System and method for validating cardiovascular parameter monitors
A system or method for validating an unvalidated cardiovascular parameter monitor can include receiving a plurality of reference cardiovascular parameter datasets, each reference cardiovascular parameter dataset associated with a patient of a plurality of patients; receiving a plurality of test cardiovascular parameter measurements, each test cardiovascular parameter measurement includes a cardiovascular parameter measured for a patient of the plurality of patients; and validating the cardiovascular parameter monitor based on an analysis of the plurality of reference cardiovascular measurements and the plurality of test cardiovascular parameter measurement.
1 . A method for determining a validation status for a blood pressure monitor comprising:
for a plurality of patients:
(a) at a first time, measuring a validated blood pressure for each patient of the plurality by:
(i) inflating and deflating a blood pressure cuff disposed on a first arm of the patient;
(ii) contemporaneously with inflating and deflating the blood pressure cuff, measuring an audio signal at a frequency of at least 10 kHz using a digital stethoscope; and
(iii) contemporaneously with inflating and deflating the blood pressure cuff, measuring a pressure dataset using a digital pressure gauge, wherein the pressure dataset comprises an instantaneous pressure of the blood pressure cuff measured at a frequency of at least 10 Hz;
(b) contemporaneously with measuring the validated blood pressure, (b) measuring an unvalidated blood pressure using the blood pressure monitor for each patient comprising:
(i) recording a time series of images of a finger region of the respective patient on a second arm of the respective patient;
(ii) determining a pulse waveform from the time series of images; and
(iii) extracting a set of fiducials from the pulse waveform, wherein the set of fiducials comprise fit parameters associated with a set of fitting functions used to fit the pulse waveform; and
(iv) transforming the set of fiducials into the unvalidated blood pressure;
(c) at a second time, comparing the validated blood pressure to the unvalidated blood pressure; and
(d) determining the validation status of the blood pressure monitor based on the comparison;
wherein the validated blood pressure comprises a systolic blood pressure approximately equal to a pressure of the pressure dataset measured when a first Korotkoff sound is detected in the audio signal and a diastolic blood pressure approximately equal to a pressure of the pressure dataset measured when a fourth or fifth Korotkoff sound is detected in the audio signal.
2 . The method of claim 1 , wherein the first time and the second time differ by at least one hour.
3 . The method of claim 1 , wherein the systolic blood pressure and the diastolic blood pressure are determined in a post-hoc analysis of the audio signal and the pressure dataset, via a computing system.
4 . The method of claim 1 , wherein comparing the validated blood pressure to the unvalidated blood pressure comprises performing a mixed model repeated analysis, via computing system, wherein the validation status is positive when a mean error of the unvalidated blood pressure is about ±5 mmHg or a standard deviation of at most about 8 mmHg.
5 . The method of claim 1 , further comprising, for each patient of the plurality of patients, contemporaneously with inflating and deflating the blood pressure cuff, measuring an ECG dataset.
6 . A method for determining a validation status for a photoplethysmography-based blood pressure monitor comprising:
(a) at a first time, measuring a validated blood pressure for a patient by:
(i) inflating and deflating a blood pressure cuff disposed on a first arm of the patient,
(ii) contemporaneously with inflating and deflating the blood pressure cuff, measuring an audio signal at a frequency of at least 10 kHz using a digital stethoscope, and
(iii) contemporaneously with inflating and deflating the blood pressure cuff, measuring a pressure dataset using a digital pressure gauge, wherein the pressure dataset comprises an instantaneous pressure of the blood pressure cuff measured at a frequency of at least 10 Hz;
(b) contemporaneously with measuring the validated blood pressure, measuring an unvalidated blood pressure using a photoplethysmography-based blood pressure monitor for the patient comprising:
(i) recording a time series of images of a finger region of a second arm of the patient,
(ii) determining a pulse waveform from the time series of images,
(iii) extracting a set of fiducials from the pulse waveform, wherein the set of fiducials comprise fit parameters associated with a set of fitting functions used to fit the pulse waveform, and
(iv) transforming the set of fiducials into the unvalidated blood pressure;
(c) at a second time, comparing the validated blood pressure to the unvalidated blood pressure;
(d) determining the validation status of the photoplethysmography-based blood pressure monitor based on the comparison; and
(e) outputting the validation status, wherein the validation status enables or disables the photoplethysmography-based blood pressure monitor for clinical use,
wherein the validated blood pressure comprises a systolic blood pressure approximately equal to a pressure of the pressure dataset measured when a first Korotkoff sound is detected in the audio signal and a diastolic blood pressure approximately equal to a pressure of the pressure dataset measured when a fourth or fifth Korotkoff sound is detected in the audio signal.
7 . A method for determining a validation status for a blood pressure monitor comprising:
(a) at a first time, measuring a validated blood pressure for a patient by:
(i) inflating and deflating a blood pressure cuff disposed on a first arm of the patient,
(ii) contemporaneously with inflating and deflating the blood pressure cuff, measuring an audio signal at a frequency of at least 10 kHz using a digital stethoscope, and
(iii) contemporaneously with inflating and deflating the blood pressure cuff, measuring a pressure dataset using a digital pressure gauge, wherein the pressure dataset comprises an instantaneous pressure of the blood pressure cuff measured at a frequency of at least 10 Hz;
(b) contemporaneously with measuring the validated blood pressure, measuring an unvalidated blood pressure using the blood pressure monitor for the patient comprising:
(i) recording a time series of images of a finger region of a second arm of the patient,
(ii) determining a pulse waveform from the time series of images,
(iii) extracting a set of fiducials from the pulse waveform, wherein the set of fiducials comprise fit parameters associated with a set of fitting functions used to fit the pulse waveform, and
(iv) transforming the set of fiducials into the unvalidated blood pressure;
(c) using a computing system, automatically processing the validated blood pressure and the unvalidated blood pressure by comparing the validated blood pressure to the unvalidated blood pressure; and
(d) using the computing system, automatically determining the validation status of the blood pressure monitor based on the comparison,
wherein the validated blood pressure comprises a systolic blood pressure approximately equal to a pressure of the pressure dataset measured when a first Korotkoff sound is detected in the audio signal and a diastolic blood pressure approximately equal to a pressure of the pressure dataset measured when a fourth or fifth Korotkoff sound is detected in the audio signal.