Clinical decision support system for patient-ventilator asynchrony detection and management
The present disclosure describes a system that automatically detects patient-ventilator asynchrony and trends in patient-ventilator asynchrony. The present disclosure describes a framework that uses pressure, flow, and volume waveforms to detect patient-ventilator asynchrony and the presence of secretions in the ventilator circuit.
1. A method for indicating patient-ventilator interaction, comprising:
a) receiving ventilator pressure and flow waveform data relating to a ventilator;
b) applying one or more normalizing constants to said ventilator pressure and flow waveform data to obtain respective normalized ventilator pressure and flow waveform data wherein the normalizing constant for the ventilator pressure waveform is the inverse of a value of pressure at approximately a time of peak flow;
c) calculating a Delta waveform based upon pairwise relationships between said respective normalized ventilator pressure and flow waveform data; and
d) using a presence or absence of peaks and/or valleys of the Delta waveform as an indication of a presence or absence of one or more asynchronous patient-ventilator interactions.
2. The method of claim 1 , wherein the ventilator pressure and flow waveform data are obtained directly or indirectly from the ventilator, directly or indirectly from another monitor or system for acquiring ventilator pressure and flow waveform data, or an archive of in vivo, ex vivo, in vitro and/or in silico data.
3. The method of claim 1 , further comprising providing a prompt for managing the present asynchronous patient-ventilator interactions.
4. The method of claim 1 , further comprising
a) extracting a feature vector from the Delta waveform; and
b) determining and indicating one or more types of asynchronous patient-ventilator interactions present, or absence thereof, associated with one or more breath cycles, which determination of asynchronous patient-ventilator interactions or absence thereof are based on the extracted feature vector.
5. The method of claim 4 , wherein extracting a set of features forming the feature vector comprises one or more of the following:
a) depth of valleys of the Delta waveform;
b) maximum value of the Delta waveform within inspiration phase;
c) area under the curve of the Delta waveform within inspiration phase;
d) maximum cross-correlation coefficient of the Delta waveform with a delivered tidal volume waveform;
e) maximum cross-correlation coefficient of the Delta waveform with the ventilator flow waveform.
6. The method of claim 4 , wherein the type of asynchronous patient-ventilator interactions present, or absence thereof, is determined and indicated by using a rule based algorithm and/or machine-learning algorithm configured to classify one or more of the breath cycles into one or more categories of asynchronous patient-ventilator interactions.
7. The method of claim 6 , wherein the rule-based algorithm and/or machine-learning algorithm classifies one or more of the breath cycles into one or more categories of asynchronous patient-ventilator interactions by comparing at least one value from the set of features with a predetermined threshold.
8. The method of claim 4 , wherein the types of asynchronous patient-ventilator interactions are chosen from one or more of inadequate ventilator support, double triggering, ineffective triggering, premature termination, delayed termination, flow starvation, air trapping, buildup of fluid in lungs and/or a ventilator circuit, and/or no asynchrony.
9. A method for indicating patient-ventilator interaction, comprising:
a) receiving ventilator pressure and flow waveform data relating to a ventilator;
b) applying one or more normalizing constants to said ventilator pressure and flow waveform data to obtain respective normalized ventilator pressure and flow waveform data wherein the normalizing constant for the ventilator flow waveform is the inverse of a value of flow at approximately a time of peak flow;
c) calculating a Delta waveform based upon pairwise relationships between said respective normalized ventilator pressure and flow waveform data; and
d) using a presence or absence of peaks and/or valleys of the Delta waveform as an indication of a presence or absence of one or more asynchronous patient-ventilator interactions.
10. The method of claim 9 , wherein the ventilator pressure and flow waveform data are obtained directly or indirectly from the ventilator, directly or indirectly from another monitor or system for acquiring ventilator pressure and flow waveform data, or an archive of in vivo, ex vivo, in vitro and/or in silico data.
11. The method of claim 9 , further comprising providing a prompt for managing the present asynchronous patient-ventilator interactions.
12. The method of claim 9 , further comprising
a) extracting a feature vector from the Delta waveform; and
b) determining and indicating one or more types of asynchronous patient-ventilator interactions present, or absence thereof, associated with one or more breath cycles, which determination of asynchronous patient-ventilator interactions or absence thereof are based on the extracted feature vector.
13. The method of claim 12 , wherein extracting a set of features forming the feature vector comprises one or more of the following:
a) depth of valleys of the Delta waveform;
b) maximum value of the Delta waveform within inspiration phase;
c) area under the curve of the Delta waveform within inspiration phase;
d) maximum cross-correlation coefficient of the Delta waveform with a delivered tidal volume waveform;
e) maximum cross-correlation coefficient of the Delta waveform with the ventilator flow waveform.
14. The method of claim 12 , wherein the type of asynchronous patient-ventilator interactions present, or absence thereof, is determined and indicated by using a rule based algorithm and/or machine-learning algorithm configured to classify one or more of the breath cycles into one or more categories of asynchronous patient-ventilator interactions.
15. The method of claim 14 , wherein the rule-based algorithm and/or machine-learning algorithm classifies one or more of the breath cycles into one or more categories of asynchronous patient-ventilator interactions by comparing at least one value from the set of features with a predetermined threshold.
16. A method for indicating patient-ventilator interaction, comprising:
a) receiving ventilator pressure and flow waveform data relating to a ventilator;
b) applying one or more normalizing constants to said ventilator pressure and flow waveform data to obtain respective normalized ventilator pressure and flow waveform data wherein the normalizing constant for the ventilator pressure waveform is the inverse of a value of pressure at an elapsed interval into a breath cycle that is approximately a rise time as set on the ventilator and/or the normalizing constant for the ventilator flow waveform is the inverse of a value of flow at the elapsed interval into the breath cycle that is approximately the rise time as set on the ventilator;
c) calculating a Delta waveform based upon pairwise relationships between said respective normalized ventilator pressure and flow waveform data; and
d) using a presence or absence of peaks and/or valleys of the Delta waveform as an indication of a presence or absence of one or more asynchronous patient-ventilator interactions.
17. The method of claim 16 , wherein the ventilator pressure and flow waveform data are obtained directly or indirectly from the ventilator, directly or indirectly from another monitor or system for acquiring ventilator pressure and flow waveform data, or an archive of in vivo, ex vivo, in vitro and/or in silico data.
18. The method of claim 16 , further comprising
a) extracting a feature vector from the Delta waveform; and
b) determining and indicating one or more types of asynchronous patient-ventilator interactions present, or absence thereof, associated with one or more breath cycles, which determination of asynchronous patient-ventilator interactions or absence thereof are based on the extracted feature vector.
19. The method of claim 18 , wherein the type of asynchronous patient-ventilator interactions present, or absence thereof, is determined and indicated by using a rule based algorithm and/or machine-learning algorithm configured to classify one or more of the breath cycles into one or more categories of asynchronous patient-ventilator interactions.
20. The method of claim 19 , wherein the rule-based algorithm and/or machine-learning algorithm classifies one or more of the breath cycles into one or more categories of asynchronous patient-ventilator interactions by comparing at least one value from the set of features with a predetermined threshold.