Simulating a wearable medical device to test a healthcare app

Wearable medical devices are just the start of a revolution in healthcare innovation. As they track symptoms and routines and take measurements, they build a body of data that can tell your healthcare team more about you than ever before.

That is, if you have a healthcare app to collect the data from the device and record it in a way that makes sense.

Developing and testing systems to get wearable device data

A client came to us with a wearable medical device meant to connect to an application on the user’s phone. It communicates with the device over Bluetooth Low Energy (BLE) to download usage statistics and give users the option to add details to the data, like the level of pain felt during an activity.

The wearable device already existed — its job was to be as unobtrusive as possible. Our job was to build out the healthcare app, the main way users would experience and interact with the device. This is where the data would be made meaningful.

To find that meaning, however, we needed a way to populate the app with every approximation of how people might use the wearable device. Physically trying out a bunch of different workout regimes was not going to cut it — there are too many combinations, too many variables and not enough time.

Instead, we had to create a way to simulate different combinations of data the wearable device would theoretically be sending. By building a simulation on a laptop and connecting the mobile application to that, instead of the wearable medical device, we can fill the app with whatever data we want. This workaround let us find and test the edge cases, bugs and other issues that have to be resolved before anyone can rely on the application as part of their healthcare plan.

Challenging the data possibilities of health wearables

With the simulated device, we can speed up the feedback cycle. Instead of wearing the device and working out for an hour, we can tell the simulation to “pretend” we’ve been working out for the last hour. That allows us to test all the data possibilities as we can hypothesize:

• When the user wears the device all day, what data gets sent? Which data if they exercise? How does the application respond to this data? How does it get presented to the user, and to the cloud server?
• What data does the device send on first use? Does the mobile application work correctly when the wearable device only has a few hours of data on it?
• If the device has been in use for a long time, will large size data syncs cause any issues with the mobile application?

The team that built the medical wearable gave us great documentation, which — alongside spot checks of the actual output — allows us to make sure all of the simulated data is an accurate representation of real-world data from the wearable device.

Building a better relationship between healthcare app and wearable device

Already, the simulated wearable is allowing us to test the healthcare app in situations that wouldn’t be possible without it. We need to know how the mobile app will behave with a year of data on it — and we need to know that now, not a year from now. We need to test all kinds of scenarios where the device loses power. Will the data save, will it connect properly the next time? Now we can test that directly, without needing to physically manipulate the device each time.

To access the data collected by the wearable, the app has to have a “communication handshake” with the device — essentially a connection — followed by repeated requests for more data. The simulated wearable allows us to quickly see how the mobile app reacts to all kinds of variations of this process. More importantly, it lets us quickly re-test the variation once we believe we have a solution in place. We can simply run the exact same scenario, tweaking until it works. More testing means that when the wearable device does make it to market, more people can rely on it.

As we get closer to a final product, we’ll be able to do it confidently, knowing that we’ve successfully tested and solved for as many pitfalls as possible. This will ultimately make user testing easier and the product safer, allowing users to truly rely on the device and its health data.