AFIB burden, measured by Apple Watch Series IX

In addition to calculating the Heart Rate Variability (Apple uses the 1 standard deviation of a series of consecutively measured R-R intervals) , the watch also calculates a weekly AFIB burden (time spent in AFIB over the previous 7 days).

Presumably the input for the AFIB burden is some level of heart rate variability, above which the watch considers the wearer to be in AFIB.

Does anyone know what that level - or threshold - is?

(for example... Is it above 35 ms or above 50 ms or above 75ms? . Apple engineers are unwilling to divulge, citing proprietary algorithm. So the only way to answer is question is with empirical data, that is, by users experience

I'm pretty sure the algorithms use more than just HRV. If they relied solely on that, things like bigeminal PACs would get flagged as afib and they usually aren't, at least not with a Kardia (mine never were). It's also pretty clear that Kardia and Apple Watch use different algorithms because AW is limited to identifying afib within certain heart rate limits (50-150 bpm currently, used to be 50-120 bpm) but Kardia has no limits.

Just to be clear, I'm not referring to the interpretation of AFIB from an ECG on the apple watch. Rather (it would seem) that the AFIB burden is calculated independently of the ECG.
It would appear that the Heart rate variability is the major (and perhaps only) input used to formulate an AFIB history. Hence my question, when is a HRV observation flagged as AFIB (for purposes of aggregating weeks-long data to produce an AFIB burden)

You may have to wear a Zio patch for 5 days to accurately compute your burden percentage. That’s how in 2019 my EP calculated my pvc burden.

If you have a PM, you could get an interrogation upload because it’s more accurate

To Carey's point, the HR vs time examples I posted here, the first is NSR with lots of PACs, the second is afib. Both would have had very high HRV. I would also be suspect of the accuracy using plethysmography vs ECG data for this. I believe when you have to touch the watch with your other hand is ECG data, otherwise the watch is using plethysmography (blood flow pulses) for the afib data.

Hi Susan, we have used ZIO patch 3 times (twice for 14 days and once for 30 days) while simultanously monitoring with the apple watch. In all three cases the ZIO calculated burden matched the Apple within +/- 5 points.
The EP was satisfied that we could monitor the burden (on a weekly basis) using the apple watch - however if he were to recommend a procedure that was covered by insurance he would have to have ZIO data to comply with the insurance criterion. OTOH, he was satisfied that he could use the Apple calculated burden to guide him in modifying prescriptions.

Perhaps we were fortunate that the RVR never exceeded ~ 140 BPM or so. I agree with Cary that the reliability of R-R intervals degrades at very high ventricular rates.