Discordance monitoring
Described herein are systems, devices, and methods for cardiac monitoring. In particular, the systems, devices, and methods described herein may be used to conveniently sense the presence of an intermittent arrhythmia in an individual. The systems, devices, and methods described herein may be further configured to sense an electrocardiogram.
1. A method of cardiac monitoring, comprising:
continuously sensing an activity level of a user with a first sensor of a wearable device comprising a wristlet or a smartwatch worn by the user;
continuously sensing a set of heart rate parameters of the user with a second sensor of the wearable device;
in response to sensing a resting activity level, comparing, using a machine learning (ML) algorithm, the resting activity level with the set of heart rate parameters to determine whether a discordance is present between the resting activity level and the set of heart rate parameters sensed when the resting activity level is sensed, wherein the set of heart rate parameters comprise an indication of a heart rate variability and an indication of a heart rate value;
in response to determining that the discordance is present:
indicating to the user, using the wearable device, a possibility of an arrhythmia being present;
initiating an ECG of the user using ECG electrodes of the smartwatch;
receiving electric signals of the user from the ECG electrodes of the smartwatch; and
displaying an ECG rhythm strip from the received electric signals; and
improving an accuracy with which the ML algorithm to identifies an arrhythmia by:
storing the resting activity level and the set of heart rate parameters as part of a historical set of parameter values, the historical set of parameter values including previously recorded resting activity levels and previously recorded heart rate parameters;
storing the ECG rhythm strip as part of a historical set of ECG data comprising previously recorded ECG rhythm strips; and
training the ML algorithm by comparing the historical set of parameter values with the historical set of ECG data.
2. The method according to claim 1 , wherein the arrhythmia is atrial fibrillation.
3. The method according to claim 1 , wherein the arrhythmia is atrial fibrillation.
4. The method according to claim 1 , wherein indicating to the user further comprises: instructing the user to record an electrocardiogram (“ECG”) using the wearable device.
5. The method according to claim 1 , wherein the arrhythmia is selected from a group consisting of atrial fibrillation, supraventricular tachycardia, and ventricular tachycardia.
6. The method according to claim 1 , wherein the first sensor comprises an accelerometer.
7. The method according to claim 1 , wherein the first sensor comprises a gyroscope.
8. A smartwatch, comprising:
a processor;
a first sensor configured to continuously sense an activity level value of a user, wherein the first sensor is coupled to the processor;
a photopletysmogram (PPG) sensor configured to continuously sense a set of heart rate parameters of the user, wherein the PPG sensor is coupled to the processor; and
a non-transitory computer readable storage medium encoded with a computer program including instructions executable by the processor to cause the processor to:
in response to sensing a resting activity level, compare, using a machine learning (ML) algorithm, the resting activity level with the set of heart rate parameters to determine whether a discordance is present between the resting activity level of the user and the set of heart rate parameters sensed when the resting activity level is sensed, wherein the set of heart rate parameters comprise an indication of a heart rate variability and an indication of a heart rate value;
in response to determining that the discordance is present:
indicate to the user a possibility of an arrhythmia being present;
initiate an ECG of the user using ECG electrodes of the smartwatch;
receive electric signals of the user from the ECG electrodes of the smartwatch if the possibility of an arrhythmia is indicated; and
display an ECG rhythm strip from the received electric signals; and
improving an accuracy with which the ML algorithm identifies an arrhythmia by:
storing the resting activity level and the set of heart rate parameters as part of a historical set of parameter values, the historical set of parameter values including previously recorded resting activity levels and previously recorded heart rate parameters;
storing the ECG rhythm strip as part of a historical set of ECG data comprising previously recorded ECG rhythm strips; and
training the ML algorithm by comparing the historical set of parameter values with the historical set of ECG data.
9. The smartwatch according to claim 8 , wherein the arrhythmia is atrial fibrillation.
10. The smartwatch according to claim 8 , wherein the arrhythmia is atrial fibrillation.
11. The smartwatch according to claim 8 , wherein indicating to the user further comprises: instructing the user to record an electrocardiogram (“ECG”) using the wearable device.
12. The smartwatch according to claim 8 , wherein the arrhythmia is selected from a group consisting of atrial fibrillation, supraventricular tachycardia, and ventricular tachycardia.
13. The smartwatch according to claim 8 , wherein the first sensor comprises an accelerometer.
14. The smartwatch according to claim 8 , wherein the first sensor comprises a gyroscope.