System and/or method for glucose sensor calibration

• US 9,289,168 B2
• Filed: 12/29/2008
• Issued: 03/22/2016
• Est. Priority Date: 12/29/2008
• Status: Active Grant

• Current Assignee: Medtronic Minimed Incorporated (Medtronic Plc)
• Original Assignee: Medtronic Minimed Incorporated (Medtronic Plc)
• Inventors: Keenan, Desmond Barry, Mastrototaro, John J.
• Primary Examiner(s): Riggs, II, Larry D
• Application Number:
  US12/345,477
• Publication Number:
  US 20100162786A1
• Time in Patent Office: 2,640 Days
• US Class Current: 1/1
• CPC Class Codes: A61B 5/14532 for measuring glucose, e.g. by tissue impedance measurement
• CPC Class Codes: A61B 5/1495 Calibrating or testing of in-vivo probes
• CPC Class Codes: A61B 5/6849 in combination with a needle set

Abstract

The subject matter disclosed herein relates to systems, methods and/or devices for calibrating sensor data to be used in estimating a blood glucose concentration. A relationship between sensor measurements and reference readings may be used to estimate a relationship between sensor measurements and blood glucose concentration. Such sensor measurements may be weighted according to a decreasing function of uncertainty associated with sensor values.

Claims

1. A method comprising:

obtaining samples of an electrical signal generated by a sensor, said samples having sample values responsive to a blood-glucose concentration in a patient;

obtaining blood glucose reference measurement values responsive to said blood-glucose concentration;

pairing at least some of the sample values with temporally associated blood glucose reference measurement values;

for at least one of said paired sample values, weighting said paired sample value based, at least in part, on a measure of statistical dispersion of said electrical signal, said measure of statistical dispersion of said electrical signal being a function of a blood glucose reference measurement value paired with said sample value;

determining one or more parameters of an estimator of said blood-glucose concentration based, at least in part, on said weighted paired sample values; and

controlling a pump to infuse a substance to said patient based, at least in part, on said determined one or more parameters of said estimator of said blood-glucose concentration.

2. The method of claim 1, wherein said determining said one or more parameters comprises determining said one or more parameters based, at least in part, on a linear regression of said weighted paired sample values and temporally associated blood-glucose reference measurement values.

3. The method of claim 2, wherein said determining said one or more parameters of said estimator further comprises:

calculating a linear regression sensitivity ratio based, at least in part, on said weighted paired sample values and temporally associated blood-glucose reference measurement values;

selecting an offset based, at least in part, on said calculated linear regression sensitivity ratio; and

calculating a modified linear regression sensitivity ratio based, at least in part, on said selected offset, said weighted paired sample values and said associated blood-glucose reference measurement values.

4. The method of claim 1, wherein said measure of statistical dispersion comprise a variance and/or approximation of a variance of said paired sample values as a function of said temporally associated blood glucose reference measurement values.

5. The method of claim 1, wherein said at least some of said paired sample values are weighted according to an inverse of said measure of statistical dispersion of said paired sample values.

6. The method of claim 1, and further comprising:

estimating a linear relationship of said measure of statistical dispersion of said electrical signal versus blood glucose concentration; and

deriving said one or more parameters based, at least in part, on said linear relationship.

7. The method of claim 1, wherein individually weighting said at least some of said paired sample values further comprises further weighting said paired sample values based on how recently said ample values are obtained.

8. The method of claim 1, and further comprising detecting a failure of said sensor based, at least in part, on a change in said determined one or more parameters.

9. The method of claim 1, and further comprising calibrating measurements from said sensor for measuring a blood-glucose concentration based, at least in part, on said determined one or more parameters.

10. The method of claim 1, wherein said individually weighting said at least some of said paired sample values comprises weighting said at least some of said paired sample values according to a decreasing function of blood glucose reference measurement values temporally associated with said weighted paired sample values.

11. An article comprising:

a non-transitory storage medium comprising machine-readable instructions stored thereon which are executable by a computing platform to;

obtain samples of an electrical signal generated by a sensor, said samples having sample values responsive to a blood-glucose concentration in a patient;

obtain blood glucose reference measurement values indicative of said blood-glucose concentration;

pair at least some of said sample values with temporally associated blood glucose reference measurement values;

for at least one of said paired sample values, weighting said paired sample value based, at least in part, on a measure of statistical dispersion of said electrical signal, said measure of statistical dispersion of said electrical signal being a function of a blood glucose reference sample value paired with said sample value;

determining one or more parameters of an estimator of said blood-glucose concentration based, at least in part, on said weighted paired sample values; and

control a pump to infuse a substance to said patient based, at least in part, on said determined one or more parameters of said estimator of said blood-glucose concentration.

12. The article of claim 11, wherein said instructions are further executable by said computing platform to determine said one or more parameters based, at least in part, on a linear regression of said weighted paired sample values and temporally associated blood-glucose reference measurement values.

13. The article of claim 12, wherein said instructions are further executable by said computing platform to:

calculate a linear regression sensitivity ratio based, at least in part, on said weighted paired sample values and temporally associated blood-glucose reference measurement values;

select an offset based, at least in part, on said calculated linear regression sensitivity ratio; and

calculate a modified linear regression sensitivity ratio based, at least in part, on said selected offset, said weighted paired sample values and said temporally associated blood-glucose reference measurement values.

14. The article of claim 11, wherein said measure of statistical dispersion comprises a variance and/or approximation of a variance of said sample values as a function of said temporally associated blood glucose reference measurement values.

15. The article of claim 11, wherein said at least some of said weighted paired sample values are weighted according to an inverse of said measure of statistical dispersion of said weighted paired sample values.

16. The article of claim 11, wherein said instructions are further executable by said computing platform to:

estimate a linear relationship of said measure of statistical dispersion of said electrical signal versus blood glucose concentration; and

derive said function based, at least in part, on said linear relationship.

17. The article of claim 11, wherein instructions are further executable by said computing platform to individually weight said paired sample values based on how recently said sample values are obtained.

18. The article of claim 11, wherein said instructions are further executable by said computing platform to detect a failure of said sensor based, at least in part, on a change in said determined one or more parameters.

19. The article of claim 11, wherein said instructions are further executable by said computing platform to calibrate measurements from said sensor for measuring a blood-glucose concentration based, at least in part, on said determined one or more parameters.

20. The article of claim 11, wherein said instructions are further executable by said computing platform to individually weight said at least some of said paired sample values according to a decreasing function of blood glucose reference values temporally associated with said paired sample values.

21. An apparatus comprising:

a sensor to generate one or more sensor signals responsive to a blood glucose concentration in a body;

a computing platform to;

obtain samples of said one or more signals, said samples having sample values responsive to said blood-glucose concentration in a patient;

obtain blood glucose reference measurement values responsive to said blood-glucose concentration;

for at least one of said paired sample values, weighting said paired sample value based, at least in part, on a measure of statistical dispersion of said electrical signal, said measure of statistical dispersion of said one or more sensor signals being a function of a blood glucose reference sample value paired with said sample value;

determine one or more parameters of an estimator of said blood-glucose concentration based, at least in part, on said weighted paired sample values; and

control a pump to infuse a substance to said patient based, at least in part, on said determined one or more parameters of said estimator of said blood-glucose concentration.

22. The apparatus of claim 21, wherein said computing platform is further to determine said one or more parameters based, at least in part, on a linear regression of said weighted paired sample values and temporally associated blood-glucose reference measurement values.

23. The apparatus of claim 22, wherein said computing platform is further to:

calculate a linear regression sensitivity ratio based, at least in part, on said weighted paired sample values and temporally associated blood-glucose reference measurement values;

select an offset based, at least in part, on said calculated linear regression sensitivity ratio; and

calculate a modified linear regression sensitivity ratio based, at least in part, on said selected offset, said weighted paired sample values and said temporally associated blood-glucose reference measurement values.

24. The apparatus of claim 21, wherein said measure of statistical dispersion comprises a variance and/or approximation of a variance of said sample values as a function of said temporally associated blood glucose reference measurement values.

25. The apparatus of claim 21, wherein said paired sample values are weighted according to an inverse of said measure of statistical dispersion of said one or more sensor signals.

26. The apparatus of claim 21, wherein said computing platform is further to:

estimate a linear relationship of said measure of statistical dispersion of said one or more sensor signals versus blood glucose concentration; and

derive said one or more parameters based, at least in part, on said linear relationship.

27. The apparatus of claim 21, wherein said computing platform is further to further weight said paired sample values based on how recently said paired sample values are obtained from said sensor.

28. The apparatus of claim 21, wherein said computing platform is further to detect a failure of said sensor based, at least in part, on a change in said estimated one or more parameters.

29. The apparatus of claim 21, wherein said computing platform is further to calibrate measurements from said sensor for measuring a blood-glucose concentration based, at least in part, on said estimated one or more parameters.

30. The apparatus of claim 21, wherein said computing platform is further to individually weight said at least some of said sample values by weighting said at least some of said sample values according to a decreasing function of blood glucose reference measurement values temporally associated with said weighted sample values.