METHOD, SYSTEM, AND COMPUTER PROGRAM PRODUCT FOR IMPROVING THE ACCURACY OF GLUCOSE SENSORS USING INSULIN DELIVERY OBSERVATION IN DIABETES

• US 20130079613A1
• Filed: 03/24/2011
• Published: 03/28/2013
• Est. Priority Date: 03/26/2010
• Status: Active Grant

• Current Assignee: University of Virginia Patent Foundation (University of Virginia)
• Original Assignee: University of Virginia Patent Foundation (University of Virginia)
• Inventors: Kovatchev, Boris P., Breton, Marc D., Patek, Stephen D., Karvetski, Colleen H.
• Granted Patent: US 9,398,869 B2
• Time in Patent Office: Days
• US Class Current: 600/365
• CPC Class Codes: A61B 2560/0223 of calibration, e.g. protocols for calibrating sensors
• 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 5/4839 combined with drug delivery
• CPC Class Codes: A61B 5/7225 Details of analog processing, e.g. isolation amplifier, gain or sensitivity adjustment, filtering, baseline or drift compensation A61B5/30, A61B5/316 take precedence; transformers for specific applications H01F38/00; amplifiers H03F
• CPC Class Codes: A61B 5/725 using specific filters therefor, e.g. Kalman or adaptive filters A61B5/316 take precedence
• CPC Class Codes: A61M 1/36 Other treatment of blood in a by-pass of the natural circulatory system, e.g. temperature adaptation, irradiation ; Extra-corporeal blood circuits
• CPC Class Codes: A61M 5/142 Pressure infusion, e.g. using pumps

Abstract

Method and System for providing a signal from an insulin pump, artificial pancreas, or another insulin delivery device as a source of information for improving the accuracy of a continuous glucose sensor (CGS). The effect of using insulin information to enhance sensor accuracy is most prominent at low blood glucose levels, i.e. in the hypoglycemic range, which is critical for any treatment. A system for providing a filtering/state estimation methodology that may be used to determine a glucose state estimate at time t-τ. The estimation may be extrapolated to some future time t and then the extrapolated value is used to extract the blood glucose component. The blood glucose component of the extrapolation and the output of the CGS are weighted and used to estimate the blood glucose level of a subject.

Claims

1. A method for improving the accuracy of a glucose measurement device, said method comprising:

using insulin delivery information to improve the accuracy of said glucose measurement device.

2. The method of claim 1, wherein said glucose measurement device is a continuous glucose device.

3. The method of claim 1, further comprising:

obtaining a readout from said glucose measurement device;

obtaining said insulin delivery information from an insulin delivery device;

performing a first intermediary computation to obtain an output of said first intermediary computation using said readout from said glucose measurement device of the subject and said insulin delivery information; and

performing a second intermediary computation to obtain an output of said second intermediary computation using said output of said first intermediary computation.

4. The method of claim 3, wherein said glucose measurement device is a continuous glucose device.

5. The method of claim 3, wherein said first intermediary computation further comprises a Kalman Filter methodology.

6. The method of claim 3, wherein said first intermediary computation further comprises a method for filtering/state estimation, such as a H-infinity filtering method, a Bayesian filtering method, a Monte Carlo method, or a least squares method.

7. The method of claim 3, wherein said readout from said glucose measurement device provides a frequent glucose readout.

8. The method of claim 7, wherein the frequency of said readout from said glucose measurement device is at least approximately 60 minute intervals.

9. The method of claim 7, wherein the frequency of said readout from said glucose measurement device is at least approximately 30 minute intervals.

10. The method of claim 7, wherein the frequency of said readout from said glucose measurement device is at least approximately 15 minute intervals.

11. The method of claim 3, wherein said insulin delivery information is provided by an insulin pump, artificial pancreas, or another insulin delivery device.

12. The method of claim 3, wherein said insulin delivery information is provided by a measuring device.

13. The method of claim 3, wherein said second intermediary computation further comprises:

inferring said output of said second intermediary computation from said output of said first intermediary computation.

14. The method of claim 13, wherein inferring said second intermediary computation further comprises:

extrapolating said output of said first intermediary computation to some future time to determine an extrapolated state vector output.

15. The method of claim 14, wherein determining said extrapolated output is provided by {circumflex over (x)}(t|t−

τ

)=A^τ

{circumflex over (x)}(t−

τ

)+A(τ

)Bu(t−

τ

)+A(τ

)Gω

(t−

τ

), wherein

{circumflex over (x)}(t|t−

τ

) refers to the extrapolated state vector output;

A is a state space matrix;

B is a state space matrix;

G is a state space matrix;

A^τ

=A·

A . . . A, i.e. the τ

-fold composition of said state space matrix A;

A(τ

)=Σ

_s=0<sup>τ

−

1</sup>A^s, with A(0)=0<sub>8×

8</sub>.

16. The method of claim 15, wherein said state space matrix A is:

17. The method of claim 15, wherein said state space matrix B is:

B^T=[−

3.05×

10<sup>−

9</sup> 1.34×

10<sup>−

10</sup> 0.9900 0.0100 6.50×

10<sup>−

5</sup> −

4.61×

10<sup>−

11</sup> 0 0].

18. The method of claim 15, wherein said state space matrix G is:

G^T=[6.76×

10<sup>−

7</sup> 0 0 0 0 1.52×

10<sup>−

8</sup> 0.9534 0.0464].

19. The method of claim 14, wherein said second intermediary computation further comprises:

extracting a blood glucose state from said extrapolated state vector output.

20. The method of claim 19, wherein extracting said blood glucose state from said extrapolated value further comprises:

applying a state space matrix to said extrapolated state vector output.

21. The method of claim 3, further comprising:

performing a third computation to reach an estimate of the blood glucose of the subject by applying a weighting scheme to said readout from the glucose measurement device and said output of said second intermediary computation.

22. The method of claim 21, wherein said weighting scheme further comprises:

applying a first weighting coefficient to said readout from the glucose measurement device and applying a second weighting coefficient to said output of said second intermediary computation.

23. The method of claim 22, wherein the weighting yields a linear combination of the readout from the glucose measurement device and said output of said second intermediary so that the sum of the first weighting coefficient and the second weighting coefficient is equal to 1.

24. The method of claim 22, wherein said weighting scheme further comprises:

applying more weight to said output of said second intermediary computation during hypoglycemia.

25. The method of claim 24, wherein said more weight is approximately 1.

26. The method of claim 22, wherein said weighting scheme further comprises:

applying more weight to said readout from the glucose measurement device during euglycemia.

27. The method of claim 26, wherein said more weight is approximately 1.

28. The method of claim 22, wherein said weighting scheme further comprises:

applying more weight to said readout from the glucose measurement device during hyperglycemia.

29. The method of claim 28, wherein said more weight is approximately 1.

30. The method of claim 22, wherein said weighting scheme further comprises:

applying a weighting scheme that accounted for the lack of accuracy of said output of said second intermediary computation immediately following meals and meal insulin boluses.

31. The method of claim 22, wherein said weighting scheme further comprises:

applying more weight on said readout from the glucose measurement device in the time following a detection of a large bolus of insulin.

32. The method of claim 22, wherein said weighting scheme further comprises:

basing said weighting scheme on an S-shaped curve with said curve increasing at approximately the threshold of hypoglycemia.

33. The method of claim 22, wherein said first weighting coefficient and said second weighting coefficient are provided by the formula w₁=LBGI(CGS−

50)/100 where LBGI is a low BG index and CGS is said readout from said continuous glucose measurement device.

34. A system for improving the accuracy of a glucose measurement device, wherein:

said glucose measurement device uses insulin delivery information to improve the accuracy of said glucose measurement device.

35. The system of claim 34, wherein said glucose measurement device is a continuous glucose device.

36. The system of claim 34, further comprising:

said glucose measurement device configured to provide a readout;

an insulin delivery device configured to provide said insulin delivery information; and

a processor, wherein said processor is configured to provide;

a first intermediary computation that provides an output of said first intermediary computation using said readout from said glucose measurement device of the subject and said insulin delivery information, and

a second intermediary computation that provides an output of said second intermediary computation using said output of said first intermediary computation.

37. The system of claim 36, wherein said glucose measurement device is a continuous glucose device.

38. The system of claim 36, wherein said first intermediary computation further comprises a Kalman Filter methodology.

39. The system of claim 36, wherein said first intermediary computation further comprises a method for filtering/state estimation, such as a H-infinity filtering method, a Bayesian filtering method, a Monte Carlo method, or a least squares method.

40. The system of claim 36, wherein said readout from said glucose measurement device provides a frequent glucose readout.

41. The system of claim 40, wherein the frequency of said readout from said glucose measurement device is at least approximately 60 minute intervals.

42. The system of claim 40, wherein the frequency of said readout from said glucose measurement device is at least approximately 30 minute intervals.

43. The system of claim 40, wherein the frequency of said readout from said glucose measurement device is at least approximately 15 minute intervals.

44. The system of claim 36, wherein said insulin delivery information is provided by an insulin pump, artificial pancreas, or another insulin delivery device.

45. The system of claim 36, wherein said insulin delivery information is provided by a measuring device.

46. The system of claim 36, wherein said second intermediary computation further comprises:

an inference of said output of said second intermediary computation from said output of said first intermediary computation.

47. The system of claim 46, wherein inferring said second intermediary computation further comprises:

an extrapolation of said output of said first intermediary computation to some future time to determine an extrapolated state vector output.

48. The system of claim 47, wherein said extrapolated output is determined by

{circumflex over (x)}(t|t−

τ

)=A^τ

{circumflex over (x)}(t−

τ

)+A(τ

)Bu(t−

τ

)+A(τ

)Gω

(t−

τ

),

wherein

{circumflex over (x)}(t|t−

τ

) refers to the extrapolated state vector output;

A is a state space matrix;

B is a state space matrix;

G is a state space matrix;

A′

=A·

A . . . A, i.e. the τ

-fold composition of said state space matrix A;

A(τ

)=Σ

_s=0<sup>τ

−

1</sup>A^s, with A(0)=0<sub>8×

8</sub>.

49. The system of claim 48, wherein said state space matrix A is:

50. The system of claim 48, wherein said state space matrix B is:

B^T=[−

3.05×

10<sup>−

9</sup> 1.34×

10<sup>−

10</sup> 0.9900 0.0100 6.50×

10<sup>−

5</sup> −

4.61×

10<sup>−

11</sup> 0 0].

51. The system of claim 48, wherein said state space matrix G is:

G^T=[6.76×

10<sup>−

7</sup> 0 0 0 0 1.52×

10<sup>−

8</sup> 0.9534 0.0464].

52. The system of claim 47, wherein said second intermediary computation further comprises:

an extraction of a blood glucose state from said extrapolated state vector output.

53. The system of claim 52, wherein said extraction of said blood glucose state from said extrapolated value further comprises:

an application of a state space matrix to said extrapolated state vector output.

54. The system of claim 36, wherein said processor is configured to provide a third computation to reach an estimate of the blood glucose of the subject by applying a weighting scheme to said readout from the glucose measurement device and said output of said second intermediary computation.

55. The system of claim 54, wherein said weighting scheme further comprises:

an application of a first weighting coefficient to said readout from the glucose measurement device and applying a second weighting coefficient to said output of said second intermediary computation.

56. The system of claim 55, wherein the weighting yields a linear combination of the readout from the glucose measurement device and said output of said second intermediary so that the sum of the first weighting coefficient and the second weighting coefficient is equal to 1.

57. The system of claim 55, wherein said weighting scheme further comprises:

an application of more weight to said output of said second intermediary computation during hypoglycemia.

58. The system of claim 57, wherein said more weight is approximately 1.

59. The system of claim 55, wherein said weighting scheme further comprises:

an application of more weight to said readout from the glucose measurement device during euglycemia.

60. The system of claim 59, wherein said more weight is approximately 1.

61. The system of claim 55, wherein said weighting scheme further comprises:

an application of more weight to said readout from the glucose measurement device during hyperglycemia.

62. The system of claim 61, wherein said more weight is approximately 1.

63. The system of claim 55, wherein said weighting scheme further comprises:

an application of a weighting scheme that accounted for the lack of accuracy of said output of said second intermediary computation immediately following meals and meal insulin boluses.

64. The system of claim 55, wherein said weighting scheme further comprises:

an application of more weight on said readout from the glucose measurement device in the time following a detection of a large bolus of insulin.

65. The system of claim 55, wherein said weighting scheme further comprises:

an application of said weighting scheme based on an S-shaped curve with said curve increasing at approximately the threshold of hypoglycemia.

66. The system of claim 55, wherein said first weighting coefficient and said second weighting coefficient are provided by the formula w₁=LBGI(CGS−

50)/100 where LBGI is a low BG index and CGS is said readout from said continuous glucose measurement device.

67. A computer program product comprising a computer useable medium having a computer program logic for enabling at least one processor in a computer system for improving the accuracy of a glucose measurement device, said computer program logic comprising:

using insulin delivery information to improve the accuracy of said glucose measurement device.

68. The computer program product of claim 67, wherein said glucose measurement device is a continuous glucose device.

69. The computer program product of claim 67, wherein said computer logic further comprises:

obtaining a readout from said glucose measurement device;

obtaining said insulin delivery information from an insulin delivery device;

performing a first intermediary computation to obtain an output of said first intermediary computation using said readout from said glucose measurement device of the subject and said insulin delivery information; and

performing a second intermediary computation to obtain an output of said second intermediary computation using said output of said first intermediary computation.

70. The computer program product of claim 69, wherein said glucose measurement device is a continuous glucose device.

71. The computer program product of claim 69, wherein said first intermediary computation further comprises a Kalman Filter methodology.

72. The computer program product of claim 69, wherein said first intermediary computation further comprises a method for filtering/state estimation, such as a H-infinity filtering method, a Bayesian filtering method, a Monte Carlo method, or a least squares method.

73. The computer program product of claim 69, wherein said readout from said glucose measurement device provides a frequent glucose readout.

74. The computer program product of claim 73, wherein the frequency of said readout from said glucose measurement device is at least approximately 60 minute intervals.

75. The computer program product of claim 73, wherein the frequency of said readout from said glucose measurement device is at least approximately 30 minute intervals.

76. The computer program product of claim 73, wherein the frequency of said readout from said glucose measurement device is at least approximately 15 minute intervals.

77. The computer program product of claim 69, wherein said insulin delivery information is provided by an insulin pump, artificial pancreas, or another insulin delivery device.

78. The computer program product of claim 69, wherein said insulin delivery information is provided by a measuring device.

79. The computer program product of claim 69, wherein said second intermediary computation further comprises:

inferring said output of said second intermediary computation from said output of said first intermediary computation.

80. The computer program product of claim 79, wherein inferring said second intermediary computation further comprises:

extrapolating said output of said first intermediary computation to some future time to determine an extrapolated state vector output.

81. The computer program product of claim 80, wherein determining said extrapolated output is provided by

{circumflex over (x)}(t|t−

τ

)=A^τ

{circumflex over (x)}(t−

τ

)+A(τ

)Bu(t−

τ

)+A(τ

)Gω

(t−

τ

),

wherein

{circumflex over (x)}(t|t−

τ

) refers to the extrapolated state vector output;

A is a state space matrix;

B is a state space matrix;

G is a state space matrix;

A(τ

)=A·

A . . . A, i.e. the τ

-fold composition of said state space matrix A;

A(τ

)=Σ

_s=0<sup>τ

−

1</sup>A^s, with A(0)=0<sub>8×

8</sub>.

82. The computer program product of claim 81, wherein said state space matrix A is:

83. The computer program product of claim 81, wherein said state space matrix B is:

B^T=[−

3.05×

10<sup>−

9</sup> 1.34×

10<sup>−

10</sup> 0.9900 0.0100 6.50×

10<sup>−

5</sup> −

4.61×

10<sup>−

11</sup> 0 0].

84. The computer program product of claim 81, wherein said state space matrix G is:

G^T=[6.76×

10<sup>−

7</sup> 0 0 0 0 1.52×

10<sup>−

11</sup> 0.9534 0.0464].

85. The computer program product of claim 80, wherein said second intermediary computation further comprises:

extracting a blood glucose state from said extrapolated state vector output.

86. The computer program product of claim 85, wherein extracting said blood glucose state from said extrapolated value further comprises:

applying a state space matrix to said extrapolated state vector output.

87. The computer program product of claim 69, further comprising:

performing a third computation to reach an estimate of the blood glucose of the subject by applying a weighting scheme to said readout from the glucose measurement device and said output of said second intermediary computation.

88. The computer program product of claim 87, wherein said weighting scheme further comprises:

applying a first weighting coefficient to said readout from the glucose measurement device and applying a second weighting coefficient to said output of said second intermediary computation.

89. The computer program product of claim 88, wherein the weighting yields a linear combination of the readout from the glucose measurement device and said output of said second intermediary so that the sum of the first weighting coefficient and the second weighting coefficient is equal to 1.

90. The computer program product of claim 88, wherein said weighting scheme further comprises:

applying more weight to said output of said second intermediary computation during hypoglycemia.

91. The computer program product of claim 90, wherein said more weight is approximately 1.

92. The computer program product of claim 88, wherein said weighting scheme further comprises:

applying more weight to said readout from the glucose measurement device during euglycemia.

93. The computer program product of claim 92, wherein said more weight is approximately 1.

94. The computer program product of claim 88, wherein said weighting scheme further comprises:

applying more weight to said readout from the glucose measurement device during hyperglycemia.

95. The computer program product of claim 94, wherein said more weight is approximately 1.

96. The computer program product of claim 88, wherein said weighting scheme further comprises:

applying a weighting scheme that accounted for the lack of accuracy of said output of said second intermediary computation immediately following meals and meal insulin boluses.

97. The computer program product of claim 88, wherein said weighting scheme further comprises:

applying more weight on said readout from the glucose measurement device in the time following a detection of a large bolus of insulin.

98. The computer program product of claim 88, wherein said weighting scheme further comprises:

basing said weighting scheme on an S-shaped curve with said curve increasing at approximately the threshold of hypoglycemia.

99. The computer program product of claim 88, wherein said first weighting coefficient and said second weighting coefficient are provided by the formula w₁=LBGI(CGS−

50)/100 where LBGI is a low BG index and CGS is said readout from said continuous glucose measurement device.