Wearable photoplethysmography device for detecting clinical decompensation based on heart rate variability
Disclosed are clinical decompensation monitoring platforms and methods. An example platform includes a wearable device to acquire photoplethysmography (PPG) waveform data representing a person's heart rate variability (HRV). Also disclosed are methods for collecting and verifying the HRV data, scoring and selecting optimal segments of HRV data for further processing that entails detecting clinically significant changes in a high frequency (HF) component, and passively monitoring clinical decompensation for providing patient feedback with individualized coping mechanisms.
1 . A method of processing photoplethysmography (PPG) waveform data from a wearable PPG device to detect clinical decompensation in a person wearing the wearable PPG device, the method comprising:
periodically acquiring the PPG waveform data;
splitting the PPG waveform data into segments;
scoring each segment based on its temporal proximity to a desired time interval for acquiring the segment;
processing PPG waveform data based on one of the segments having a best score based on a number of errors it has and its temporal proximity;
performing beat detection on the PPG waveform data;
verifying detected beats by checking motion data measured by the wearable PPG device and comparing the detected beats to a corresponding reference signal configured to determine common fiducials representing a repeating cardiogenic artifact;
in response to the verifying, calculating a heart rate variability (HRV) metric from the detected beats;
comparing the HRV metric to a clinical decompensation threshold so as to detect the clinical decompensation in the person; and
in response to detecting the clinical decompensation, providing a notification.
2 . The method of claim 1 , in which the HRV metric includes at least one of: HF component, HF/LF ratio, RMSSD, and standard deviation in width of a Poincaré plot.
3 . The method of claim 1 , further comprising:
determining after a predetermined period whether the HRV metric indicates an improvement to the clinical decompensation; and
alerting a third party in response to insufficient improvement.
4 . The method of claim 1 , in which the providing the notification includes indicating an intervention.
5 . The method of claim 1 , in which the providing the notification includes generating a message to cause the wearable PPG device to notify the person or a caregiver of a risk of suicidality.
6 . The method of claim 1 , in which the wearable PPG device has a smartwatch form factor.
7 . The method of claim 1 , further comprising providing the PPG waveform data to a remotely located server configured to process the PPG waveform data.
8 . The method of claim 1 , further comprising scoring each segment based on beat picking errors.
9 . The method of claim 1 , further comprising:
detecting a sequence of beat intervals in the PPG waveform data; and
removing errors in the sequence.
10 . The method of claim 1 , further comprising temporarily halting the periodically acquiring when the motion data indicates the person was moving.
11 . The method of claim 1 , further comprising filtering the HRV metric to remove a portion acquired while the motion data indicates the person was moving.
12 . The method of claim 1 , in which the clinical decompensation indicates a decline in mental health.
13 . The method of claim 1 , in which the intervention comprises an activity that does not require participation of a clinician.
14 . The method of claim 13 , in which the intervention comprises one of therapeutic biofeedback or a personalized solution for the user.