Temperature prediction system and method

• US 20060224349A1
• Filed: 04/01/2005
• Published: 10/05/2006
• Est. Priority Date: 04/01/2005
• Status: Active Grant

• Current Assignee: Carefusion 303 Incorporated (Becton, Dickinson & Co)
• Original Assignee: Cardinal Health 303 Incorporated (Cardinal Health, Inc.)
• Inventors: Butterfield, Robert D.
• Granted Patent: US 7,318,004 B2
• Time in Patent Office: Days
• US Class Current: 702/130
• CPC Class Codes: G01K 7/42 Circuits effecting compensation of thermal inertia; Circuits for predicting the stationary value of a temperature

Abstract

A thermometer system and method that rapidly predict body temperature based on the temperature signals received from a temperature sensing probe when it comes into contact with the body. A nonlinear, multi-parameter curve fitting process is performed and depending on the errors in the curve fit, parameters are changed or a prediction of the temperature is made. Criteria exist for the differences between the curve fit and the temperature data. The processor switches to a Continuous Monitor State if the curve fit over a limited number of time frames is unacceptable. Determining the start time on which the measurement time frame for prediction is based is performed by tissue contact threshold coupled with a prediction time delay.

Claims

1. A thermometer for determining the temperature of a subject, comprising:

a sensor that provides temperature signals in response to sensing the temperature of the subject, the temperature signals varying in time; and

a processor configured to;

monitor the temperature signals;

adapt a variable shape, nonlinear, multi-parameter model temperature curve to the monitored temperature signals;

compare the model temperature curve to the monitored temperature signals; and

if the differences between the model temperature curve and the monitored temperature signals do not exceed a threshold, predict the temperature of the subject based on the model curve.

2. The thermometer of claim 1 wherein the processor is further configured to determine the differences between the model temperature curve and the monitored temperature signals through a sum squared error.

3. The thermometer of claim 1 wherein the configuration of the processor in adapting the model temperature curve comprises:

4. The thermometer of claim 3 wherein the processor is further configured to determine the differences between the model temperature curve and the monitored temperature signals by means of sum squared error and uses the sum squared error in selecting one or more of the parameters.

5. The thermometer of claim 4 wherein the processor uses the sum squared error in selecting the A and B parameters given the C and E parameters.

6. The thermometer of claim 3 wherein the processor is further configured to determine the differences between the model temperature curve and the monitored temperature signals by means of sum squared error and uses the sum squared error in selecting the values of all of the selectable parameters.

7. The thermometer of claim 1 wherein the processor is further configured to:

monitor the temperature signals over a first time frame; and

select a second time frame within which to monitor the temperature signals in the event that the comparison of the model temperature curve to the monitored temperature signals shows the differences to exceed the threshold.

8. The thermometer of claim 7 wherein the processor is further configured to select the second time frame to overlap the first time frame.

9. The thermometer of claim 7 wherein the processor is further configured to select the second time frame to not overlap the first time frame.

10. The thermometer of claim 1 wherein the processor is further configured to compare the nonlinear curve fitting to a predetermined integrity criterion and if the predicted temperature from the curve fitting does not meet the integrity criterion, the processor excludes the monitored temperature data from a prediction of the subject'"'"'s temperature.

11. The thermometer of claim 1 wherein the processor is further configured to compare characteristics of the monitored temperature data to integrity criteria and if any characteristic of the monitored temperature data does not satisfy the integrity criteria, the processor does not use the monitored temperature data to predict the temperature of the subject.

12. The thermometer of claim 11 wherein one of the integrity criteria comprises a curvature quality of the monitored temperature data.

13. The thermometer of claim 12 wherein the curvature quality is determined by the existence of a section of a curve based on the actual temperature measurement data that is considered for use in estimation of temperature that is not concave downward.

14. The thermometer of claim 11 wherein one of the integrity criteria comprises a slope limit of the monitored temperature data.

15. The thermometer of claim 14 wherein the slope limit comprises a maximum slope limit.

16. The thermometer of claim 1 wherein the processor is further configured to compare curvature quality of the monitored temperature signals to a quality threshold and if the curvature quality exceeds the quality threshold, the processor will not predict the temperature of the subject based on the data forming the basis for the curvature quality comparison.

17. The thermometer of claim 1 wherein the processor is further configured to compare a slope of the monitored temperature signals to a slope threshold and if the slope exceeds the slope threshold, the processor will not predict the temperature of the subject based on the data forming the basis for the slope comparison.

18. The thermometer of claim 1 wherein one of the multiple parameters for the nonlinear model used by the processor is a curvature index.

19. The thermometer of claim 18 wherein the curvature index is constrained to a family of values appropriate to a measurement site.

20. A method for determining the temperature of a subject, comprising:

sensing the temperature of the subject;

providing temperature signals based on the sensing, the temperature signals varying in time;

monitoring the temperature signals;

adapting a variable shape, nonlinear, multi-parameter model temperature curve to the monitored temperature signals;

comparing the model temperature curve to the monitored temperature signals; and

if the differences between the model temperature curve and the monitored temperature signals do not exceed a threshold, predicting the temperature of the subject based on the model curve.

21. The method of claim 20 further comprising determining the differences between the model temperature curve and the monitored temperature signals through a sum squared error.

22. The method of claim 20 wherein adapting the model temperature curve comprises:

23. The method of claim 22 wherein determining the differences between the model temperature curve and the monitored temperature signals comprises taking a sum squared error and using the sum squared error in selecting one or more of the parameters.

24. The method of claim 25 further comprising using the sum squared error in selecting the A and B parameters given the C and E parameters.

25. The method of claim 22 further comprising determining the differences between the model temperature curve and the monitored temperature signals by means of sum squared error and using the sum squared error in selecting the values of all of the selectable parameters.

26. The method of claim 20 further comprising:

monitoring the temperature signals over a first time frame; and

selecting a second time frame within which to monitor the temperature signals in the event that the comparison of the model temperature curve to the monitored temperature signals shows the differences to exceed the threshold.

27. The method of claim 26 wherein selecting a second time frame comprises selecting the second time frame to overlap the first time frame.

28. The method of claim 26 wherein selecting the second time frame comprises selecting the second time frame to not overlap the first time frame.

29. The method of claim 20 further comprising comparing the nonlinear curve fitting to a predetermined integrity criterion and if the predicted temperature from the curve fitting does not meet the integrity criterion, excluding the monitored temperature data from a prediction of the subject'"'"'s temperature.

30. The method of claim 20 further comprising comparing characteristics of the monitored temperature data to integrity criteria and if any characteristic of the monitored temperature data does not satisfy the integrity criteria, excluding the monitored temperature data from prediction of the temperature of the subject.

31. The method of claim 30 wherein one of the integrity criteria comprises a curvature quality of the monitored temperature data.

32. The method of claim 31 wherein the curvature quality is determined by the existence of a section of a curve based on the actual temperature measurement data that is considered for use in estimation of temperature that is not concave downward.

33. The method of claim 30 wherein one of the integrity criteria comprises a slope limit of the monitored temperature data.

34. The method of claim 33 wherein the slope limit comprises a maximum slope limit.

35. The method of claim 20 further comprising comparing curvature quality of the monitored temperature signals to a quality threshold and if the curvature quality exceeds the quality threshold, not predicting the temperature of the subject based on the data forming the basis for the curvature quality comparison.

36. The method of claim 20 further comprising comparing a slope of the monitored temperature signals to a slope threshold and if the slope exceeds the slope threshold, not predicting the temperature of the subject based on the data forming the basis for the slope comparison.

37. The method of claim 20 wherein adapting a variable shape, nonlinear, multi-parameter model temperature curve comprises selecting a curvature index as one of the multiple parameters for the nonlinear model.

38. The method of claim 37 wherein the curvature index is constrained to a family of values appropriate to a measurement site.