Google account for Fitbit

[Triplennn]

Hi! We have several clinical studies utilizing Fitbit, each involving hundreds of patients. Many of our patients are seniors, and to ensure smooth account management, we usually set up Fitbit accounts for them before sending out the devices. However, we've run into difficulties due to Fitbit's migration to Google accounts. Now we're faced with the challenge of creating a large number of Google accounts without running into any blocks from Google. We would greatly appreciate any guidance or suggestions. Thank you in advance for your assistance!

[Gordon-C]

Hi @Triplennn 

With the migration to Google Accounts, we came across some unexpected challenges which have mostly impacted researchers who configured Fitbit accounts on behalf of their participants.  We are working with our product teams to determine if we can resolve this.   At the moment, the best solution we have is to ask each of your study participants to create the Google account themself and link it to Fitbit.   Some researchers are personally walking their participants through this process while others are providing detailed instructions or resources.

Creating a resolution will not easy as it involves working with multiple teams at Google.   Keep in mind that we take our user's data privacy and security very serious at Google, so we appreciate your patience during this time.

Gordon

[bklounge]

Hi Gordon,

Similar to original poster @Triplennn, technology is a barrier for many patient populations in our research studies. For older, or physically / cognitively impaired study participants, technology in general can be a challenge. For lower income participants, it's the lack of access to technology. 

Asking these study participants to create Google accounts doesn't solve the problem. Instead of creating Fitbit accounts for study-provided Fitbits, we're still stuck with manually creating Google accounts for study-provided Fitbits. Thus, the importance of having the ability to create Fitbit or Google accounts in bulk for study-provided Fitbits. It's labor-intensive and tedious to do and keep track of everything and ultimately limits the size and diversity of study populations.

BK

[Gordon-C]

Hi @bklounge 

I can understand the challenges this new normal is creating for research.   Our developer and privacy policies clearly state the user is in control of their data.  This includes the ability to consent certain data types and revoke consent to a research study.   Some participants might need some extra hand holding by your research team to onboard with the Google and Fitbit accounts.

If there are some additional online resources you believe you need to help onboard your participants, please post them here.  I can see what we have to share or need to create.

Gordon

[Kevin_H]

I'm glad you're aware of researchers having challenges with the Google acquisition around creating study-owned accounts, thank you for addressing it. 

Would you please give us all a status update periodically? I would be willing to devote a big chunk of my time to helping out.

In case no one has had this idea yet, perhaps you could expand your researcher registry (the one you use to allow intra-day series) to capture an application email address, then allow the Gmail "+" hack for creating sub-accounts?

In terms of sensitivity to data ownership, our studies allow participants to adopt the Fitbit account at the end of the study, so they would then have full access to their data.

For the last 10 years I've been creating research study systems using Fitbit, where a step in the on-boarding process is for the researcher to create a Fitbit account using the Gmail "+" hack where anything in the address from a "+" up to the "@" is ignored. This was a very elegant solution as it allowed us to manage the Fitbit correspondence with a single Gmail address.   

We've tried other work-arounds, but it seems it's not possible to create a large number of actual Gmail accounts from the same device. I have three studies going live in the next two months and could really use your help.

[bklounge]

---

@Gordon-C wrote: If there are some additional online resources you believe you need to help onboard your participants, please post them here.  I can see what we have to share or need to create.

---

Now that you've asked @Gordon-C! Here are some resources or capabilities that would be really helpful:

• Make Google accounts extensible in the same manner that Gmail accounts are using "base+ID @Gmail.com, for example, study+12345 @Gmail.com, so only one parent account (study @Gmail.com) is needed" as described by @sol_38 in this community post and echoed by @Kevin_H above.
• Most of our studies use more than one wearable or connected device (we're a large academic medical center). As a result, we're creating a mobile app to help orchestrate the participant device onboarding process. When the study provides a) Fitbit + Google acct credentials or b) Fitbit, it would be very helpful if there was an SDK for Fitbit mobile app so that we could embed it into our app and make the device onboarding experience seamless and simple. Short of that, the following would still greatly improve the onboarding experience:
  
  • Fitbit mobile app login page supported autofill. For example, our mobile app passes study-provided acct credentials to Fitbit mobile app login so participant doesn't have to navigate back-and-forth between apps (copy-and-paste twice) or re-key from study instructions paperwork.
  • Fitbit mobile app allows 3rd party apps to request that it generate local notification (pop-up) that, if tapped, redirects participant 3rd party app. For example, we have no way of knowing if participant has paired Fitbit is ready to return to our mobile app in order to consent sharing device data with study.

Happy to jump on the phone to discuss.

Appreciate you asking us, Gordon.

[Samantha_G]

Hi @Gordon-C ,

Our study team has been asking participants to create Google accounts themselves and linking it to Fitbit, but we're running into issues with a limitation on the number of times a phone number can be used to verify an account. We've had multiple participants who have tried creating a study Google account, but they are told that their number can not be used to verify the account because it has been used too many times. It seems there may be a max on the number of accounts that someone can create using the same phone number which presents an issue for us if participants already have other google accounts besides the study account we ask them to create. Can you please advise how we can get participants who have already used their phone number the max number of times set up with study specific google accounts? Appreciate your help!

[Gordon-C]

Hi @Samantha_G 

It is not necessary for the user to create a new Google account for your research study if they already have a Google account created.   The users can use their existing Google account and link it to their Fitbit device and your research study.   I believe in this scenario, the account will already be verified.   

Gordon

[Samantha_G]

Hi @Gordon-C ,

Thanks for your response!

Do you have any recommendations on how to proceed if we do not want participants to link their personal google accounts. Creating study accounts allows us to ensure participant privacy since they are de-identified. Asking participants to grant us access to Fitbit data connected to their personal Google accounts presents participant privacy concerns. Appreciate your help!

[afolarin]

Hi @Gordon-C I think the guidance from Fitbit here specifically recommends against doing this, i.e. using personal email addresses for reasons of anonymisation this is the guidance we also received from our ethics boards. It is still therefore important that we are able to create multiple de-identified email addresses for the participants or be able to use aliased "+" addresses as mentioned previously. 

https://healthsolutions.fitbit.com/research-pledge/#:~:text=If%20you%20create%20Fitbit%20accounts%20....

[Sandynnn]

Hi there! We are recruiting 600 participants for our study and have already purchased 600 watches. We are now urgently tasked with manually creating 600 Gmail addresses. Have we found a solution for how to create multiple de-identified Gmail accounts as required of ethics review? @Gordon-C Look forward to your response! 

[Gordon-C]

Hi everyone,

We have come to the conclusion with the recommended solution that your participants should use their own Google account with your studies.  We do not have a way for researchers to create a bunch of accounts.  There are limits in place that previous a single user to creating a bunch of unique accounts and getting them verified.  So, you should not ask your participants to create a unique account for your study and verify it again their phone number.

@afolarin brought up the statement "If you create Fitbit accounts for research subjects, do so in a way that does not reveal their identity. For example, do not use their real name or email address, and default their account settings to the most privacy-protective option available."   This can be done using the user's personal Google account.

The only scope that provides the user's personal information is "profile".  The profile scope will allow your application access to the Get Profile endpoint which contains the user's personal information.  See Get Profile Response.  As long as your application does not request this scope, it is technically not possible to get the user's name, email address, etc.   When the user completed the authorization flow, the only information you get about the user is their Fitbit user id and the scopes they enabled during consent.   Here's an example of what that looks like, taken from the documentation.

{
    "access_token": "<access_token>",
    "expires_in": 28800,
    "refresh_token": "<refresh_token>",
    "scope": "activity heartrate nutrition oxygen_saturation respiratory_rate settings sleep temperature weight",
    "token_type": "Bearer",
    "user_id": "<user_id>"
}

In the past when you were creating anonymized accounts, you were getting this same information.   After discussing this solution with several research partners, they were in agreement this was sufficient.

If there are questions about this, please let me know.   

[afolarin]

Hi @Gordon-C 

I think in principle - yes people could use their personal gmail accounts, create a personal FitBit account then use this to authenticate and authorise data collection. However, there are a few issues with this we have encountered with this workflow (without exposing their personal gmail).

1. There is no way to "blind" the participant from the data (e.g. locked 2nd Android acc), otherwise this can be a confounding factor.
2. It can be very useful for the research team to have access to the data for verification of data collection from the REST-API. Troubleshooting etc.
3. It is commonly the case that we will set everything up for the participant, including the FitBits for data collection, this way they can be sent to the participants preconfigured for the study.

Suggestions / workarounds welcome. 

[Gordon-C]

Hi @afolarin,

Here is some feedback on your concerns 

• There is no way to "blind" the participant from the data (e.g. locked 2nd Android acc), otherwise this can be a confounding factor.

I believe blind studies do not adhere to our research guidelines.  The Fitbit users are suppose to have access and control over the data they share with any commercial or research application.   In the same Research Pledge you referenced earlier, we state

https://healthsolutions.fitbit.com/research-pledge .

• It can be very useful for the research team to have access to the data for verification of data collection from the REST-API. Troubleshooting etc.

Through the Web API, you will have access to the user's data.  Many cases, you will get more information out of the Web API than directly logging into the user's account.  There is information throughout our documentation to help you troubleshoot issues.   Anything you do not see in the documentation, contact our team for assistance.  We'll be glad to help.

• It is commonly the case that we will set everything up for the participant, including the FitBits for data collection, this way they can be sent to the participants preconfigured for the study.

Also, it is a common practice for researchers to provide written instructions and/or in-person (i.e. clinic, phone, chat, etc) assistance to help their participants through the Fitbit setup and consent.   We encourage you to do the same.  If there are additional resources that you need, please let us know.   

[afolarin]

HI Gordon, a further question came up from the ethics on this matter. 

If we use the participant's personal Google account to sign up, their personal details will leave the EU/EEA (i.e., they will be associated with their FitBit accounts). Previously, we were not able to create a dummy account for the participants. Is this unavoidable?

Kind regards,

[mlcollins914]

Hello @Gordon-C, I recently saw a webinar where Google shared they would be soon releasing a clinical-research focused product/capability that would enable researchers to create larger numbers of Google accounts for this specific use case of using Fitbits and de-identified email addresses for logging into those device in a trial. Can you speak to the status of this, and when it will be released?

[afolarin]

@GordonFitbit I was also wondering, as it has been a number of months now if there is any update on this. It would be helpful to know even if a broad expected date is available.

[ddempsey]

Hi @mlcollins914, would you happen to have a link to that webinar?