Residual based monitoring of human health

• US 8,795,170 B2
• Filed: 11/29/2006
• Issued: 08/05/2014
• Est. Priority Date: 11/29/2005
• Status: Active Grant

• Current Assignee: Prolaio, Inc.
• Original Assignee: Venture Gain LLC
• Inventors: Pipke, Robert Matthew
• Primary Examiner(s): Jackson, Gary
• Assistant Examiner(s): JIAN, SHIRLEY XUEYING
• Application Number:
  US11/564,679
• Publication Number:
  US 20070149862A1
• Time in Patent Office: 2,806 Days
• Field of Search: 703/11
• US Class Current: 600/301
• CPC Class Codes: A61B 5/0205 Simultaneously evaluating both cardiovascular conditions and different types of body conditions, e.g. heart and respiratory condition
• CPC Class Codes: A61B 5/02055 Simultaneously evaluating both cardiovascular condition and temperature
• CPC Class Codes: A61B 5/021 Measuring pressure in heart or blood vessels A61B5/0205 takes precedence
• CPC Class Codes: A61B 5/02455 provided with high/low alarm devices
• CPC Class Codes: A61B 5/0816 Measuring devices for examining respiratory frequency measuring frequency of electric signals G01R23/00
• CPC Class Codes: A61B 5/11 Measuring movement of the entire body or parts thereof, e.g. head or hand tremor, mobility of a limb for measuring pulse A61B5/02 ;A61B5/1038 takes precedence; motion detection to correct for motion artifacts in physiological signals A61B5/721
• CPC Class Codes: A61B 5/145 Measuring characteristics of blood in vivo, e.g. gas concentration, pH value; Measuring characteristics of body fluids or tissues, e.g. interstitial fluid, cerebral tissuemeasuring of blood pressure or blood flow A61B5/02; non-radiation detecting or locating of foreign bodies in blood A61B5/06
• CPC Class Codes: A61B 5/14532 for measuring glucose, e.g. by tissue impedance measurement
• CPC Class Codes: A61B 5/14551 for measuring blood gases
• CPC Class Codes: A61B 5/318 Heart-related electrical modalities, e.g. electrocardiography [ECG]
• CPC Class Codes: A61B 5/352 Detecting R peaks, e.g. for synchronising diagnostic apparatus; Estimating R-R interval
• CPC Class Codes: A61B 5/7264 Classification of physiological signals or data, e.g. using neural networks, statistical classifiers, expert systems or fuzzy systems neural networks per se G06N3/00; expert systems per se G06N5/00
• CPC Class Codes: A61B 5/7267 involving training the classification device
• CPC Class Codes: G06F 18/22 Matching criteria, e.g. proximity measures
• CPC Class Codes: G06F 9/545 where tasks reside in different layers, e.g. user- and kernel-space
• CPC Class Codes: G16B 5/00 ICT specially adapted for modelling or simulations in systems biology, e.g. gene-regulatory networks, protein interaction networks or metabolic networks
• CPC Class Codes: G16B 5/30 Dynamic-time models
• CPC Class Codes: G16B 99/00 Subject matter not provided for in other groups of this subclass
• CPC Class Codes: G16H 50/20 for computer-aided diagnosis, e.g. based on medical expert systems
• CPC Class Codes: G16H 50/30 for calculating health indices; for individual health risk assessment
• CPC Class Codes: G16H 50/50 : for simulation or modelling of medical disorders
• CPC Class Codes: G16H 50/70 : for mining of medical data, e.g. analysing previous cases of other patients

Abstract

Improved human health monitoring is provided in the context of sensor measurements of typical vital signs and other biological parameters, by a system and method using an empirical model of the parameters and disposed to estimate values of the parameters in response to actual measurements. Residuals resulting from the difference between the estimates and actual measurements are analyzed for robust indications of incipient health issues. Residual analysis is both more robust and more sensitive than conventional univariate range checking on vital signs.

Claims

1. A method for monitoring the health status of a human, comprising the steps of:

obtaining an input observation of multiple vital signs from the human, a vital sign being a direct sensor measurement of a biological signal or a transformation thereof;

providing in a computer accessible memory a set of exemplary observations of said multiple vital signs characteristic of the interrelationships between said multiple vital signs under normal health conditions for the monitored human;

generating with a computer an estimate of at least one of said multiple vital signs as a function of said input observation, using a linear combination of at least some of the exemplary observations, determined from a kernel-based comparison of the input observation and at least some exemplary observations;

differencing with a computer said estimate and the corresponding measured value of the at least one of said multiple vital signs in said input observation to yield a residual value upon which a determination is made related to the health status of the human.

2. The method according to claim 1, further comprising the step of using said input observation of multiple vital signs to localize on a subset of said exemplary observations to generate said estimate.

3. The method according to claim 2, wherein said subset is determined with a computer by a kernel-based comparison of said input observation with said exemplary observations.

4. The method according to claim 3, wherein said localizing kernel-based comparison of said input observation with said exemplary observations marks for inclusion in said subset a predetermined number of those exemplary observations that have a highest kernel value when compared with said input observation.

5. The method according to claim 3, wherein said localizing kernel-based comparison of said input observation with said exemplary observations marks for inclusion in said subset those exemplary observations that have a kernel value above a predetermined threshold and that contain at least one vital sign value that is closest and less than, or closest and greater than the corresponding vital sign value of said input observation, from among all exemplary observations.

6. The method according to claim 1, wherein said residual value is compared to a threshold with a computer in order to make said determination of the health status of the human.

7. The method according to claim 6, wherein the determination of the health status of the human is made based on a sequence of comparisons of said residual value to said threshold for successive said input observations of multiple vital signs.

8. The method according to claim 6, wherein residuals are compared to respective thresholds for a plurality of said multiple vital signs and the results of the comparisons are used in combination to make said determination of the health status of the human.

9. The method according to claim 1, wherein said step of generating an estimate is performed using a similarity-based model.

10. The method according to claim 1, wherein said step of generating an estimate is performed using a kernel regression estimator.

11. The method according to claim 1, wherein said kernel-based comparison in said estimate generating step is determined based on a norm of the vector difference between its multivariate input observations treated as vectors.

12. The method according to claim 11, wherein the output value K_h of said kernel-based comparison is determined according to:

13. The method according to claim 11, wherein the output value K_h of said kernel-based comparison is determined according to:

14. The method according to claim 11, wherein the output value K_h of said kernel-based comparison is determined according to:

15. The method according to claim 1, wherein said kernel-based comparison in said estimate generating step is determined based on respective differences between like elements of its multivariate input observations.

16. The method according to claim 15, wherein the output value K_h of said kernel-based comparison is determined at least in part based on an elemental kernel value according to:

17. The method according to claim 15, wherein the output value K_h of said kernel-based comparison is determined at least in part based on an elemental kernel value according to:

18. The method according to claim 15, wherein the output value K_h of said kernel-based comparison is determined at least in part based on an elemental kernel value according to:

19. The method according to claim 1, wherein said multiple vital signs includes a measure of heart rate variability.

20. The method according to claim 1, wherein said multiple vital signs includes a measure of a time lag between a respiratory activity and resulting oxygenation changes in the blood.

21. The method according to claim 1, wherein said multiple vital signs includes a measure of a time lag between a QRS event and a subsequent blood pressure event.

22. The method according to claim 1, wherein said multiple vital signs includes a measure of the difference between the ambient air temperature and the temperature of a location of the body.

23. The method according to claim 1, wherein said multiple vital signs includes a measure of breathing depth.

24. The method according to claim 1, further comprising the step of testing whether the input observation is characteristic of at least one activity state, in order to use the input observation for monitoring of the human.

25. The method according to claim 24, wherein said at least one activity state includes a state of resting.

26. The method according to claim 24, wherein said testing step comprises classifying the activity state of the input observation using a kernel-based inferential model.

27. The method according to claim 24, wherein said testing step comprises classifying the activity state of the input observation using a clustering technique to identify which of a predetermined set of data clusters the input observation is associated with.

28. The method according to claim 24, wherein said at least one activity state includes a state of sleeping.

29. The method according to claim 1, wherein said set of exemplary observations of said multiple vital signs comprises a plurality of subsets each characteristic of a distinct human baseline class, and further comprising the step of selecting for use in said estimate generating step a subset appropriate to the human baseline class to which the human being monitored belongs.

30. The method according to claim 29, wherein each said human baseline class is distinguished at least by a measure of age.

31. The method according to claim 29, wherein each said human baseline class is distinguished at least by gender.

32. The method according to claim 29, wherein each said human baseline class is distinguished at least by a measure of weight.

33. The method according to claim 29, wherein each said human baseline class is distinguished at least by medical condition.

34. The method according to claim 1, wherein at least some of said set of exemplary observations have been previously acquired from the human who is monitored.

35. The method according to claim 1, further comprising the step of:

filtering said input observation to exclude it from use in monitoring the health status of the human if the input observation coincides with an interval which exhibits a disrupted respiratory pattern.

36. The method according to claim 35, wherein said filtering step excludes said input observation if it coincides with an interval which exhibits at least one of the set of coughing and talking.

37. The method according to claim 1, wherein said multiple vital signs includes a measure of blood glucose.

38. The method according to claim 1, wherein said multiple vital signs includes a measure of activity.

39. The method according to claim 1, further comprising the step of adapting said stored set of exemplary observations of said multiple vital signs to include an input observation.

40. The method according to claim 39, further comprising the steps of:

providing a web page interface for displaying to a user notifications of any deviation of the health of the human based on tests of residuals; and

adapting said stored set of exemplary observations of said multiple vital signs to include at least one monitored observation from data corresponding to a notification identified by the user for adaptation.

41. The method according to claim 1, wherein the determination of the health status of the human is made based on testing with a computer sequences of residuals with a sequential probability ratio test.

42. The method according to claim 1, further comprising the step of providing a web page interface for displaying to a user notifications of any deviation of the health of the human based on tests of residuals, said web page interface configured to enable a user of the system to annotate any notification as corresponding to one of the set of statuses selected from the group consisting of (a) the notification is being monitored and investigated;

(b) the notification is overruled; and

(c) the notification is confirmed.

43. The method according to claim 1, wherein said stored set of exemplary observations of said multiple vital signs is characteristic of the behavior of at least three parameters selected from a heart rate, a blood oxygenation value, a respiration metric, a blood pressure metric, and a temperature metric.

44. The method according to claim 43, wherein input observations of said multiple vital signs are made every 1 minute, and said multiple vital signs comprise at least a measure of average heart rate, a measure of average respiration rate, and a measure of average blood pressure.