements. For example, referring to Figure 17 flowchart 145 of, if a previous sensor session showed generally reliable results (step 147), then the same calibration parameters can simply be used from the previous session for the new sensor session (step 149). Specifically, the calibration parameters from the old sensor session can be transmitted to the new sensor session in several ways, such as by using glucose signal transmitter technology if the calibration parameters are stored on the sensor electronics, or by passing a calibration status variable to the new session if the calibration parameters are stored in a monitoring device (e.g., a smart phone). This technique can be particularly useful if the sensors are related in some way, such as from the same batch, the same package, of the same type, or the like.

[0223] This technique is not necessarily limited to the use of a single previous sensor session, such as returning exactly one session. For example, by analyzing a number of previous sensor sessions, repetitive patterns can be learned, such as historical patterns. For example, the system can learn that a user is accustomed to eating pizza on Fridays and has done so in the last seven sessions, and the algorithm can thus learn not to treat these events as outliers. Patient habits can also be learned, such as a patient preferring to eat many small meals as opposed to only a few large meals. Failure modes can also be learned, such as whether a patient's trends are towards a particular failure mode due to a particular way of installing and / or using their device. As discussed below in conjunction with Figure 18 it is advantageous to learn certain glucose trace characteristics that constitute repeatable events from previous sensor sessions, and in many cases multiple previous sensor sessions are necessary to distinguish common events from outliers. Additionally, in cases where other event data is available, such as meal or exercise data, the correlation between these repeatable events and the input meal or exercise data can be learned.

[0224] Next, referring to Figure 18 , a flowchart 178 for another calibration method is shown. Specifically, it is known that certain glucose trace characteristics indicate repeatable events that reoccur at known and repeatable glucose values (if they reoccur). As an example, steady-state values, certain trend values that include certain slopes, etc. tend to be reproducible for a given patient. These repeatable events tend to result in repeatable and detectable characteristic glucose trace signatures and / or patterns. For example, it is a characteristic of many biological systems that analyte values (if not changing rapidly, i.e., being in a steady state) are highly reproducible. In other words, if an analyte value is in a steady state at a first time and then in a steady state in a second time period, then the value of the analyte (e.g., concentration) is generally at or near the same value in each steady-state time period. This concept can be used to calibrate an analyte sensor.

[0225] For example, if a user is in a steady state with respect to an analyte, the value of the analyte can be measured and stored. When the user is again in a steady state, it is likely that their analyte value is the same as the previously measured value, and thus the sensor that reads the analyte concentration value can be calibrated.

[0226] Different analytes can achieve different steady states in various ways. For example, the uric acid concentration rarely changes over a typical full day. If a user has not exercised and has not eaten for several hours, then their glucose value can be in a steady state. If a user has not exercised for several hours, then their lactate value can be in a steady state. Generally speaking, if a biological system does not change state significantly over a time period, then many analyte values, including glucose, will achieve a steady state. As described, the steady state values are reproducible, especially for pre-diabetic or non-diabetic patients, and for those who use the system primarily for weight loss optimization or exercise optimization. Therefore, whenever a steady state is detected in an analyte value, the sensor that measures the analyte can be calibrated.

[0227] In some cases, the system can prompt the user to fast or refrain from exercise in order to achieve a steady state, and then the steady state can be measured and subsequently used for these calibrations. Moreover, the system can detect the steady state but prompt the user to perform a test, for example by asking the user "Have you fasted?"

[0228] In some cases, the steady state value can be determined based on the user's demographics, so no measurement is needed at all. For example, for non-diabetic patients, the typical glucose value can be between about 80 and 100 mg / dL. In many cases, if a person is non-diabetic, then their value can be about 80, and if they are progressing towards pre-diabetes, then their value may approach 100. Therefore, simply providing the system with certain information about the user can allow a certain degree of calibration to be performed, especially for certain applications, including cases where the user does not need to have the accuracy of the value determined to an exact value, but only needs to be accurate to a specific range, such as hypoglycemia, hyperglycemia, or normoglycemia.

[0229] In addition to steady states, other repeatable events that can be employed include slopes, responses after a typical or similar meal, such as the response after breakfast in the case where the user eats the same breakfast every day. Other repeatable events include, for example, the range from a low measurement value to a high measurement value on a daily basis, i.e., the daily high-to-low range, certain types of free, certain types of transient patterns, decay rates and slopes, rates of change, and the like.

[0230] As a specific example, if a user has eaten a characteristic breakfast and the characteristic breakfast results in a characteristic post-meal glucose trend, then if a change in the trend occurs, it can be assumed that at least to some extent a change in the sensitivity of the sensor has occurred and the sensor needs to be recalibrated.

[0231] Additional data can also be used to assist in calibration. For example, if the user is a diabetic and measures their blood glucose several times a day in any manner, then these values can be used as calibration values. This is especially true if their blood glucose values vary significantly. In this case, if the system detects a local steady state, then the system can prompt the user to measure and enter the blood glucose value in order to correlate the value with the steady state.

[0232] In some cases, historical data can be used to determine steady state calibration values, such as previous data from blood records, data from previous sessions, estimates from measurements such as A1C that track long-term glucose averages, or similar data.

[0233] In some cases, it is not necessary to focus on a specific value. A determination by the user within a range of values may be sufficient for a type II user. The range can be determined and used to provide the user with information about whether a goal is met, or other information about the program they are in.

[0234] The use of the steady state assumption can be used not only for initial calibration, but also for updating calibration. That is, whenever the system is seen to be in a subsequent steady state, the values measured during the subsequent steady state can be assumed to be the values originally determined for the user. Other calculations can also be employed, including the use of weighted averages, slow moving averages (see below), and the like.

[0235] See Figure 18 , the method of flow chart 178 allows calibration of glucose concentration or other analyte values using information generated directly from the device (i.e., the analyte concentration signal itself).

[0236] Thus, referring to flow chart 178, the first step in a calibration routine according to these principles is to detect that the system is in a steady state (step 180), such as a first known steady state.

[0237] Steady state can typically be detected without any action required from the user at all (step 182). However, in some cases, steady state can be prompted by waiting a predetermined time period after an event to allow steady state to be achieved (step 184), such as by waiting a predetermined time period after a meal or exercise routine (step 186). The system or routine can ask for or prompt for user information, such as regarding fasting, such as prompting the user to enter the duration since their last meal or related parameters. In the case of glucose, the routine can be configured to look for a low rate of change or a rate of change below a predetermined value, such as less than 0.25 mg / dL per minute. In some cases where calibration has not occurred, the rate of change can be based on counts or microamperes or other "raw" signal values. The rate of change can be determined by calculation of the derivative.

[0238] Once a steady state has been detected, calibration parameters such as sensitivity can be determined (step 188) based on the measured number of counts (or on-current) at steady state and a known steady state glucose value (which may be known or assumed). Subsequently, another steady state can be detected (step 192), and in the case where drift has occurred, the second steady state will be associated with a different number of counts. The different number of counts can be used to determine the degree of drift, and the system can be recalibrated (step 194) using the new (second) number of counts measured at the new (second) steady state together with the previously known steady state glucose value.

[0239] In some cases where the change is substantial (e.g., exceeds a predetermined threshold), the user can be prompted for additional data (step 196), such as finger sticks, data on exercise or meals, etc. The difference can also be used to determine the cause of the drift (step 198).

[0240] A specific example is now described. A continuous glucose monitor can be used for patients without diabetes. In patients without diabetes, their glucose values typically range from 70 to 90 mg / dL. Their glucose levels deviate from this range only after meals or during extreme exercise, and in some cases these deviations can be detected by changes in the sensor signal or via auxiliary measurements using a heart rate monitor and an accelerometer. During these glucose excursions, the current measured by the sensor will change rapidly, which can be easily detected by the monitoring device. A rate of change can be calculated using, for example, a FIR filter on the glucose values over the last 20 minutes, but can extend to a simple rate of change, such as defined as the difference between two glucose values at the start and end of a time period divided by the duration of the time period. During a rapid rate of change, a system and method according to the principles of the present invention can avoid using this rapidly changing data for calibration if steady state occurrences are used as a repeatable event for calibration, but instead can wait until the value stabilizes and then perform a calibration event (as noted, some calibration methods can take advantage of this situation, e.g., as described above, in some cases, repeatable events available for calibration can include transient noise events or patterns, i.e., certain aspects of high change regions or peaks can be useful for non-steady state repeatable event calibration, and in many cases, transient events including rapid or slow rates of change of analyte concentration values can form signal characteristics from which patterns can be inferred and used as repeatable events). In one example, an absolute rate of change threshold can be set at 0.25 or 0.5 mg / dL / minute on the glucose data over the last 25 minutes. As noted above, uncalibrated units can also be employed. Thus, if the absolute rate of change threshold is exceeded, calibration can be prohibited. In other embodiments, different calibration values can be used, and they can also be configured to depend on the direction of the rate of change, or the calibration value used can vary with the rate of change itself.

[0241] See Figure 19 , and as described in step 194 with respect to Figure 18 , a steady state can be used to update calibration values and to determine initial calibration values. Also, the system can be employed even when calibration parameters change, such as when the sensor enzyme layer changes over time in use, or when other drifts occur. For example, see Figure 19 in graph 202, where the analyte value can be related to a calibrated or uncalibrated value at time t0 (axis 206), and this steady state value can be assumed to be reproducible whenever the user is in a steady state in the analyte value. This knowledge can lead to an initial calibration line 208, the slope of which is the sensor sensitivity when the axis 204 has count units. At a subsequent time t1, also assumed to be in a steady state, knowledge of the same steady state value allows the subsequent calibration line 210 to be plotted and thus the system can be recalibrated. The degree of recalibration required can be highly useful in determining the cause of the drift and dealing with it.

[0242] Systems and methods according to the principles of the present invention can be most effective when the baseline or background signal is stable and predictable enough (or eliminated by advanced thin film or sensor technology). When the sensor is activated, a current is generated. If the baseline is small enough or estimated with sufficient accuracy, then the remaining current will be from the analyte of interest. The algorithm can measure the current over a set time period, and if the current is stable, such as within a specified limit, then the algorithm can assume that the glucose (or other analyte) is not changing and is within a narrow range. The algorithm can then use a glucose value of, for example, about 80 mg / dL (or any value determined to be typical for the user) and correlate it to the current generated during that time period in the case of stable glucose to automatically calibrate the device. In one embodiment, the glucose value is set at 100 mg / dL (it should be reiterated that this can vary depending on several factors, such as the rate of change, wear duration, time of day, characteristics of the user, such as the state of progression towards diabetes, etc.). The system and method can employ a regression model to calculate the slope and baseline with two points. The first point is the current generated during stable glucose (and the approximate glucose level of a non-diabetic user), and the second point is zero glucose (using an estimated value of the background signal). The slope of the line can be determined using, for example, a weighted average of the regression slope (counts / assumed to be BG) and a previous slope estimate. After calibration, the glucose data can be presented to the user. The baseline of this embodiment is assumed to be zero, however, different non-zero baseline values can be used. Depending on the embodiment, calibration can also be updated periodically, such as every few minutes or every few hours.

[0243] The above techniques can be employed in combination with factory calibration information generated during the manufacture of the device, or it can also be used with externally generated glucose information, and sensitivity can be varied over time by incorporating a predefined drift curve or other drift compensation techniques, as described in more detail in U.S. Patent Application USSN 13 / 446,848, filed Apr. 13, 2012 and published as US2012-0265035-A1, which is owned by the assignee of the present application and incorporated herein by reference in its entirety.

[0244] Systems and methods in accordance with the principles of the present invention can further be used to calibrate one sensor using calibration information about another sensor, such as a neighboring sensor, e.g., sensors under the same thin film. This calibration can be performed as a drift parameter, which can be assumed to be the same for both sensors if caused by the thin film. For example, if two sensors are under the same thin film layer, such as a glucose sensor and a lactate sensor, and if one or more calibration parameters are determined in vitro, then the calibration parameters can be assumed to have a similar relationship in vivo, and thus the measurements of one sensor can be used to determine the measurements of the other sensor. For example, if the lactate sensor has a known offset (or other relationship or scaling or correlation factor) from the glucose sensor as measured in vitro during calibration, then in vivo, the determination of the calibration of the glucose sensor can be used to calibrate the lactate sensor. For example, if the calibration of the glucose sensor is seen to drift by 50%, then the calibration of the lactate sensor can be assumed to have drifted by 50%. Thus, an update of one or more calibration parameters of one sensor can result in an update of one or more calibration parameters of the other sensor.

[0245] Additional details of these aspects can be found in U.S. Patent Application USSN 12 / 770,618, filed Apr. 29, 2010 and published as US-2011 / 0004085-A1; and USSN 12 / 829,264, filed Jul. 1, 2010 and published as US-2011 / 0024307-A1 and [545PR], all of which are owned by the assignee of the present application and incorporated herein by reference in their entirety.

[0246] Additionally, systems and methods in accordance with the principles of the present invention can start with factory calibration information and then incorporate automatic calibration techniques over time to obtain more accurate glucose information. If the signal does not follow or is outside of pre-specified parameters, the system can use known techniques to request a calibration value, such as SMBG or finger stick calibration. The system and method can then incorporate this glucose information into the original parameters to adjust the set point from, for example, 100 mg / dL to a more appropriate and accurate value. That is, while the above techniques established for use in certain applications can generally be configured to avoid the need for finger stick calibration, if this is available, systems and methods in accordance with the principles of the present invention can advantageously also be applied for calibration purposes and other purposes.

[0247] Systems and methods in accordance with the principles of the present invention can be configured to determine a confidence level or range, and the confidence level or range can change when the resolution or accuracy of the data changes. More specifically, the display can generate value and trend graphs, or this can show ranges or other UI elements. The range can change over time and can shrink or expand when the confidence in the accuracy changes. For example, during an initial warm-up period, factory calibration values can be utilized. However, their accuracy may not be as precise as in the case with additional information. During this time period, the display can show a range instead of a value.

[0248] Another feature of systems and methods in accordance with the principles of the present invention is that they can request information when the user is set up within the system and adjust the techniques to be used depending on the information. For example, the system can prompt the user to enter whether they have type I diabetes, type II diabetes, or are non-diabetic, and can select different techniques depending on the response. Systems and methods in accordance with the principles of the present invention can further ask the user if they are interested in weight loss optimization, exercise and fitness optimization, or other similar optimization routines, and can adjust the algorithm accordingly. The device can also be used, for example, in a "blind" mode for an extended time period, such as 14 days, and only accept blood glucose values. These blood glucose values can be used to understand what the patient's resting blood glucose is, which can better guide the assumption of an automatically calibrated steady-state blood glucose value. After the extended time period, the user can then use the device in an automatic calibration mode.

[0249] In addition to using the nature of the steady-state value of the analyte to gather additional information, "slow moving averages" can also be employed, i.e., averages taken over, for example, 1 to 3 days, as these slow moving averages are also generally constant, especially in the use of sensor sessions. Thus, the change in the slow moving average can be used qualitatively and quantitatively to detect and quantify drift. For example, the value of the slow moving average of the glucose concentration is shown by Figure 20 the graph 220 in Figure 21The more schematic graph 212 therein shows that sensor counts are shown on axis 214 and time is shown on axis 216. As can be seen, the slow moving average G1 measured at time t1 can be reduced to the slow moving average G2 at time t2. The slow moving average can be used to quantify drift because the selectivity of the advanced sensor for glucose is high. Therefore, the only contributor to the change in the slow moving average is sensitivity.

[0250] Although the details of the use of the slow moving average are described in detail below, it should be noted here that these do not need to be contiguous time periods. For example, the slow moving average can be obtained by sampling over a common time period of several days. For example, the slow moving average of the nighttime time period can be obtained, and this can then consider only a time period such as from 11 pm to 7 am. The slow moving average will then constitute the average of the data obtained only during this time period but over the course of several days. Other exemplary time periods during which this discrete or non - contiguous slow moving average can be obtained can include, for example, the post - meal time period and so on. These events can be time - based, where the user has a very consistent timing of these events, or event - based, such as where the event is marked or tracked by one or more sensors. For example, an exercise event can be marked by an accelerometer, a meal event can be marked by detecting a glucose spike, and so on.

[0251] More specifically, instead of using the daily average, a time period can be adopted where the average of a time period specific to a qualitative or quantitative type, which is specific and important to the user, is obtained, such as making the time window for observing the average a specific one. For example, the time window can be considered as "four hours after a meal". The time period data can be used as measured over several days, but only the specific time period is observed, and thus only the change in the average from the glucose response during the specific set of time periods is measured. In other words, the average can be determined by stitching together all the glucose values from individual time periods over the course of several days and taking the average. In this way, if drift is detected, it can be defined relative to the average obtained by taking the average over these similar time periods. For example, the daily average of a patient may be 100, but their nighttime average may be 85, and their "regular" waking daytime average is 120. The measured value of drift can be obtained relative to this defined "localized" average. Exemplary time periods can include, for example, after dinner, from 9 am to noon, during sleep, and so on.

[0252] Slow moving averages can be obtained for calibrated or even uncalibrated values, but the slow moving average itself may require additional data to perform an initial calibration. For this reason, embodiments can include using other methods to determine an initial glucose value, and obtaining valuable data for example over a day, from which an initial slow moving average can be determined and then compared to subsequently measured slow moving averages to determine for example sensitivity drift corrections and the like. These systems and methods can be particularly beneficial because the slow moving average can be checked very frequently, for example every 5 minutes, compared to prior systems, which contrasts with SMBG calibration, which can only be done as frequently as the user desires to take a reading.

[0253] Figure 22 Flowchart 222 of FIG. illustrates one embodiment of the use of a slow moving average. In a first step, an analyte value is measured using a sensor (step 224). A first slow moving average can then be determined (step 226). Analyte value measurements can then continue (step 228), and this can form the basis for a displayed value of the analyte concentration (step 232), where the displayed value is based on an initial (or subsequent) value of sensitivity. Second and subsequent slow moving averages can be determined (step 230), where the period of the slow moving average is generally greater than 1 / 2 day, for example 1 to 3 days. The slow moving average can be obtained at a desired frequency, for example every 5 minutes, every hour, etc. In some embodiments, the time constant of the filter can be changed, for example if the user has a high actual value and thus the effect (the user's high analyte concentration) does not represent an actual drift or change in the sensitivity of the filter. The use of a slow moving average can also be replaced by other forms of filtering, such as sequential statistical filtering or time domain filtering.

[0254] If the first and second values of the slow moving average (or second or subsequent values, or indeed any set of values) change, then in one embodiment, drift can be assumed to have occurred and the calibration can be adjusted based on the drift (for example, the difference between the slow moving averages) (step 234). Other potential causes of changes in the slow moving average are discussed in the embodiments described below. The display can be updated based on the re-calibrated sensor (re-calibrated at least in part based on the measured drift) (step 236). In some cases, the seed value (or other value used as the basis for the display) can also be modified (step 238) to reflect (and compensate for) the drift.

[0255] See Figure 23AFor the flow chart 240, recalibration (or other recalibration) based on data collected from changes in the slow moving average can also be used in post - processing to update the display of historical (here defined as "already displayed") values based on the recalibration. In a first step, the value of the analyte concentration can be measured by a sensor (step 242). The measured value can be displayed as a clinically based value based on calibration, such as a previously determined value based on sensitivity (step 244). The sensitivity can change based on sensor signal information (step 246), including exclusively based on, for example, changes in the slow moving average or steady - state value. Then the display can be updated based on the change (step 248), and in particular, the historical already - displayed values can be redisplayed based on the recalibration so as to more accurately represent the historical values. Other forms of detecting sensitivity changes (or drifts) using signal characteristics (or features) include the coefficient of variation (CV), standard deviation, or the inter - quartile range of the signal and the corresponding glucose parameters related thereto.

[0256] Once a change has been detected, it can be analyzed or "discriminated" to determine the cause and / or magnitude of the change. Usually, sensitivity changes due to drift are found, but this can also have other causes, including pump problems, such as blocked tubing, or other problems, such as membrane rupture or the like. And it is further desirable to discriminate these changes from those due to actual changes in the glucose value. At least as a first step in such a determination, the measured change in the glucose value can be compared with physiologically feasible thresholds for such changes. If the change is not physiologically feasible, then the change can be considered to be at least partially due to drift or system failure.

[0257] Another way to discriminate the signal drift behavior is by comparing the signal drift curve with known signal drift curves, and in particular with a plurality of these curves or the envelope of these curves. Figure 4 One such curve is illustrated, but for a given type of sensor, an envelope of these curves exists, i.e., there is a pattern of how the sensitivity changes. If the way the sensitivity is changing follows one of these curves, then it can be inferred that the change is due to sensitivity drift rather than an actual glucose value change or system failure. Additional details regarding the sensitivity distribution curves are described, for example, in U.S. Patent Application No. 13 / 796,185, titled "Systems And Methods For Processing Analyte Sensor Data", filed on September 19, 2013, which is owned by the assignee of the present application and incorporated herein by reference in its entirety.

[0258] If the sensitivity is changed by shifting to another one in a known sensitivity curve, the known calibration parameters for the curve can be adopted in subsequent data analysis and display. If the sensitivity is changed outside the bounds of a known sensitivity curve, it can be inferred that the change is due to a system problem or failure as described above, such as an error or artifact. However, in some embodiments, a certain sensor can have a known sensitivity curve within a band. Known failure modes can cause the sensitivity to shift to another curve within the band or to another band, i.e., the sensitivity can shift to one of the curves of known separate discrete bands in a known failure mode.

[0259] It should be noted in this regard that, generally speaking, a given type of sensor will function to meet the required objectives up to a certain tolerance. For example, 80% of the sensors in a batch can work as required (see Figure 23B ). On the other hand, the remaining 20% (see Figure 23C ) may not follow the expected sensitivity curve. A larger percentage (e.g., 75%) of these remaining sensors can follow a known failure mode, which causes them to follow a known alternative sensitivity curve, a group of curves, or a curve band. By identifying which of these sensors follow the alternative sensitivity curve and adjusting the calibration of the sensors accordingly, the "failure" of the sensors can be mostly remedied. For example, if the failure mode causes all of the 75% to have signal values that tend to decrease in the same way, then after the determination of the failure mode, the "failure" can be remedied by "upward" adjusting the sensor readings. These aspects can be particularly important as sensor sessions become longer, e.g., from 7 sessions to 14-day sessions. In the Figure 23C failure mode illustrated, the sensitivity decrease starts at about the eighth day. The ability to detect and remedy these failure modes is particularly important because even though the start of the session sensor failure mode is becoming better characterized, the end of the session sensor failure mode, especially with longer sessions, is still difficult to quantify.

[0260] In some embodiments, the consideration of the combination of glucose signal variability and a slow moving average can be employed to distinguish between glucose signal changes and sensor sensitivity fluctuations. For example, if the slow moving average decreases but the variability remains the same, or remains within a predefined range or band, then the reason for the decrease in the slow moving average may be a change in sensitivity. Alternatively, if the slow moving average decreases but the variability also decreases, then the reason for the decrease may be a real and actual change in glucose concentration.

[0261] Non-physiologically viable changes, variations, errors, artifacts, and other signal manifestations may be the cause of various remedial actions of the system, and some of these actions have been mentioned above. For example, recalibration may be performed, and the results can be propagated back to historical data. The user may be prompted to provide fingerstick calibration. Part of the remedial action algorithm can be to determine whether to correct via recalibration or whether to prompt for fingerstick or other calibration points. For example, if a signal outside of physiologically viable boundaries is received, then the user may be prompted for additional calibration points, such as fingerstick. Alternatively, the user may be prompted to provide other types of additional information, such as entering data corresponding to a recent exercise or meal or other recent changes in the user's behavior. As a specific example, if the slow moving average of the user's glucose concentration was 100 mg / dL for the first three days of a session, but suddenly changed to 200 on the fourth day, then the system and method according to the principles of the present invention may prompt for fingerstick calibration. If the slow moving average on the fourth day was 105 mg / dL, then the system may scale or adjust the sensitivity accordingly, such as bringing the value back down to 100 mg / dL.

[0262] In the case of performing fingerstick calibration, what the fingerstick calibration data is used for can vary depending on the use of the device. For example, if the device is additionally used, i.e., used non-therapeutically (as is the case for many type II users), if fingerstick indicates drift, then the sensor may still be used in cases where the drift is not significant. In many cases, the techniques described herein can be used to remedy the effects of drift and still allow for the display of accurate readings. If the device is not additionally used, such as for insulin-using type I patients, then if fingerstick indicates drift, then more aggressive remedial action or adjustment, such as recalibration, may be performed, and if this cannot be done, or if an accurate sensor reading cannot be obtained even after recalibration, then an indication to stop using the sensor may be displayed to the user.

[0263] Pattern analysis may be performed to determine whether the change or variation is of a known type, such as a characteristic of a known sensitivity change. Pattern analysis can determine whether the change or variation meets criteria known for certain manifestations, such as exceeding certain thresholds. As described above, the performance for a day can be used to determine the slow moving average. If the data for that day is determined by the system to be untrustworthy after analysis, then the data for another day may be used. The data may be displayed in several ranges or bands or as a confidence interval or other indication, rather than as a high-precision numerical value. In the case of using a slow moving average, the time constant of the slow moving filter can be adjusted to include or exclude shorter-term variations.

[0264] Figure 24 These aspects are summarized in flowchart 250 of

[0265] First, referring to Figure 24 , values of analyte concentration can be measured by a sensor (step 252). These values are generally measured as current, for example as amperes (picoamperes), and equivalently as counts. A slow moving average can be defined by measuring the counts over a long time period, for example over several hours, half a day, 24 hours, or two to three days. Since the sensor varies between units, the slow moving average will generally only be meaningful once an initial period (e.g., 24 hours) has passed. Thus, the next step is to determine a first slow moving average in a first time period (step 254).

[0266] In one embodiment, an initial value of the first apparent sensitivity can be assumed via a seed value in the manner described above. The first apparent sensitivity can then be determined based on the first slow moving average (step 256). Specifically, if the first slow moving average is assumed, the first apparent sensitivity can be based on the relationship between the assumed first slow moving average and the measured average.

[0267] Subsequently, a second slow moving average in a second time period can be determined (step 258). And again, optionally, a second apparent sensitivity can be based on the second slow moving average (step 260).

[0268] If it is found that the slow moving average or the apparent sensitivity has changed between the first and second time periods, then a remedial action can be required. Thus, the next step is to determine whether the change matches a predetermined criterion (step 262). The predetermined criterion can include several elements (step 270), such as known sensitivity changes over time, the envelope of the sensitivity distribution curve, physiologically feasible changes, changes associated with errors such as pump failures, or the like. For example, this step can require a determination as to whether the change matches a criterion associated with sensitivity drift, failure, or an actual change in the average glucose concentration value (step 264). If it is consistent with drift, for example if it is determined by comparison with a known sensitivity distribution curve that the sensitivity has drifted, then correction can be made automatically. In any case, the sensitivity can be adjusted at least in part based on the difference between the two slow moving averages (step 268), since the difference provides a quantitative indication of the degree of change or drift that the sensor has undergone.

[0269] If the change is not consistent with the drift, then it can be determined whether the change is consistent with other causes for which a predetermined criterion exists. If the change is not consistent with known behavior, e.g., does not match a predetermined criterion, then the user can be prompted to enter information to explain the change (step 266), such as meal or exercise information, fingerstick calibration values, data from other external sources, or similar information. In some cases, the user-entered data can be employed in a recalibration routine along with the slow moving average or the quantization difference of the sensitivity, to, e.g., determine a new or updated sensitivity.

[0270] Figure 25 Another flowchart 286 of an exemplary method that employs a slow moving average or a steady state value is illustrated. In a first step, after an initial calibration, the analyte concentration value continues to be monitored with the sensor (step 288). The initial calibration can be based on a number of factors (step 290), including population averages, data from previous sessions, bench data, in vitro data, or other prior data.

[0271] Based on the measured value and based on the initial value, the clinical value of the analyte concentration is calculated and / or displayed. An updated calibration can then be calculated based only on the measured value, e.g., based only on the signal from the sensor (step 294). The adjustment can be based on a change in the slow moving average, a change in the steady state value, or other basis.

[0272] After the updated calibration, the clinical value can be calculated and / or recalculated based on the updated calibration (step 298), and then the display can be updated (step 300), including updating a previously displayed (historical) value to an updated value based on the updated calibration.

[0273] Figure 26 A flowchart 302 of yet another embodiment of the principles of the present invention is illustrated. A first step is to receive a seed value of a calibration parameter (step 304), e.g., sensitivity, on the monitoring device. The seed value can be based on a number of factors (step 306), including user self-characterization of disease state, population averages, data from previous sessions, bench data, in vitro data, or other prior data.

[0274] The monitoring device then continues to receive sensor data and can detect when the analyte concentration value is at a steady state (step 308). For example, this can occur when a set of received signals within a predetermined time period is within a predetermined range or band of values. A correlation of the measured signal value (e.g., in current or counts) with a known or assumed steady state value can then occur (step 310).

[0275] After the correlation step, the monitoring device continues to receive signals from the sensor (step 312). The clinical value of the analyte concentration is calculated and displayed based on the received signal, the known or assumed steady state value (even if the subject is no longer at a steady state), and the seed value (step 314).

[0276] Performance outside of pre-specified parameters can be detected, as described above in connection with Figure 24 that described, and the user can be prompted to input external data, such as a fingerstick calibration value (step 316). Recalibration and / or recalculation can be performed based on the received signal, external data, and optionally a seed value, followed by subsequent display (step 318). In some embodiments, known steady state values and / or seed values can be reset based on the calculations performed (step 320), and historical values recalculated and redisplayed.

[0277] Figures 27 to 33 FIG. illustrates a detailed method for determining calibration parameters (e.g., sensitivity and baseline) using a probabilistic approach. Certain aspects of the probabilistic approach are described in U.S. Patent Application No. 13 / 827,119, filed Mar. 14, 2013, titled "Advanced Calibration For Analyte Sensors", which is owned by the assignee of the present application and incorporated herein by reference in its entirety. In this application incorporated by reference and containing what is herein referred to as a "signal-based calibration algorithm", prior calibration distribution information is modified with real-time inputs and converted to posterior calibration distribution information, thereby determining calibrated data points. In this way, calibration errors can be avoided, where for example regression results in incorrect sensitivity and / or baseline values due to inappropriate assumptions about reference data. In addition to the manner described in the application incorporated by reference above, other ways of determining calibration parameters such as sensitivity and baseline can also be employed. These ways include techniques that make factory calibration available during the life of a sensor session and the like.

[0278] In Figures 27 to 33 again, distributions for calibration parameters such as sensitivity and baseline are employed, but they are optimized based on subsequently known data, such as the sensor count distribution obtained during the first 24 hours of use of a sensor session. Referring first to Figure 27 the flowchart 322 of, the first step is to receive an initial value of analyte concentration from the sensor, and to receive an initial value or distribution of sensitivity and optionally baseline (step 324). In some cases, the influence of the baseline can be reduced to essentially zero, thereby simplifying the calculations. The initial value of sensitivity can be from the sources described above (step 326), such as user input, derived from population averages, transferred from a previous session, or other sources of seed values. The initial value can also be used as the basis for a slow moving average filter or part of a calculation. Where the initial value of sensitivity or baseline is a distribution of these values, then it can be formed at least initially from considerations of population statistics.

[0279] Next, the analyte signal from the sensor is monitored (step 328). Next, a plurality of clinical values are calculated and displayed based on the monitored signal and the initial value or distribution of sensitivity (step 330) or alternatively based on the initial value of average glucose. For simplicity, the baseline is assumed to be zero or negligible. Assume an initial value or distribution of sensitivity such that an analyte concentration value can be provided to the user, even if the value is not as accurate during this initial time period as it will be when additional data is collected at subsequent times.

[0280] Next, the value distribution of the monitored signal can be determined (step 332). A representative value of the value distribution of the monitored signal can be calculated (step 334), such as an average value, a median value, a mean value, etc. As the initial sensitivity, based on the initially assumed sensitivity, the representative value can be divided by the initial value of the analyte concentration (step 338).

[0281] Next, the initial value or value distribution of sensitivity and / or the plurality of concentration values can be optimized to match the value distribution of the monitored signal (step 336). More specifically, the sensor count is the product of the sensitivity and the analyte concentration, and thus the concentration is equal to the sensor count divided by the sensitivity. The median sensor count can be determined, for example, after one day of data has been obtained, and a search can be performed that optimizes or provides the best fit of the distribution of the sensor count given a distribution or sample from the sensitivity parameter space and the baseline parameter space. For example, if the long-term glucose value range of the user over one day is from 100 to 200, then certain limitations on what the sensitivity and the baseline can be can be inferred. Thus, the sensitivity and the long-term glucose value within the distribution can be selected such that their product optimally optimizes the measured representative value of the sensor count. Additionally, the sensitivity and the long-term glucose value can be selected (step 340) such that their values are the'most likely', where'most likely' means that their values are closest to the center of their respective distributions. The slow moving average can be monitored (step 342), and changes therein can be detected and analyzed as described above (step 344).

[0282] In other words, after the first day, data exists regarding the initially assumed average glucose value or sensitivity and the distribution of the sensor count. From this distribution, the median sensor count can be obtained.

[0283] The sensor count SC = f SC (SC), which has a normal distribution.

[0284] The sensitivity equation has the following form:

[0285] y = mx + b

[0286] If it is assumed that b = 0 and the equation is further specified as an average value, then:

[0287] Median sensor count = m * average glucose value

[0288] And both the sensor count and the sensitivity are considered to be normally distributed:

[0289] f SC (SC) = m * GV

[0290] Moreover, it is known that m is a slowly varying function of time due to drift, and thus:

[0291] f SC (SC) = m(t) * GV

[0292] And thus it is clear that the sensor count and the glucose value are linked by a multiplicative "constant" which is actually a slowly varying function of time.

[0293] The distribution of the sensor count can also reveal aspects of the underlying distribution of the sensitivity, i.e., the underlying initial sensitivity m, and specifically:

[0294] Average GV = (Median SC / m median )

[0295] And thus:

[0296] m median = Median SC / Average GV

[0297] For example, if the median SC is 131,000 and the average GV is 131, then m is 1000 counts / (mg / dL). And the distribution of m can be examined to determine whether this value is reasonable or not. And a similar determination can be made for the case where 'b' is not negligible.

[0298] The representative value of the monitored signal (sensor count) can be converted to an estimate of long-term glucose:

[0299] Long-term glucose = (Long-term sensor count) / (Sampled sensitivity) – Sampled baseline, in mg / dL

[0300] Figure 28 Graph 346 of shows an exemplary distribution of the sample sensitivity, from Figure 29 Graph 348 of shows an exemplary distribution of the sample baseline, and Figure 30 Graph 350 of shows an exemplary distribution of the sampled long-term glucose values. As an example, if the representative value of the sensor count is 113,536, then given the constraints of the above three distributions, the best slope is 890 counts / (mg / dL), the best estimate of long-term glucose is 153.5685, and the best baseline is -26 mg / dL. These points are in Figure 31 、 32and are illustrated at points 354, 358, and 362 respectively on the same set of graphs 346, 348, and 350 reproduced on 33.

[0301] In the variations, the distribution can be made more 'granular' such that different distributions can be provided for different demographic groups or clusters. Other variations will also be understood.

[0302] As described above, a slow moving average filter can be used as part of drift quantization because the selectivity of the advanced sensor for glucose is high. Thus, the main contributing factor to the change in the slow moving average value is sensitivity. To initially seed the slow moving average filter, the initial seed value of the average glucose concentration can be multiplied by the average sensitivity to obtain an initial count number. Subsequently:

[0303] S t = αFilter t α * S t-1 +(1 – α)S TX t sensor t

[0304] After a time period (e.g., 1 day), enough data can be received such that it can be corrected to the actual average and used for drift determination. Then the above steps can be repeated to determine the subsequent optimal combination of slope, baseline, and glucose in the above manner that will give the best raw count to match the estimate of the raw count determined from e.g., the first day's data.

[0305] In some embodiments, the distribution is modified based on the measured data as more data is obtained. In this way, better calibration can be obtained. At the start, only a hypothesized distribution is employed. Subsequently, actual measured data becomes available and can be used. The filter can be reseeded with the median or other representative sensor count, and the distribution will generally converge to the actual measured data. If finger stick data is available, it can be used for even faster convergence.

[0306] Seeding or reseeding can be performed in several ways and customized to achieve faster convergence of the drift curve based on filter seed parameters. For example, the initial seed value of the filter can be based on the expected signal level as estimated from the expected average glucose level and sensor sensitivity. The initial seed value of glucose can also be based on the subject demographics such as the user's age and duration of diabetes. The initial seed value can also be based on data such as medical record information or laboratory tests, e.g., fasting glucose level, hemoglobin A1C (A1C) test, current diabetes treatment (e.g., oral medications, basal insulin use, or intensive insulin therapy), or downloaded self - monitored blood glucose values.

[0307] In another embodiment, an initial seed value can be used first to start a filter operating in the forward direction. After a representative set of sensor readings has been collected, e.g., after 24 hours of sensor readings, a second filter can then be run in the reverse direction. When a representative set of sensor readings is available, a typical signal value can be used to seed the forward or reverse filter, e.g., the median sensor reading, or a typical signal value adjusted for expected drift can be used to seed the filter, e.g., starting the forward filter at 0.9*median and the reverse filter at 1.1*median. The advantage of these techniques is that when two or more filters are used, e.g., a forward filter and a reverse filter, their seed values can be further optimized to minimize the difference between the two filter outputs. For example, an exemplary way is to minimize the mean square error between the two filter outputs.

[0308] In systems and methods according to the principles of the present invention as described above, redefining the seed value daily helps to minimize the mean absolute error in the signal domain. In one embodiment, on a daily basis, using a rough estimate of the signal trajectory from the first day, for example, the smooth trend of the noisy filter output can produce the trajectory of the drift for the next day. The drift rate estimates can be compared from two or more different methods, and the difference or error between the two can be fed into an algorithm, e.g., a signal-based calibration algorithm as disclosed in the patent application incorporated by reference above (No. 13 / 827,119), which is particularly related to determining the sensitivity distribution of the deterministic interval. Also in this way, signal features can be extracted, including features corresponding to noise, level, drift model, power, energy, etc. In this way, it can be determined whether the drift correction is on the appropriate trajectory. For example, if the drift slope is quite different from what is indicated by the factory calibration model, e.g., differing by more than a predetermined threshold percentage, then the user can be prompted for feedback or can be prompted to provide a finger stick calibration.

[0309] In some embodiments, a smoothing filter can be employed to compensate for signal drift in real time. In one case, a double exponential smoothing filter is used. This filter can assume a multiplicative decay trend without seasonality; however, an additive or multiplicative seasonality can be assumed to improve performance. The double exponential filter operates to recover the potential changes over time in the sensor signal, i.e., the drift. There are three main potential equations governing the double exponential smoothing filter:

[0310]

[0311] A table of the parameters in the equations can be seen in Table I below:

[0312] Table I

[0313]

[0314] In the above equation, and in this setting, α and β can be considered small. α is small because it is desired to make the filter insensitive to free glucose. β is small because the underlying trend being recovered is moving slowly in nature. A set of exemplary results (Table II (assuming a five-minute sampling rate)) are generated using the following parameters:

[0315] Table II

[0316]

[0317] In the above equation, the slope can be the initial sensitivity calculated by the algorithm or a value determined from any one of another method or a method used to determine the above sensitivity value. The average glucose in the above equation can be the mean of historical glucose values from a previous session or, for example, based on the A1C value reported by the user. In one embodiment, data is generated using the initial sensitivity estimated by the algorithm with 2-hour calibrations and self-reported A1C numbers from a group of test subjects.

[0318] The drift correction curve is estimated using the following equation:

[0319]

[0320] The sensor signal is then drift-corrected using the following equation:

[0321]

[0322] The glucose value is then calculated using the following equation:

[0323]

[0324] In the above equation, the slope estimated by the algorithm at initial calibration and baseline is the slope multiplied by 1 mg / dL.

[0325] To illustrate the efficacy of the double exponential filter, Figure 34 and 35 exemplary glucose trajectories are illustrated. Figure 34 The CGM trajectory 364 is shown along with reference and calibration values. Figure 35 The raw sensor signal 366 is shown along with the output 367 from the double exponential filter. In this case, the sensor is calibrated once using two starting values input by the user.

[0326] The estimated drift curve of the above sensor is through Figure 36The curve 368 therein is visible. As can be seen, the drift curve is easily visible in both the upward and downward directions, and the knowledge of the drift detected and quantified by the double exponential filter allows for the correction of the drift. An advantage of this embodiment is that it does not rely on any known curve shape to correct the drift.

[0327] Figures 37 to 39 is an additional graph that illustrates drift correction according to the above principles.

[0328] In addition to the use of the double exponential filter, other filters can also be used. For example, a Kalman filter can be employed, which includes process noise (also known as model noise) as well as measurement noise estimation. Gaussian filters, classical Butterworth low-pass filters, moving median filters, moving average filters, etc. can also be used, as long as the filter helps to reveal the underlying trend. A filter bank or a series of filters can be used, which combines multiple filters in order to obtain the average or combined trend in the underlying signal. In the case of using multiple filters, different types of filters can be employed within a single group, and the filter settings can vary between the filters. By using multiple filters, signal correction can be improved at the end of subsequent time periods, such as at the end of the valuable data on the second day, at the end of the valuable data on the third day, etc. Without wishing to be bound by theory, it is assumed in the use of these filters that the variability of the average glucose measurement per day is negligible compared to the slope change or the change in the underlying signal.

[0329] In other variations, the signal can be preprocessed before drift estimation filtering to remove data gaps and outliers. Additionally, the calibration can be automatically updated in a manner to reduce the occurrence of unexpected jumps in the CGM readings. These manners include making changes when the signal is stable, such as changing the calibration settings at midnight, or slowly blending the calibration change into the current setting over a time period (e.g., an hour or longer).

[0330] Other useful techniques that can be employed together with filtering include various decomposition techniques. For example, empirical mode decomposition can be used to decompose the signal into a series of intrinsic mode functions over time. Other time- and frequency-based decompositions can be used, including Fourier transform and wavelet transform.

[0331] In another variation, other signal-based parameters can be determined and used in the calibration. For example, referring to Figure 40 , it can be seen that the coefficient of variation (CV) of the sensor signal has a strong correlation with the glucose change in the signal, such as a strong correlation with the glucose standard deviation. Therefore, in determining the calibration parameters, this correlation can be used to select the calibration parameters that satisfy the signal CV - glucose standard deviation error model.

[0332] More specifically, prior information, as already described, can be used in factory calibration and this includes information obtained prior to a particular calibration. For example, this information includes information from a previous calibration, information obtained prior to sensor insertion, and so on. The calibration information includes information for calibrating a continuous glucose sensor, such as but not limited to: sensitivity (m), sensitivity change (dm / dt), and other signal and time derivative aspects as already described above. Also importantly, the prior information includes distribution information such as range, distribution function, and distribution parameters including mean and standard deviation.

[0333] The standard deviation of the distribution specifically with respect to glucose values can be advantageously used, for example, to determine the boundaries of possible glucose values and possible combinations of sensitivity and baseline. The standard deviation can also be used to determine a level of certainty from prior calibration distribution information, for example, in the case where the information is feedback of posterior calibration distribution information from a previous calibration (where the level of tightness or looseness of the distribution is quantified by the standard deviation). In the same way, the level of certainty can be determined from posterior calibration distribution information, for example, based on the level of tightness or looseness of the distribution, which can again be quantified by the standard deviation.

[0334] As a specific example, Figure 41 shows the glucose signal over a 10-day period. From this, the signal standard deviation and signal mean can be calculated. The coefficient of variation of the signal can then be determined as follows:

[0335] Signal CV = Signal standard deviation / Signal mean

[0336] In the Figure 41 case, the signal CV is calculated to be 0.4230.

[0337] If the relationship between the signal CV and the glucose standard deviation has been determined, for example, see the line in Figure 42 , then the calculated signal CV determined above can be used to determine the expected glucose standard deviation, which can then be used in the above calculations and other calculations. For example, this correlation and the error model used in calibration in U.S. Patent Application No. 13 / 827,119, which is incorporated herein by reference, can be used to build an error model.

[0338] Figure 43 shows an exemplary distribution of the difference between the measured standard deviation and the expected standard deviation.

[0339] Other relationships can also be employed. For example, see Figure 44, the relationship between the average glucose value and the glucose standard deviation can be seen. Specifically, it can be seen that users with a higher standard deviation tend to have a higher average glucose. This relationship can be used to determine, set, infer, or otherwise select a calibration value. As another example, and referring to Figure 45 , another useful indicator is the patient type. Specifically, Figure 45 illustrates a clear difference in glucose standard deviation between non-diabetic, type I diabetic, and type II diabetic patients. Thus, based on the patient type, the model selected for a patient population can be changed based on the population type, e.g., the standard deviation can be tightened on the CV error model, the mean can be shifted, etc.

[0340] As described above, the system can adjust the data for the time lag of data from a previous time period (e.g., remove the time lag caused by real-time filtering), and can display a graphical representation (e.g., a trend curve graph) of the time-lag adjusted data for the time period. The system can also compensate for the time lag. For example, Figure 46 shows data points separated by a time lag, where Δ represents the individual rate of change between two adjacent points. These time lags can be compensated for by using, for example, the following or a similar equation:

[0341] Glucose(t) = [DriftCorrectedSensor(t) + 5*ROC(t)] / m – b (mg / dL)

[0342] For example, if the current point is in mild noise, then all filtered sensor counts can be employed.

[0343] Example

[0344] An exemplary calibration routine is now described, the steps of which can be seen in the flowchart 400 of Figure 47 . In a first step, the sensitivity distribution is characterized over time in a bench test that measures the sensor response in a glucose solution over one or more days (step 402). Since this is typically a destructive test, it can be run on a representative set of sensors from a manufacturing batch or manufacturing line. This test can be repeated periodically or when the process changes, e.g., when there are changes in raw materials or equipment. A non-destructive bench test can then be used to measure the sensitivity of each sensor (step 404). The results of steps 402 and 404 are used to estimate the in vivo initial and final sensitivity of each sensor (step 406).

[0345] Note that here, destructive testing measures the long-term drift of a group of sensors, e.g., determining that a sensor drifts 10% in the first two days. Non-destructive bench testing provides a starting point for each sensor in the log, e.g., the body sensor may have a starting sensitivity of 20. Combining the two tests can determine, e.g., that the body sensor starts at 20 and is expected to drift 10%, e.g., to reach 22.

[0346] Thus, this step maps or transforms the bench values to in-vivo values with a function that is trained or optimized on well-characterized in-vivo data such as clinical trials. As another example, the body sensor may start at 24 and drift to 26 in vivo.

[0347] During manufacturing, the electrical characteristics of the transmitter (e.g., gain and offset) are calibrated (step 408) and these calibration factors are stored on the transmitter (step 412) so that the raw sensor signal can be corrected for part-to-part variations in the electronics, and then the algorithm is run.

[0348] The sensor is then encapsulated with a single-use transmitter (step 414). The identifier of the sensor (e.g., identification number) is read by means of an optical barcode and its estimated in-vivo sensitivity is retrieved from the manufacturing database and written to the transmitter using, e.g., wireless (NFC or ) communication (step 416).

[0349] When the transmitter first detects that it is connected to a functioning sensor, a session timer is started and the algorithm begins (step 418). The algorithm begins by converting the sensor signal to glucose using a CGM model (step 420), where the model parameters are set to prior information regarding the sensor sensitivity recorded in step 416.

[0350] When a set of representative signal data (e.g., 24 hours) is available, the seed parameters for the forward and reverse filters can be set using the median signal value and an assumed amount of drift (step 422). These seed values are then further optimized to minimize the mean squared error between the two signal filters (step 424).

[0351] The signal-based calibration algorithm uses the forward and reverse filter signals as well as the mean of the raw sensor signal, and in doing so adjusts the model parameters, e.g., sensitivity and baseline, to meet several criteria (step 426). In doing so, time-based input data is used to update the algorithm. An exemplary signal-based calibration algorithm is the algorithm disclosed in the patent application incorporated by reference above (No. 13 / 827,119) and specifically in

[0188] (i.e., Example 4), which describes a Bayesian learning method for drift estimation and correction.

[0352] In Figure 47In an embodiment, the model is adjusted to meet the criterion that the average glucose value is consistent with the expected diabetes mean, and the model can be further adjusted to meet another criterion that the CGM glucose variability is consistent with the average glucose level. To calculate these metrics, the algorithm has an assumed relationship between average glucose and glucose variability. For example, a non-diabetic person may have an average glucose level of 85 mg / dL and a standard deviation of 15 mg / dL. A diabetic patient may have an average glucose of 170 mg / dL and a standard deviation of 65 mg / dL.

[0353] The CGM model uses the optimized model parameters to convert subsequent sensor readings into glucose values (step 428), which are then displayed.

[0354] Steps 424 to 428 are repeated when a new set of signal data is available.

[0355] A similar method can be used to detect an unacceptable amount of sensor change (usually a loss of sensitivity after day 7) and stop displaying readings that may be inaccurate.

[0356] In a preferred embodiment, the analyte sensor is an implantable glucose sensor, such as described in U.S. Patent No. 6,001,067 and U.S. Patent Publication No. US-2005-0027463-A1. In another preferred embodiment, the analyte sensor is a transcutaneous glucose sensor, such as described in U.S. Patent Publication No. US-2006-0020187-A1. In still other embodiments, the sensor is configured to be implanted in a body vessel or in vitro, such as described in U.S. Patent Publication No. US-2007-0027385-A1, co-pending U.S. Patent Application No. 11 / 543,396 filed on October 4, 2006, co-pending U.S. Patent Application No. 11 / 691,426 filed on March 26, 2007, and co-pending U.S. Patent Application No. 11 / 675,063 filed on February 14, 2007. In an alternative embodiment, the continuous glucose sensor includes a transcutaneous sensor such as described in U.S. Patent No. 6,565,509 to Say et al. In another alternative embodiment, the continuous glucose sensor includes a subcutaneous sensor such as described in reference to U.S. Patent No. 6,579,690 to Bonnecaze et al. or U.S. Patent No. 6,484,046 to Say et al. In another alternative embodiment, the continuous glucose sensor includes a refillable subcutaneous sensor such as described in reference to U.S. Patent No. 6,512,939 to Colvin et al. In another alternative embodiment, the continuous glucose sensor includes an intravascular sensor such as described in reference to U.S. Patent No. 6,477,395 to Schulman et al. In another alternative embodiment, the continuous glucose sensor includes an intravascular sensor such as described in reference to U.S. Patent No. 6,424,847 to Mastrototaro et al.

[0357] The connections between the elements shown in the figures illustrate exemplary communication paths. Additional communication paths, which may include direct or via intermediaries, may be included to further facilitate the exchange of information between the elements. The communication paths may be two-way communication paths, allowing the elements to exchange information.

[0358] The various operations of the methods described above may be performed by any suitable means capable of performing the operations, such as various hardware and / or software components, circuits, and / or modules. In general, any operation illustrated in the figures may be performed by a corresponding functional means capable of performing the operation.

[0359] The various illustrative logical blocks, modules, and circuits described in connection with the present disclosure (e.g., Figure 2 and 4The blocks) can be implemented or executed by a general-purpose processor, a digital signal processor (DSP), an application specific integrated circuit (ASIC), a field programmable gate array (FPGA) or other programmable logic device (PLD), discrete gate or transistor logic, discrete hardware components, or any combination thereof designed to perform the functions described herein. The general-purpose processor may be a microprocessor, but in the alternative, the processor may be any commercially available processor, controller, microcontroller, or state machine. The processor may also be implemented as a combination of computing devices, such as a combination of a DSP and a microprocessor, multiple microprocessors, one or more microprocessors in conjunction with a DSP core, or any other such configuration.

[0360] In one or more aspects, the described functionality may be implemented in hardware, software, firmware, or any combination thereof. If implemented in software, the functionality may be stored on or transmitted via a computer-readable medium as one or more instructions or code. The computer-readable medium includes both computer storage media and communication media including any medium that facilitates transfer of a computer program from one place to another. The storage media may be any available media that can be accessed by a computer. By way of example and not limitation, this computer-readable medium may comprise various types of RAM, ROM, CD-ROM or other optical disk storage, magnetic disk storage or other magnetic storage devices, or any other medium that can be used to carry or store the desired program code in the form of instructions or data structures and that can be accessed by a computer. Also, any connection is properly termed a computer-readable medium. For example, if software is transmitted from a website, server, or other remote source using coaxial cable, fiber optic cable, twisted pair, digital subscriber line (DSL), or wireless technologies such as infrared, radio, and microwave, then the coaxial cable, fiber optic cable, twisted pair, DSL, or wireless technologies such as infrared, radio, and microwave are included in the definition of the medium. As used herein, disk and disc include compact disc (CD), laser disc, optical disc, digital versatile disc (DVD), floppy disk and optical disc, where disks typically reproduce data magnetically, while discs reproduce data optically with lasers. Thus, in some aspects, the computer-readable medium may include non-transitory computer-readable media (e.g., tangible media). Additionally, in some aspects, the computer-readable medium may include transitory computer-readable media (e.g., signals). Combinations of the above should also be included within the scope of computer-readable media.

[0361] The methods disclosed herein include one or more steps or acts for implementing the described methods. Without departing from the scope of the claims, the method steps and / or acts may be interchanged with one another. In other words, unless a specific order of the steps or acts is specified, the order and / or use of specific steps and / or acts may be modified without departing from the scope of the claims.

[0362] Certain aspects may include a computer program product for performing the operations presented herein. For example, such a computer program product may include a computer-readable medium having instructions stored (and / or encoded) thereon, which instructions can be executed by one or more processors to perform the operations described herein. For certain aspects, the computer program product may contain packaging material.

[0363] Software or instructions may also be transmitted via a transmission medium. For example, if software is transmitted from a website, server, or other remote source using coaxial cable, fiber optic cable, twisted pair, digital subscriber line (DSL), or wireless technologies such as infrared, radio, and microwave, then the coaxial cable, fiber optic cable, twisted pair, DSL, or wireless technologies such as infrared, radio, and microwave are included in the definition of the transmission medium.

[0364] Furthermore, it should be understood that modules and / or other suitable components for performing the methods and techniques described herein may be downloaded and / or otherwise obtained by a user terminal and / or a base station, where applicable. For example, such a device may be coupled to a server to facilitate the transfer of components for performing the methods described herein. Alternatively, the various methods described herein may be provided via a storage component (e.g., RAM, ROM, a physical storage medium such as a compact disc (CD) or a floppy disk, etc.), such that the user terminal and / or the base station may obtain the various methods after coupling or providing the storage component to the device. Moreover, any other suitable techniques for providing the methods and techniques described herein to the device may be utilized.

[0365] It should be understood that the claims are not limited to the exact configurations and components described above. Various modifications, changes, and variations may be made in the arrangement, operation, and details of the methods and apparatuses described above without departing from the scope of the claims.

[0366] Unless otherwise defined, all terms (including technical and scientific terms) shall be given their ordinary and customary meaning to one of ordinary skill in the art and shall not be limited to a special or customized meaning, unless expressly so defined herein. It should be noted that the use of a particular term when describing certain features or aspects of the present disclosure should not imply that the term is redefined herein to be limited to any specific characteristic that encompasses the features or aspects of the present disclosure associated with the term. The terms and phrases used in this application and their variations, particularly in the appended claims, should be construed as open-ended rather than limiting, unless expressly stated otherwise. As an example of the foregoing, the term 'comprising' should be understood to mean 'including without limitation', 'including but not limited to', or a similar meaning; the term 'including' as used herein is synonymous with 'comprising', 'containing', or 'characterized by', and is inclusive or open-ended and does not exclude additional, unrecited elements or method steps; the term 'having' should be interpreted as 'having at least'; the term 'comprising' should be interpreted as 'comprising but not limited to'; the term 'example' is used to provide an exemplary instance of the item under discussion, rather than an exhaustive or limiting list; adjectives such as terms of 'known', 'normal','standard', and similar meanings should not be construed as limiting the item described to a given time period or to items available with respect to a given time, but should be understood to cover known, normal, or standard techniques that are available or known at present or at any time in the future; and the use of terms such as 'preferably', 'preferred', 'desired', or 'desirable', and words of similar meanings should not be understood to imply that certain features are critical, essential, or even important to the structure or function of the invention, but only to highlight alternative or additional features that may or may not be utilized in a particular embodiment of the invention. Similarly, a group of items associated with the conjunction 'and' should not be understood to require that each and every one of those items be present in the group, but should be understood as 'and / or', unless expressly stated otherwise. Similarly, a group of items associated with the conjunction 'or' should not be understood to require mutual exclusivity among the group, but should be understood as 'and / or', unless expressly stated otherwise.

[0367] Where a range of values is provided, it is understood that the embodiments cover the upper and lower limits of the range and every intervening value therebetween.

[0368] Regarding the use, in general, of any plural and / or singular terms herein, those skilled in the art may translate from plural to singular and / or from singular to plural, as appropriate, in the context and / or applications. For clarity, various singular / plural permutations may be set forth explicitly herein. The indefinite article “a” does not exclude a plurality. A single processor or other unit may fulfill the functions of several items recited in the claims. The mere fact that certain measures are recited in mutually different dependent claims does not indicate that a combination of these measures cannot be used to advantage. Any reference signs in the claims should not be construed as limiting the scope.

[0369] Those skilled in the art will further understand that if a specific number of introduced claim recitations is intended, such intention will be expressly recited in the claims and absent such recitation no such intention exists. For example, as an aid to understanding, the appended claims may contain the use of the introductory phrases "at least one" and "one or more" to introduce claim recitations. However, the use of such phrases should not be construed to imply that the introduction of a claim recitation by the indefinite article "a" limits any particular claim containing such introduced claim recitation to an embodiment containing only one such recitation, even when the same claim includes the introductory phrases "one or more" or "at least one" and an indefinite article such as "a" (e.g., "a" should typically be construed to mean "at least one" or "one or more"); the same holds true for the use of the definite article to introduce a claim recitation. Additionally, even if a specific number of introduced claim recitations is expressly stated, those skilled in the art will recognize that such statement should typically be construed to mean at least the stated number (e.g., the mere statement of "two recitations" without further qualification typically means at least two recitations, or two or more recitations). Further, in those instances where a convention such as "at least one of A, B, and C, etc." is used, generally it is intended that those skilled in the art will understand the significance of the convention, e.g., to include any combination of the listed items, including single members (e.g., "a system having at least one of A, B, and C" will include but not be limited to a system having only A, only B, only C, A and B together, A and C together, B and C together, and / or A, B, and C together, etc.). In those instances where a convention such as "at least one of A, B, or C, etc." is used, generally it is intended that those skilled in the art will understand the significance of the convention (e.g., "a system having at least one of A, B, or C" will include but not be limited to a system having only A, only B, only C, A and B together, A and C together, B and C together, and / or A, B, and C together, etc.). Those skilled in the art will further understand that virtually any disjunctive word and / or phrase presenting two or more alternatives should be understood to contemplate the possibility of one of the items, any one of the items, or both items, whether in the description, claims, or drawings. For example, the phrase "A or B" will be understood to contemplate the possibilities of "A" or "B" or "A and B".

[0370] All numbers expressing quantities of ingredients, reaction conditions, and the like used in the specification should be understood to be modified in all instances by the term "about." Accordingly, unless indicated to the contrary, the numerical parameters set forth herein are approximations that may vary depending upon the desired properties sought to be obtained. At the very least, and not as an attempt to limit the application of the doctrine of equivalents to the scope of any claims in any application claiming the benefit of priority of this application, each numerical parameter should be construed in light of the number of significant digits and the ordinary rounding method.

[0371] All references cited herein are incorporated herein by reference in their entirety. To the extent that the disclosures of incorporated publications and patents or patent applications conflict with the disclosure contained in the specification, the specification is intended to supersede and / or take precedence over any such conflicting material.

[0372] The headings contained herein are for reference and to assist in locating various sections. These headings are not intended to limit the scope of the concepts described therein. These concepts may have applicability throughout the specification.

[0373] In addition, while the foregoing has been described in some detail for purposes of clarity and understanding by way of illustration and example, it will be apparent to those of ordinary skill in the art that certain changes and modifications may be practiced. Accordingly, the description and examples should not be construed as limiting the scope of the invention to the particular embodiments and examples described herein, but rather also cover all modifications and alternative embodiments that fall within the true scope and spirit of the invention.

Claims

1. A method for calibrating an analyte concentration sensor using a signal from an analyte concentration sensor within a biological system, characterized in that, Under steady state, the analyte concentration value within the biological system is a known value, and the method includes: Receiving, on a monitoring device, a seed value of a calibration parameter; Detecting, on the monitoring device, when the analyte concentration value measured by an analyte concentration sensor disposed in the biological system is at steady state; and Correlating, on the monitoring device or on a device or server operatively coupled to the monitoring device, the measurement of the analyte concentration value when it is detected that the analyte concentration value measured by the analyte concentration sensor is at steady state with the known value; After the correlation, receiving a signal from the sensor; and Calculating and displaying a calculated value corresponding to the received signal, the calculated value being based on the received signal, the known value, and the seed value.

2. The method according to claim 1, wherein The received seed value is received from a source including factory calibration information.

3. The method according to claim 1, wherein The method further includes: Detecting a manifestation of the received signal outside of a pre-specified parameter; and Prompting a user to input external calibration information.

4. The method according to claim 3, characterized in that, The displayed calculated value is further based on the external calibration information.

5. The method according to claim 3, wherein The external calibration information is received from SMBG or finger stick calibration.

6. The method according to claim 3, characterized in that, The method further includes resetting the known value to a new known value, the resetting being at least partially based on the external calibration information.

7. The method according to claim 3, wherein The method further includes resetting the seed value to a new seed value, the resetting being at least partially based on the external calibration information.

8. The method according to claim 1, characterized in that The method further includes changing the display based on a determined accuracy of the calculated value.

9. The method according to claim 8, wherein The changing the display includes displaying a range instead of a value, or displaying a value instead of a range.

10. The method according to claim 1, characterized in that, The seed value of the received calibration parameter is a user-inputted characterization of a disease state.

11. The method according to claim 10, wherein The user-inputted characterization of the disease state includes an indication of type I diabetes, type II diabetes, non-diabetes, or pre-diabetes.

12. The method according to claim 1, wherein The seed value of the received calibration parameter is a value based on one or more user-inputted blood glucose values.

13. The method according to claim 1, characterized in that, The display of the calculated value corresponding to the received signal includes displaying a graph or table indicating current measured and historical values of the analyte concentration, and the method further includes: Detecting that a calibration change has occurred; Adjusting, based on the calibration change, one or more calibration parameters of the analyte concentration sensor; and After the adjustment, updating the display of the graph or table indicating current measured and historical values of the analyte concentration according to the adjusted calibration parameters.

14. The method according to claim 13, wherein The detecting that a calibration change has occurred includes: Detecting a change in a slow moving average; or Detecting a change in the known value.

15. The method according to claim 13, wherein The adjustment is configured to occur when the sensor readings are generally stable, or within a predetermined range of readings over a threshold time period, thereby reducing the occurrence of unexpected jumps in the readings.

16. An analyte concentration sensor system, characterized in that, The system includes: An analyte sensor configured to be calibrated in a biological system using a signal from the analyte sensor, wherein under steady state, the analyte concentration value within the biological system is a known value; A processor including components of a monitoring device or a device or server operatively coupled to the monitoring device, the processor being configured to: Receive a seed value of calibration parameters; Detect when an analyte concentration value measured by an analyte concentration sensor disposed in a biological system is in a steady state; Correlate a measurement of the analyte concentration value when the analyte concentration value measured by the analyte concentration sensor is detected to be in a steady state with the known value; after the correlation, receive a signal from the sensor; and Calculate a calculated value corresponding to the received signal, the calculated value being based on the received signal, the known value, and the seed value; and A display device, wherein the display device is configured to display the calculated value.

17. The system according to claim 16, wherein, The received seed value is received from a source including factory calibration information.

18. The system according to claim 16, wherein, The processor is configured to: Detect a performance of the received signal outside of a predetermined parameter; and Prompt a user to input external calibration information.

19. The system according to claim 18, wherein The processor is configured to reset the known value to a new known value, the reset being at least partially based on the external calibration information.

20. The system according to claim 18, wherein The processor is configured to reset the seed value to a new seed value, the reset being at least partially based on the external calibration information.

21. The system according to claim 16, wherein The processor is configured to change the display based on a determined accuracy of the calculated value.

22. The system according to claim 16, wherein, The display device is configured to display a graph or table indicating current measured and historical values of the analyte concentration, and the processor is further configured to: Detect that a calibration change has occurred; Adjust one or more calibration parameters of the analyte concentration sensor according to the calibration change; and after the adjustment, Update the display of the graph or table indicating current measured and historical values of the analyte concentration according to the adjusted calibration parameters.

23. The system according to claim 22, wherein The processor is configured to detect that the calibration change has occurred by detecting a change in a slow moving average or detecting a change in the known value.

24. The system according to claim 16, characterized in that, The processor is configured to adjust one or more calibration parameters at a time when sensor readings are generally stable or within a predetermined range of readings of a threshold time period, thereby reducing the occurrence of unexpected jumps in the readings.

25. A method for calibrating and compensating for drift in an analyte concentration sensor using only signals from an implanted analyte concentration sensor within a biological system, wherein in a steady state, an analyte concentration value within the biological system is a known value, the method comprising: a. On a monitoring device, detect when an analyte concentration value measured by an analyte concentration sensor disposed in a biological system is in a steady state; b. On the monitoring device or on a device or server operatively coupled to the monitoring device, correlate a measurement of the analyte concentration value when the analyte concentration value measured by the analyte concentration sensor is detected to be in a steady state with the known value; c. Determine a first slow moving average of the analyte concentration values measured in a first time period, and calibrate the sensor at least partially based on the first slow moving average and based on the known value; d. After the determination in step c, determine a second slow moving average of the analyte concentration values measured in a second time period; and e. Adjust the calibration of the sensor at least partially based on the difference between the first slow moving average and the second slow moving average.

26. The method according to claim 25, wherein the adjustment of the calibration is configured to occur when the sensor readings are substantially stable, or within a predetermined range of readings over a threshold time period, thereby reducing the occurrence of unexpected jumps in the readings.

Images