Calibration of glucose monitoring sensor and/or insulin delivery system

• US 9,089,292 B2
• Filed: 03/26/2010
• Issued: 07/28/2015
• Est. Priority Date: 03/26/2010
• Status: Active Grant

• Current Assignee: Medtronic Minimed Incorporated (Medtronic Plc)
• Original Assignee: Medtronic Minimed Incorporated (Medtronic Plc)
• Inventors: Roy, Anirban, Keenan, Barry, Kremliovsky, Michael
• Primary Examiner(s): BERHANU, ETSUB D
• Application Number:
  US12/748,341
• Publication Number:
  US 20110237917A1
• Time in Patent Office: 1,950 Days
• Field of Search: 600/365
• US Class Current: 1/1
• CPC Class Codes: A61B 2560/0223 of calibration, e.g. protocols for calibrating sensors
• CPC Class Codes: A61B 5/1451 for interstitial fluid
• CPC Class Codes: A61B 5/14532 for measuring glucose, e.g. by tissue impedance measurement
• CPC Class Codes: A61B 5/14865 invasive, e.g. introduced into the body by a catheter or needle or using implanted sensors
• CPC Class Codes: A61B 5/1495 Calibrating or testing of in-vivo probes
• CPC Class Codes: A61B 5/4839 combined with drug delivery
• CPC Class Codes: A61B 5/6849 in combination with a needle set
• CPC Class Codes: A61B 5/7203 for noise prevention, reduction or removal
• CPC Class Codes: A61B 5/7221 Determining signal validity, reliability or quality A61B5/7207 and A61B5/7217 take precedence
• CPC Class Codes: A61B 5/725 using specific filters therefor, e.g. Kalman or adaptive filters A61B5/316 take precedence
• CPC Class Codes: A61B 5/7275 Determining trends in physiological measurement data; Predicting development of a medical condition based on physiological measurements, e.g. determining a risk factor
• CPC Class Codes: A61M 2005/1726 the body parameters being measured at, or proximate to, the infusion site
• CPC Class Codes: A61M 5/1723 using feedback of body parameters, e.g. blood-sugar, pressure measurement of body parameters A61B5/00

Abstract

Disclosed are methods, apparatuses, etc. for calibrating glucose monitoring sensors and/or insulin delivery systems. In certain example embodiments, blood glucose reference samples may be correlated with sensor measurements with regard to a delay associated with the sensor measurements. In certain other example embodiments, one or more parameters of a probability model may be estimated based on blood glucose reference sample-sensor measurement pairs. Based on such information, function(s) for estimating a blood-glucose concentration in a patient may be determined.

Claims

1. A method comprising:

obtaining, from a blood-glucose sensor, one or more blood glucose sensor measurements;

obtaining one or more reference sample signals based at least in part on one or more blood glucose reference samples;

combining, via a processor, said one or more reference sample signals and said one or more blood glucose measurements to provide a plurality of blood glucose reference sample-sensor measurement pairs;

estimating one or more parameters of a plurality of probability models based, at least in part, on said blood glucose reference sample-sensor measurement pairs;

determining a plurality of functions for estimating a blood-glucose concentration in a patient from sensor measurements based, at least in part, on said estimated one or more parameters of said plurality of probability models;

for each of said plurality of functions, estimating a sensor glucose concentration value by application of said function to provide a plurality of estimated sensor glucose concentration values, each of said estimated sensor glucose concentration values being generated by a corresponding one of said plurality of functions; and

determining a composite sensor glucose concentration value for the patient based, at least partly, on a combination of said plurality of estimated sensor glucose concentration values.

2. The method of claim 1, wherein said obtaining comprises:

taking sensor measurements for said plurality of blood glucose reference sample-sensor measurement pairs using at least one subcutaneous current sensor.

3. The method of claim 1, wherein one or more of said functions are defined at least partly by a slope and an offset.

4. The method of claim 1, wherein said one or more parameters for at least one of said functions comprise a calfactor variable and an offset variable.

5. The method of claim 4, wherein a relationship between current sensor measurements corresponding to said blood glucose reference sample-sensor measurement pairs and sensor glucose concentration for the patient may be represented by a linear model, said linear model associated with a slope and an offset;

and wherein said slope and said offset of said linear model are determinable from said calfactor variable and/or said offset variable.

6. The method of claim 1, wherein said estimating one or more parameters of at least one of said probability models further comprises:

estimating said one or more parameters of said at least one of said probability models using a linear Kalman filter estimator in which process noise and measurement noise are modeled as being constant.

7. The method of claim 1, wherein said estimating one or more parameters of at least one of said probability models further comprises:

estimating said one or more parameters of said at least one of said probability models using a linear Kalman filter estimator in which process noise is adapted as a function of model performance.

8. The method of claim 1, wherein said estimating one or more parameters of at least one of said probability models further comprises:

estimating said one or more parameters of said at least one of said probability models using a linear Kalman filter estimator in which measurement noise is adapted as a function of model performance.

9. The method of claim 1, wherein said estimating one or more parameters of at least one of said probability models further comprises:

estimating said one or more parameters of said at least one of said probability models using a Bayesian estimator in which sensor measurements corresponding to said blood glucose reference sample-sensor measurement pairs comprise an independent variable and blood glucose reference samples corresponding to said blood glucose reference sample-sensor measurement pairs comprise a measured variable.

10. The method of claim 1, wherein said determining a composite sensor glucose concentration value for the patient comprises:

weighting said plurality of sensor glucose concentration values based, at least in part, on a plurality of quality indicators indicative of an accuracy of said plurality of probability models.

11. The method of claim 10, wherein said plurality of quality indicators comprise a plurality of likelihood values;

and wherein said method further comprises;

calculating said plurality of likelihood values based, at least partly, on error differences between sensor glucose concentration values estimated from said plurality of probability models and reference blood glucose values from blood glucose reference samples.

12. The method of claim 1, wherein determining said composite sensor glucose concentration value comprises:

fractionally weighting said estimated sensor glucose concentration values based, at least in part, on likelihood values based on performance of said plurality of functions; and

combining said fractionally weighted estimated glucose concentration values to provide said composite sensor glucose concentration value.

13. The method of claim 1, wherein said plurality of probability models are to estimate said one or more signals to represent said one or more parameters relatively or effectively simultaneously.

14. An apparatus comprising:

a calibration unit to receive one or more signals based on blood-glucose sensor measurements, said calibration unit comprising one or more processors programmed to;

obtain a plurality of blood glucose reference sample-sensor measurement pairs;

estimate one or more parameters of a plurality of probability models based, at least in part, on said blood glucose reference sample-sensor measurement pairs;

determine a plurality of functions for estimating a blood-glucose concentration in a patient from sensor measurements based, at least in part, on said estimated one or more parameters of said plurality of probability models;

for each of said plurality of functions, estimate a sensor glucose concentration value by application of said function to provide a plurality of estimated sensor glucose concentration values, each of said estimated sensor glucose concentration values being generated by a corresponding one of said plurality of functions; and

determine a composite sensor glucose concentration value for the patient based, at least in part, on a combination of said plurality of estimated sensor glucose concentration values.

15. The apparatus of claim 14, wherein said one or more processors are further programmed to obtain said plurality of blood glucose reference sample-sensor measurement pairs by taking sensor measurements for said plurality of blood glucose reference sample-sensor measurement pairs using at least one subcutaneous current sensor.

16. The apparatus of claim 14, wherein at least one of said functions is defined at least partly by a slope and an offset.

17. The apparatus of claim 14, wherein at least one of said functions is defined, at least in part, by a calfactor variable and an offset variable.

18. The apparatus of claim 17, wherein a relationship between current sensor measurements corresponding to said blood glucose reference sample-sensor measurement pairs and sensor glucose concentration for the patient may be represented by a linear model, said linear model associated with a slope and an offset;

and wherein said slope and said offset of said linear model are determinable from said calfactor variable and/or said offset variable.

19. The apparatus of claim 14, wherein said one or more processors are further programmed to estimate said one or more parameters of at least one of said probability models by:

estimating said one or more parameters using a linear Kalman filter estimator in which process noise and measurement noise are modeled as being constant.

20. The apparatus of claim 14, wherein said one or more processors are further programmed to estimate said one or more parameters of at least one of said probability models by:

estimating said one or more parameters using a linear Kalman filter estimator in which process noise is adapted as a function of model performance.

21. The apparatus of claim 14, wherein said one or more processors are further programmed to estimate said one or more parameters of at least one of said probability models by:

estimating said one or more parameters using a linear Kalman filter estimator in which measurement noise is adapted as a function of model performance.

22. The apparatus of claim 14, wherein said one or more processors are further programmed to estimate said one or more parameters of at least one of said probability models by:

estimating said one or more parameters using a Bayesian estimator in which sensor measurements corresponding to said blood glucose reference sample-sensor measurement pairs comprise an independent variable and blood glucose reference samples corresponding to said blood glucose reference sample-sensor measurement pairs comprise a measured variable.

23. The apparatus of claim 14, wherein said one or more processors are programmed to determine said composite sensor glucose concentration value for the patient by:

weighting said plurality of sensor glucose concentration values based, at least in part, on a plurality of quality indicators indicative of an accuracy of said plurality of probability models.

24. The apparatus of claim 23, wherein said plurality of quality indicators comprise a plurality of likelihood values;

and wherein said one or more processors are further programmed to;

calculate said plurality of likelihood values based, at least partly, on error differences between sensor glucose concentration values estimated from said plurality of probability models and reference blood glucose values from blood glucose reference samples.

25. The apparatus of claim 14, further comprising:

one or more blood-glucose sensors adapted to be coupled to the patient to obtain blood-glucose sensor measurements and adapted to provide said one or more signals based on said blood-glucose sensor measurements, wherein said one or more processors are further programmed to obtain said plurality of blood glucose reference sample-sensor measurement pairs via said one or more blood-glucose sensors using said one or more signals.

26. A system comprising:

means for obtaining a plurality of blood glucose reference sample-sensor measurement pairs;

means for estimating one or more parameters of a plurality of probability models based, at least in part, on said blood glucose reference sample-sensor measurement pairs;

means for determining a plurality of functions for estimating a blood-glucose concentration in a patient from sensor measurements based, at least in part, on said estimated one or more parameters of said plurality of probability models;

for each of said plurality of functions, means for estimating a sensor glucose concentration value by application of said function to provide a plurality of estimated sensor glucose concentration values, each of said estimated sensor glucose concentration values being generated by a corresponding one of said plurality of functions; and

means for determining a composite sensor glucose concentration value for the patient based, at least partly, on a combination of said plurality of estimated sensor glucose concentration values.

27. An article comprising:

at least one non-transitory storage medium having stored thereon instructions executable by one or more processors to;

obtain a plurality of blood glucose reference sample-sensor measurement pairs;

estimate one or more parameters of a plurality of probability models based, at least in part, on said blood glucose reference sample-sensor measurement pairs;

determine a plurality of functions for estimating a blood-glucose concentration in a patient from sensor measurements based, at least in part, on said estimated one or more parameters of said plurality of probability models;

for each of said plurality functions estimate a sensor glucose concentration value by application of said function to provide a plurality of estimated sensor glucose concentration values, each of said estimated sensor glucose concentration values being generated by a corresponding one of said plurality of functions; and

determine a composite sensor glucose concentration value for the patient based, at least part, on a combination of said plurality of estimated sensor glucose concentration values.