How do I collect wearables data through a custom mobile app on a budget?
This questions was asked to us by an external collaborator, who had already written a mobile app to collect specific Apple HealthKit / SensorKit data but had a very limited budget, and needed to figure out how to get the data without incurring a huge cost. The answer would probably be helpful to other researchers, so I decided to post it here.
Answer (1)
The answer depends on what resources are available at your organization or academic institution. For the most part, you need a mobile app to send data to an API, then a way to move the data from there to appropriate, secure storage. You’ll also need to get the app through the appropriate app store verification process, which is simpler if the app follows standard security practices.
Getting the data from an app to you requires hosting an API somewhere (AWS, Azure, etc.). That’s about half of the cost, and you can keep it down by just keeping the data there temporarily until a different process gets it out. I’ve seen some researchers simply email the data to themselves, but that poses a security risk and requires users to actually hit “send” on the email.
Then, to store the data, it needs to land in a database or file server somewhere. If you stay within the same cloud offering, you need to add storage cost (tiny) and the cost of getting the data out (that’s where they get you!). Also, add the compute cost for moving the data from the API to storage. (The API could technically push the data to your storage server directly, but I would advice against it to avoid performance and data sync issues with the app).
If you don’t stay within the same cloud environment, then there is also the matter of corporate firewalls – it is very difficult to get an IT department to open up the firewall for incoming connections, so it's best to host software internally that would make an outgoing connection to the API.
In MiNap, we used an internal server to host a script to reach out to the API and retrieve the data, then an internal database to store it. Both the server and database would be free to us in our institution. If your IT department doesn’t offer that, you could check if they have an automation tool like Microsoft PowerAutomate or TDX iPaaS. With those, you can schedule a workflow to get the data from the API and store it in something like SharePoint, Dropbox, or internal storage (if the automation tool is setup to cross the firewall). Just make sure both the automation tool and your storage location have approval from your organization's IT security, compliance, or information assurance department.
https://github.com/DepressionCenter/MiNap
Another option that might work if you don’t have a lot of concurrent participants is Guava Health. It’s an iOS and Android app that collects data from Apple Health, Google Health Connect, EHRs, and user-entered symptom tracking. Users can share data with you via a provider link, and from there you would manually download the data. I’m not sure if breathing disruptions are available – you can always install it, sync to your apple watch, and see what data it shows! Their development team is very responsive to enhancement requests. But, again, make sure it is approved by your organization.
Later on, as you get more established and get a bigger budget, I would highly encourage you to consider a study management system (e.g. Avicenna, MetricWire, MyDataHelps, ExpiWell, etc.). These are cloud-hosted apps that can manage your entire study including enrollment, surveys, app notifications, form data collection, sensor & wearable data collection. On average, they can run around $10k/year for small-to-medium studies.
Gabriel Mongefranco
Mobile Data Architect
Eisenberg Family Depression Center, University of Michigan