No longer able to retroactively choose Sensitive Sleep

[bluefoxicy]

Sleep actigraphy is roughly 97% similar to reference polysomnography for tracking total sleep time if it is less-sensitive to movement when sleep is efficient, and more-sensitive to movement when sleep is inefficient.  In a practical sense, this means you can only get an accurate view of sleep time by setting the Fitbit tracker to "Normal" or "Sensitive" after waking up, depending on how you slept.

I no longer see the option to change the sleep tracker used in the sleep graphs.  As I can't predict if I'm going to have a bad night, I can't reconfigure the Fitbit to use the appropriate tracker prior to sleeping.

Moderator Edit: Clarified Subject

[SunsetRunner]

@bluefoxicy - To get to the sleep sensitive settings, where you can change from Normal to Sensitive, go to Account --> Advanced Settings --> Sleep Sensitivity. Hope this helps.

[bluefoxicy]

You have to change the sleep sensitivity for individual days, after those days have been recorded.

"To change the sleep sensitivity setting for a given date, select the sleep data for that date, and set its sensitivity."

July 17, 2016:  Sleep sensitivity "Normal".

July 18, 2016:  Sleep sensitivity "Normal".

July 19, 2016:  Sleep sensitivity "Normal".

July 20, 2016:  Restless night, bad sleep.  Sleep tracker says 85% efficiency, 6 hours asleep; I spent 2 hours asleep and 5.5 hours awake!  Retroactively change sleep sensitivity to "Sensitive".  Now sleep tracker says 2 hours 11 minutes asleep.

July 21, 2016:  Slept well.  Sleep sensitivity "Normal".

...

Here's how it works:

Fitbit "Normal" sensitivity tracker:  This tracker will be accurate on nights when you sleep well.

Fitbit "Sensitive" sensitivity tracker:  This tracker will be accurate on nights when you sleep poorly.

Swap these settings and you will get wildly-inaccurate sleep measurements.  The only way to get a correct measurement is to assign the tracker based on how well you slept that night.

Moderator edit: format

[SunsetRunner]

@bluefoxicy - I did not mean to insult you. And I did indeed read your entire posting. It is evident that I am not the right person to try to help with this question. Hopefully someone else can. Or you can contact Support. 

[MatthewFitbit]

---

Here's how it works:

Fitbit "Normal" sensitivity tracker:  This tracker will be accurate on nights when you sleep well.

Fitbit "Sensitive" sensitivity tracker:  This tracker will be accurate on nights when you sleep poorly.

---

Hi @bluefoxicy - it sounds like there might be a misunderstanding about this feature, and I'll be happy to clear it up.

We recommend that all Fitbit customers use the "Normal" sleep setting unless they are specifically tracking sleep for medical purposes or gathering data about a potential sleep disorder. The "Normal" setting will always provide the most accurate information, while "Sensitive" will show you every single subtle movement that you made during the night, resulting in an extreme vantage of the sleep data. They are not intended for the purpose suggested above, and we do not advise switching this setting retroactively. Unless you are working with a physician to study your sleep, you can safely leave it set to "Normal" all the time.

You can read more in this help article. I hope that helps!

[squirrelyone]

Matthew, as a lot of the people objecting to the changes to the sleep tracker have been pointing out all day, many of us *are* seeing physicians for sleep disorders.  This is a leading reason many of us bought Fitbits.  Bluefoxicy began his post comparing actigraphy to polysomnagraphy, for pity's sake.  Maybe it's not a feature Fitbit intended when it was implemented, but it's a feature that was there, that users saw was there, and that many of us went out, purchased, and recommended to our friends because of it.  We bring our reports to our neurologists.  We bring them to our full medical sleep studies.  They're not substitutes for those studies, but they're features we bought our devices with every reasonable expectation of having for the foreseeable future.  Surely you can understand why we're frustrated now.

[bluefoxicy]

Yeah, that's not how that setting works.

When you switch from "Normal" to "Sensitive", Fitbit actually claims less total sleep time.  It registers periods of lesser movement as "Awake" than it does when in "Normal" mode.  It's not just a more-detailed graph; it fundamentally analyzes the data in a different manner.

I know this because I've done it repeatedly, in both directions, and compared the results to nightly sleep journaling.  I've also read research papers on sleep actigraphy which explain that this is standard medical procedure.

@MatthewFitbit, you are giving technically-incorrect advice.  The link you cited also says you're wrong:

• The normal setting counts significant movements as being awake (such as rolling over) and is appropriate for most users.
• The sensitive setting will cause your tracker to record nearly all movements as time spent restless or awake. This setting may be helpful for users who wake up feeling tired even though their sleep history shows sufficient rest.

The sensitive setting actually reduces total sleep time given the same sleep data.  It reduces it significantly, frequently by half or more in my case.

[MatthewFitbit]

@bluefoxicy I think we're saying the same thing in different ways. The Sensitive option registers even tiny movements as restless time, with the intention of giving a more high-contrast view of your night's activity. This has the result of showing more time as restless and less time as asleep.

As I mentioned above, this isn't the most valuable vantage for the majority of Fitbit users, and we recommend just using the Normal setting to see proper sleep data. You shouldn't be switching back and forth or changing it retroactively if you want to see accurate information.

[bluefoxicy]

@squirrelyone I wasn't just comparing it to polysomnography.  Sleep actigraphy is a well-studied practice, and has been compared to polysomnography in many papers.

It's well-known that sleep actigraphy is generally good at detecting sleep (above 87% accurate) but poor at detecting wakefulness (barely over 30%).  Some of the literature actually does observe that a large amount of long wakeful episodes in which the patient doesn't attempt to leave bed will register as (restless) sleep, while using an algorithm detecting wakefulness at a lower threshold will register wakefulness episodes when the patient was asleep.

For insomniacs, it's been shown that a lower threshold reads nights of poor sleep with relatively-high accuracy, and a higher threshold reads nights of good sleep with relatively-high accuracy.  This is accurate for non-morbid cases, as well; however, cases involving apnea or restless limb syndrome have different concerns.

Fitbit isn't a precision medical tool; it's of sufficiently-high quality to provide trend data, which allows individuals to better-understand what's happening to them and medical professionals to analyze patient testimony and other diagnostic criteria in a better context.  EEG and polysomnography would be far-better data, and properly-handled Fitbit data remains of high enough quality to have proliferated its use in sleep disorder research.

I've developed protocol to get the best quality data available from the Fitbit; that protocol is now impossible.

[bluefoxicy]

@MatthewFitbit Let me put this simply:  you're wrong.  You are also insulting me with your ignorance.  I have actually used the Fitbit, I've compared the measurements in both modes to actual sleep journaling, and I've compared it to scientific literature on the topic of sleep actigraphy.  I've gotten consistent results.  That means I have put effort into undestanding how this thing works, and you are telling me I'm stupid and delusional.
 
The "Sensitive" option is more accurate in the case of more-restless sleep.  If you didn't predict having sleeping difficulties that night, slept 3 hours, and stayed in bed for 8, you'll only get something substantially-close to 3 hours by changing the tracker used to "Sensitive".  Otherwise you get something that says you slept for 6 or 7 hours.
 
I've been journaling my time each night for nearly half a year, comparing the general amount of resltlessness on bad nights to what's in the Fitbit, and evaluating the different settings for their accuracy both in total time asleep and in how they register spans of time when I know I was asleep or awake.  I've read papers on how sleep actigraphy works.  I've put a lot of research into this, and I've gotten consistent results.
 
You're over there saying people should predict the future and make decisions to outcomes not in their control before-hand.
 
Changing the sensitivity option to match the type of sleep I've had has consistently given me the most-accurate information I've derived from Fitbit.
 
If I have a night where I spend most of my time lying awake in the middle of the week, Fitbit might drop me from a 90% sleep efficiency to an 86% sleep efficiency, even though I actually spent 5 additional hours awake that night.  A look at the entire week shows sleep of consistently-high quality every single day.
 
If I set the tracker to "Sensitive" for every day that week, Fitbit shows me sleeping at around 36%-47% efficiency every single day, even though I spend over 90% of my time in bed asleep.  A look at the entire week shows sleep of consistently-low quality every single day.
 
If I switch the tracker to "Sensitive" for the bad day, Fitbit shows me as awake during the times I'm actually awake (because I check the time when I WASO into high alertness, and when I can't get back to sleep for an extended period), and actually records me as sleeping less because I'm awake during those times.  That's more accurate.  Meanwhile, I leave the days where it accurately and correctly recorded me as sleeping most of the night set to "Normal", because setting it to "Sensitive" those days will record several hours of time spent asleep as time spent awake.
 
You can keep saying that it doesn't work that way, and you can also go jump off the roof of the Hilton and keep telling everyone you can fly the whole way down.  You'd  be just as correct either way.

Moderator edit: Format

[squirrelyone]

I apologize for oversimplifying!  My points were only that:

A) it should have been very clear that your initial post was not spitballing in the wind, you're obviously very well read on this matter and shouldn't be blown off with the general customer service bland response and

B) these report *are* helpful to those of us with sleep disorders.  They aren't replacements for sleep studies.  But they're ways we can track things at home and keep a general thumb on our situation, so we can bring a graph with general trends to our doctors--"Since I've been on this new medication, I sleep for an hour and a half straight at the beginning, then in nothing more than 10 minute increments for the rest of the night"--and let them decide if that, along with our other tools, our sleep journals, observations from our partners, etc., are causes for other studies.  As you said - far better than I could - the trend data is helpful.

A lot of words over here to say I agree with you, I've agreed with all your posts, and you are my hero of the day.

[WavyDavey]

Just in case it wasn't stated, there is a board called Feature Suggestions here where you can request new features. The community can vote on it, and with enough votes it may be implemented. There's a difference between "I need help on a feature" and "I want it to do this thing it doesn't do (or used to do)".

It's an interesting discussion, but if you want the developers to hear you, posting in the Feature Suggestion board is the way to do it. If we're just batting the ball around, great. But the developers don't monitor this forum, so we're basically talking it out. Carry on.

[MatthewFitbit]

@bluefoxicy You are of course welcome to believe what you choose, but I want to ensure that other users don't get confused with misinformation. My post above explains the feature accurately.

@WavyDavey Thanks for that thought - normally I would 100% agree with you that the Feature Suggestion board is the place for such discussion. However, in this circumstance it's probably more helpful to post your thoughts here so that I can relay them directly to the relevant team. No need to create a Feature Suggestion for this.

I'm going to go ahead and close this thread, as the question has been answered and I don't see anything else worthwhile happening here.