SIGNAL PROCESSING FOR CONTINUOUS ANALYTE SENSOR

• US 20100030484A1
• Filed: 09/23/2009
• Published: 02/04/2010
• Est. Priority Date: 12/09/2003
• Status: Active Grant

Abstract

Systems and methods for dynamically and intelligently estimating analyte data from a continuous analyte sensor, including receiving a data stream, selecting one of a plurality of algorithms, and employing the selected algorithm to estimate analyte values. Additional data processing includes evaluating the selected estimative algorithms, analyzing a variation of the estimated analyte values based on statistical, clinical, or physiological parameters, comparing the estimated analyte values with corresponding measure analyte values, and providing output to a user. Estimation can be used to compensate for time lag, match sensor data with corresponding reference data, warn of upcoming clinical risk, replace erroneous sensor data signals, and provide more timely analyte information encourage proactive behavior and preempt clinical risk.

Claims

1. A method for estimating an analyte value from a continuous analyte sensor, the method comprising:

receiving a data stream from the continuous analyte sensor for a first time period; and

estimating at least one analyte value for a second time period based on the data stream, wherein the step of estimating at least one analyte value comprises selecting an algorithm from a plurality of algorithms based at least in part on an analysis of the data stream.

2. The method of claim 1, the step of selecting an algorithm is conditional upon at least one value selected from the group consisting of analyte concentration, rate of change, acceleration, and an individual historical pattern of the data stream.

3. The method of claim 1, wherein the step of selecting an algorithm comprises employing a plurality of algorithms on the data stream and determining which of the plurality of algorithms best correlates with the data stream.

4. The method of claim 1, wherein the plurality of algorithms comprise an algorithm selected from the group consisting of polynomial regression, autoregressive algorithms, Fourier transform, Wavelet transform, neural network-based mapping, fuzzy logic based pattern matching, and Genetic-Algorithms based pattern matching.

5. The method of claim 1, wherein the step of selecting an algorithm further comprises applying a physiological boundary to the selected algorithm.

6. The method of claim 1, further comprising a step of evaluating a selected algorithm by applying an evaluation function prior to employing the selected algorithm to estimate the analyte value, wherein the evaluation function is selected from the group consisting of a data association function, a curvature formula, and a physiological boundary.

7. The method of claim 1, wherein the second time period is a future time period.

8. A system for estimating an analyte value from a continuous analyte sensor, the system comprising:

an input module operatively connected to the continuous analyte sensor that receives a data stream for a first time period; and

a processor module comprising programming configured to select an algorithm from a plurality of algorithms based on an analysis of the data stream, wherein the processor further comprises programming configured to estimate at least one analyte value using the selected algorithm.

9. The system of claim 8, wherein the programming configured to select an algorithm from a plurality of algorithms selects the algorithm conditional upon at least one value selected from the group consisting of analyte concentration, rate of change, acceleration, and an individual historical pattern of the data stream.

10. The system of claim 8, wherein the programming configured to select the an algorithm from a plurality of algorithms employs a plurality of algorithms on the data stream and determines which of the plurality of algorithms best correlates with the data stream.

11. The system of claim 8, wherein the plurality of algorithms comprise an algorithm selected from the group consisting of polynomial regression, autoregressive algorithms, Fourier transform, Wavelet transform, neural network-based mapping, fuzzy logic based pattern matching, and Genetic-Algorithms based pattern matching.

12. The system of claim 8, wherein the programming configured to select an algorithm from a plurality of algorithms applies a physiological boundary to the selected algorithm.

13. The system of claim 8, wherein the processor module further comprises programming configured to evaluate a selected algorithm by applying an evaluation function prior to employing the selected algorithm to estimate the analyte value, wherein the evaluation function is selected from the group consisting of a data association function, a curvature formula, and a physiological boundary.

14. The system of claim 8, wherein the second time period is a future time period.

15. A method for estimating an analyte value from a continuous analyte sensor, the method comprising:

receiving a data stream from the continuous analyte sensor for a first time period;

selecting at least one algorithm from a plurality of algorithms based on an analysis of the data stream; and

evaluating the algorithm based on at least one parameter selected from the group consisting of a statistical parameter, a physiological parameter, and a clinical parameter; and

employing the selected algorithm based on the step of evaluating to estimate at least one analyte value.

16. The method of claim 15, wherein the step of employing the selected algorithms comprises estimating at least one analyte value for a second time period based on the data stream.

17. The method of claim 16, wherein the second time period is a future time period.

18. A system for estimating analyte values from a continuous analyte sensor the system comprising:

an input module operatively connected to the continuous analyte sensor that receives a data stream for a first time period; and

a processor module comprising programming configured to select an algorithm from a plurality of algorithms based on an analysis of the data stream, that evaluates the algorithm based on a parameter selected from the group consisting of statistical parameters, physiological parameters, and clinical parameters, and that employs a selected algorithm based on the algorithm evaluation to estimate at least one analyte value.

19. The system of claim 18, wherein the programming configured to employ the selected algorithms estimates at least one analyte value for a second time period based on the data stream.

20. The system of claim 19, wherein the second time period is a future time period.

21. The system of claim 19, wherein selecting an algorithm is conditional upon at least one value selected from the group consisting of analyte concentration, rate of change, acceleration, and an individual historical pattern of the data stream.

22. The system of claim 19, wherein the processor is configured to select an algorithm by employing the plurality of algorithms on the data stream and determining which of the plurality of algorithms best correlates with the data stream.

23. The system of claim 19, wherein the processor is configured to select an algorithm by employing at least one of polynomial regression, autoregressive algorithms, Fourier transform, Wavelet transform, neural network-based mapping, fuzzy logic based pattern matching, and Genetic-Algorithms based pattern matching.

24. The system of claim 19, wherein the processor is configured to select an algorithm by applying a physiological boundary to the selected algorithm.

25. The system of claim 19, wherein the processor is configured to evaluate the selected algorithm by applying an evaluation function prior to employing the selected algorithm to estimate the analyte value, wherein the evaluation function is selected from the group consisting of a data association function, a curvature formula, and a physiological boundary.