Sleep score is not accurate

[pwung0327]

One night I got almost 7 hours of sleep and I get a score of 72 and the other night I barely get 5 hours of sleep and I get a score of 76?

Also, do you have any literature which justifies Fitbit to apply the HRV rather than the brain monitoring to estimate the sleep stages?

Moderator edit: subject for clarity 

[WilsonFitbit]

Hello @pwung0327, it's great to see you around! 

Your overall sleep score is a sum of your individual scores in "sleep duration, sleep quality, and restoration", for a total score of up to 100. Most people get a score between 72 and 83. Sleep score ranges are:

• Excellent: 90-100
• Good: 80-89
• Fair: 60-79
• Poor: Less than 60

Duration: Time asleep and awake: the more you sleep, the better your score. But take into consideration that the two other scores below, count. 

Quality: Deep & REM sleep: how much time you spent in deep and REM sleep, so the more time you spend in these sleep stages, the better your score. 

Restoration: Sleeping heart rate & restlessness: How relaxed you were during sleep, a high sleeping heart rate or too much tossing and turning lowers your score. For more information, please check: What's sleep score in the Fitbit app?

For your inquiry about HRV, please check How do I track heart rate with my Fitbit device? and for sleep stages, please see: What should I know about Fitbit sleep stages?

Sleep stages are traditionally measured in a lab using an electroencephalogram to detect brain activity along with other systems to monitor eye and muscle activity. While this method is the gold standard for measuring sleep stages (source), your device can estimate your sleep stages in a more comfortable, convenient way.

Hope this helps!

[pwung0327]

Yeah, I can read the FAQ as well as anyone.

My question is: does the sleep score derive from the percentage of time spent in say REM sleep as compared to the total time spent sleeping? If that is the case, the sleep score discounts the time in light sleep. So it doesn't mean anything.

 Second, I want to know whether using the HRV is equivalent to using the "electroencephalogram to detect brain activity along with other systems to monitor eye and muscle activity"? What is the correlation between the two methods, if any.  I also want to know just how far off the HRV method is in determining the stages of sleep.

Thanks

[WilsonFitbit]

@pwung0327, thank you for the information you've shared! 

Your understanding is very appreciated but we're unable to answer your questions because it would involve the release of proprietary information. Currently, we are only able to share information that is available on our help articles. 

See you around. 

[t.parker]

@pwung0327 you are right. Sleep score probably means nothing. Why? Because it cannot be verified as it is a "proprietary" metric (how convenient...). If something cannot be verified then it could be just random number given and it would mean exactly the same. Sleep score is just another made-up metric not backed by any science.

Your second question is very interesting and there is a lot of research done discussing such correlation (it is really interesting topic). As for HRV, it requires having accurate sensor. Wrist-based optical HR sensors (any, not just Fitbit) very unlikely are able to provide accuracy.

[WilsonFitbit]

@t.parker, thank you very much for taking the time to provide your feedback! 

Currently, to see additional insights into your sleep data, including analysis of sleeping heart rate, a Premium subscription is needed. For more information, see What should I know about Fitbit Premium? Here you can download your complete Fitbit data at any time. Note that the data is raw data without further analysis. For more information, see How do I export my Fitbit account data?

On a side note, please note that our team is always working to enhance your experience and your feedback is a big part of that process. Definitely it helps a lot! 

See you around. 

[pwung0327]

As someone older, I am having to get up at night to go to the restroom. That is part of my life. The sleep algorithm built into the sleep function trashes the sleep that I had before getting up to do my business. The sleep data that the app is showing underestimates the amount of sleep that I am actually getting. If I were a conspiracy theorist I would say that Fitbit is deliberately giving every customer the idea that they are getting much less sleep than they actually are getting. 

The explanations for the sleep analysis does not dig into the facts of the matter, things like how you correlate the heart beat data with the brain wave activity is nonexistent. That is the basis of your justification of doing this kind of analysis using a substtute measure, and yet the explanation is ambiguous and the justification is: You just have to trust us. 

Really not science and very duplicitous.

[WilsonFitbit]

@pwung0327, thank you very much for your reply! 

I'm sorry for any inconvenience caused, your comments haven't gone unnoticed and be sure they are very important to us. I'd like to let you know that the data provided is for educational purposes only and is not intended as a substitute for medical diagnosis or treatment. The recommendation is not use this information to diagnose or treat a health problem or condition. If you have any concerns about your health or before changing your diet, altering your sleep habits, taking supplements, or starting a new fitness routine, always check with your doctor. 

Thanks for your understanding. 

[hawkmeister]

While I'd agree that it's not extremely accurate,,a sensor on the wrist can't generate the rigorous scientific detail of a sleep study.   IMO it's more of a guideline.   

I think you can get a relative idea of your sleep on a given night compared to other nights.

[pwung0327]

I totally agree with you, but I dislike Fitbit trying to tell me that their analytics are just as accurate as those generated in a sleep lab. I also take issue with the fact that the reports ignore chunks of sleep data because I had to get up to go to the bathroom and I am having to log in those times because the sensor stop recording. This isn't why I paid to get the data.

[WilsonFitbit]

Hello @hawkmeister, welcome to the community forums! Thank you for your input! 

@pwung0327, currently, it is expected that when you wake up or get up at night your Fitbit device will stop recording your Sleep Activity, therefore; your Sleep stages. When you go back to Sleep, then it might start a 2nd Sleep log but take into consideration that this happens when your body is completely at rest and you haven’t moved for about an hour, your Fitbit device records that you’re asleep.

As a workaround, on the Mobile Fitbit App, you can begin a sleep log by going into your Sleep tile and clicking on the plus (+) sign. Choose “Begin Sleep Now” and then whenever you wake up, just click on the “I’m Awake” button to stop logging. Therefore; this options prevents Sleep Stages and Sleep Score to be received. 

Thank you very much for your feedback! 

See you around. 

[pwung0327]

I am doing that now, logging the missing sleep. My question to you is: why do I need to do that? The system can retieve my sleep data from the Fitbit during that time when I have not rested for an hour continuously, why cut that out of the reporting and making me, the customer to retrieve it by hand?

[WilsonFitbit]

@pwung0327, the situation is that your Fitbit device automatically detects your movements. For this reason, when you wake up or get up at night, your Fitbit device will finish the Sleep tracking and it will start recording a second log after it detects that there hasn't been any movement. This is expected. 

I'll be around. 

[Joe.FIT]

My biggest problem with sleep (and Fitbit tracking my sleep) is I wake up in the middle of the night, usually around 3:00 AM.  I've taken to working a puzzle such as SODUKO until my eyes get heavy.  I then roll over and fall asleep again.

Fitbit records this as two separate sleep cycles.  Sometimes the second sleep cycle is short (about 2 hours) and I get no sleep score for it.  Sometimes the cycle is long enough that I get two sleep scores for the night, but both will typically be somewhat low numbers such as 60.

Is there anyway to resolve the sleep score average for the night?

  I tend to ignore the sleep score due to this and work with just the total number of hours of sleep.  I use the formula of total number of hours asleep divided by the total number of hours in bed to get a percentage of my sleep time.  80% is believed to be a healthy number.

Thanks

[WilsonFitbit]

Hello @Joe.FIT Welcome to the community forums! 

Thank you very much for the detailed information! To help improve your sleep score, work on your overall sleep habits, heart health, and activity levels:

• Exercise earlier in the day or at least 3 to 4 hours before you go to sleep (source).
• Maintain a consistent sleep schedule (source). For more information, see How do I track my sleep with my Fitbit device?
• Meet the CDC’s recommended 150 active minutes per week (source).

For additional tips on improving your sleep routine and quality, see How can I use my Fitbit device to sleep better?. On a side note, Sleep score ranges are:

• Excellent: 90-100
• Good: 80-89
• Fair: 60-79
• Poor: Less than 60

Hope this helps! 

[kabella66]

I have this same issue -- it's not accurate.  Everyday I have to "add" in sleep time after I've gotten up to let dog out and gone back to bed.  But I don't know why I even do it -- because the tracker doesn't change.  It records the time I put in on the summary page, but shows nothing on the sleep graph.  It doesn't even revert back to the basic sleep graph.  It just shows the time before I got up and then nothing afterward ... except on rare occasions like this morning where it recorded an hour of light sleep ... when I was sitting at my desk, reading my email.   

[pwung0327]

I have given up on the accuracy of the sleep score, I ignore it. Indeed, I have asked time and again for proof, in the form of a published and juried paper correlating the Fitbit classification of the sleep stages through the heartrates with the brain wave studies which we assume to be the reference. No response.

And the moderator's attempt to obfuscate the issues that is plaguing this function is frustrating. They know they have an issue with the population that gets up in the middle of the night and who are not getting accurate readings and all we get is: you should sleep better, like normal people. How condescending.

I also just got a firmware update from Firbit, so I know they do updates on the devices, why not correct the problem through the firmware? Ob viously they don't care about the minority of the users who have this problem.

[Joe.FIT]

I'm not having that much of a problem.  Which Fitbit do you have?  I understand it does make a difference.I have a slightly older Fitbit Charge 3.  I compare it to my Garmin watch every day.  I wear both to bed, LOL.My problem is just with the sleep score.  When I wake and then go back to sleep, I get the right amount of sleep hours, but I get 2 separate sleep scores; both show fair to poor sleep.I think I should be getting one combines sore showing fair to good sleep.
Thanks for the comment. 

[S.Gray]

I belong to the same group that pwung0327 does -- old men who have to get up in the middle of the night.  My takeaway from reading through this topic is that for those of us in that group the Fitbit sleep score and sleep log are not accurate and cannot be accurate because it's programmed to start your sleep over every time you get up instead of logging a period of wakefulness in the middle of your sleep.  I do not expect an inexpensive wrist-worn device to equal or even approximate the accuracy of a sleep study, but it would have been nice to get a useful approximation.  To that end this thread has been helpful and I will lower my expectations.

[Sleep_deprived]

The algorithm almost certainly is only tracking time in bed. I sleep cycle disorder meaning I don’t cycle through stages and I wake up 7 to 8 times per night for no reason just to go back to sleep. It tells me I got great sleep because I got 45 minutes of deep sleep at the beginning of the night, 80 REM at the end, and a total “time asleep” of eight hours. Didn’t complete a single REM cycle, feel like crap, and could sleep another 3-4 in the morning. 
It is a very basic algorithm, it has to be.