Method and apparatus for distinguishing between clinically significant changes and artifacts in patient physiological information

• US 8,070,677 B2
• Filed: 05/31/2006
• Issued: 12/06/2011
• Est. Priority Date: 06/09/2005
• Status: Active Grant

• Current Assignee: Koninklijke Philips Electronics N.V. (Koninklijke Philips N.V.)
• Original Assignee: Koninklijke Philips Electronics N.V. (Koninklijke Philips N.V.)
• Inventors: Ali, Walid S. I.
• Primary Examiner(s): Johnson, III, Henry M
• Assistant Examiner(s): Archer, Marie
• Application Number:
  US11/916,767
• Publication Number:
  US 20080208012A1
• Time in Patent Office: 2,015 Days
• Field of Search: 600300-301, 600544-547, 600372-385, 600/481, 600/429, 600/336, 128/920, 128900-902, 128921-925, 700/2
• US Class Current: 600/301
• CPC Class Codes: G06F 2218/18 by plotting the signal segments against each other, e.g. analysing scattergrams
• CPC Class Codes: G16H 40/63 for local operation

Abstract

A patient physiological information monitoring system includes a plurality of patient monitoring devices (6) and a physiological information analyzer (2). The plurality of patient monitoring devices (6) monitor physiological information from a patient and generate corresponding physiological signals. The physiological information analyzer (2) processes the monitored physiological information and determines whether a physiological change is a clinically significant event or an artifact. The physiological information analyzer includes at least one receiver (4) that receives the physiological signals from the patient monitoring devices; a signal correlator (10) that generates morphograms from pairs of the received physiological signals; a signature generator (12) that applies a wavelet decomposition to each morphogram to compute a signature for each morphogram; and a decision component (14) that compares the morphogram signatures within and across sampling intervals and determines if a physiological change is a clinically significant change or an artifact.

Claims

1. A patient physiological information monitoring system, comprising:

a plurality of patient monitoring devices that monitor physiological information from a patient and generate corresponding physiological signals; and

a physiological information analyzer including one or more processors that processes the monitored physiological information and determines whether a physiological change is a clinically significant event or an artifact, the one or more processors being programmed to;

receive the physiological signals from the patient monitoring devices;

generate morphograms of one physiological signal versus another physiological signal from pairs of the received physiological signals within sampling intervals;

apply a wavelet decomposition to each morphogram to compute a signature for each mapping morphogram;

compare the morphogram signatures within and across the sampling intervals for a shape perturbation; and

determine whether the shape perturbation is a result of a clinically significant change or an artifact.

2. The patient physiological information monitoring system as set forth in claim 1, wherein the one or more processors are further programmed to:

determine that the physiological change is an artifact in response to the shape perturbation being only in a subset of the signatures with at least one common signal.

3. The patient physiological information monitoring system as set forth in claim 1, wherein the one or more processors are further programmed to:

determine that an alarm is valid in response to the shape perturbation which indicates a physiological change.

4. The patient physiological information monitoring system as set forth in claim 1, further including:

a display component that visually presents at least some of the signatures for observation by a clinical staff.

5. The patient physiological information monitoring system as set forth in claim 1, wherein the one or more processors are further programmed to:

discard signatures that include an artifact and stores signatures that include clinically significant physiological changes in a storage component.

6. The patient physiological information monitoring system as set forth in claim 1, wherein the wavelet is a Symlet wavelet.

7. The patient physiological information monitoring system as set forth in claim 1, wherein the signature of each morphogram is represented by one or more coefficients that fully describe a shape variability of the morphogram.

8. The patient physiological information monitoring system as set forth in claim 7, wherein the one or more coefficients describe a horizontal component, a vertical component, and a diagonal component.

9. The patient physiological information monitoring system as set forth in claim 1, wherein the wavelet decomposition includes a plurality of levels of decomposition.

10. A patient physiological information monitoring system comprising:

a plurality of patient monitoring devices that monitor physiological information from a patient and generate corresponding physiological signals; and

a physiological information analyzer including one or more processors that processes the monitored physiological information and determines whether a physiological change is a clinically significant event or an artifact, the one or more processor being programmed to;

receive the physiological signals sampled by the patient monitoring devices during successive sampling intervals;

generate morphograms from pairs of the received physiological signals in each of the sampling intervals;

compute a signature for each morphogram; and

compare the morphogram signatures of a current and prior sampling intervals for a shape perturbation to determine if a physiological change is a clinically significant change or an artifact,

wherein the sampling intervals occur at a sampling rate that is less than or equal to an alarm latency period of the corresponding patient monitor device, the alarm latency period corresponding to a time over which an alarm condition persists in order to trigger an alarm with a high degree of correctness.

11. The patient physiological information monitoring system as set forth in claim 10, wherein the monitored physiological information includes at least one or more electrocardiogram (ECG) signals and an arterial blood pressure (ABP) signal.

12. The patient physiological information monitoring system as set forth in claim 10, wherein the physiological information analyzer sounds an alarm and/or prematurely invokes a monitoring device to alarm when it determines a clinically significant physiological change has occurred.

13. The patient physiological information monitoring system as set forth in claim 10, wherein the physiological information analyzer prevents or turns off a patient monitoring device alarm when the analyzer determines a monitored physiological change is an artifact.

14. A method for determining whether a physiological change captured by a patient monitoring device is a clinically significant change or an artifact, comprising:

receiving physiological signals captured by a patient monitoring device indicative of physiological information monitored from a patient;

with one or more processors, generating a mapping of one physiological signal versus another physiological signal for pairs of the physiological signals as the physiological information for each sampling interval is received;

with the one or more processors, applying a wavelet decomposition on each mapping to generate a signature for each mapping;

with the one or more processors, comparing the mapping signatures within and across sampling intervals for signature perturbations;

with the one or more processors, determining whether the signature perturbations are a result of a clinically significant change or an artifact; and

with the one or more processors, in response to determining that the signature perturbations are the result of an artifact, overriding an alarm of the patient monitoring device; and

with the one or more processors, in response to determining that the signature perturbations are the result of a clinically significant change, invoking the alarm of the patient monitoring device.

15. The method as set forth in claim 14, wherein the wavelet is a Symlet wavelet.

16. The method as set forth in claim 14, wherein each signature is a unique descriptor represented as coefficients that describe a shape variability of a corresponding mapping.

17. The method as set forth in claim 15, further including displaying at least the signatures for observation by clinical staff.

18. A non-transitory computer readable medium having program code embodied therein that when executed by one or more processors, causes the one or more processor to perform a method comprising:

receiving physiological signals captured by a patient monitoring device indicative of physiological information monitored from a patient;

with one or more processors, generating a mapping of one physiological signal versus another physiological signal for pairs of the physiological signals as the physiological information for each sampling interval is received;

with the one or more processors, applying a wavelet decomposition on each mapping to generate a signature for each mapping;

with the one or more processors, comparing the mapping signatures within and across sampling intervals for signature perturbations;

with the one or more processors, determining whether the signature perturbations are a result of a clinically significant change or an artifact; and

with the one or more processors, in response to determining that the signature perturbations are the result of an artifact, overriding an alarm of the patient monitoring device; and

with the one or more processors, in response to determining that the signature perturbations are the result of a clinically significant change, invoking the alarm of the patient monitoring device.

19. A patient monitoring system, comprising:

a patient monitoring device;

a computer programmed to;

generate morphogram signatures of related patient monitoring signals versus each other from the patient monitoring device;

decompose the morphogram signatures using wavelet decomposition;

compare changes in the decomposed morphogram signatures relative to each other or over time; and

based on the morphogram signature comparisons, determine whether a clinically significant event or an artifact has occurred.

20. The patient monitoring system as set forth in claim 19, wherein the patient monitor includes an alarm and the computer is further programmed to:

invoke the alarm in response to an event being determined clinically significant; and

override the alarm in response to determining that an artifact has occurred.