When Apple vice president of technology Kevin Lynch announced at WWDC last week that Apple Watch would add an automatic handwashing-detection feature in its next update, many people naturally assumed the feature was prompted by the COVID-19 pandemic – a public health emergency that has strongly underscored the importance of handwashing.
Not so, Lynch told MobiHealthNews in a recent interview. The feature has been in development for years, he said, and the timeliness of the launch is just serendipity.
“With the presence of Apple Watch on your body, we feel a responsibility to explore other ways we can help support people with their health and wellness.” Lynch said. “And we’re really thoughtful about doing that in a gentle way, a useful way.”
The Watch team came upon handwashing after looking for a use case where Apple could be valuable in the domain of preventative care.
Handwashing detection turned out to be challenging for a number of reasons. It turns out the movements people use to wash their hands are not consistent – some people favor a rotary motion, for example, while others move top to bottom and back and forth.
The solution lay in an approach that’s come to define the Apple Watch team’s approach: what Lynch calls sensor fusion – combining data from multiple sensors to create one clear signal. For instance, the Apple Watch combines heart-rate data with motion sensing to do calorimetry on a variety of activities.
“Sensor fusion is a technique that we’ve been using since our early days on Apple Watch, and it provides a rich view into different activities,” he said. “We’re constantly looking for ways that we can combine the views from multiple sensors we have or perhaps new sensors that would add perspective that we don’t have yet.”
For handwashing, the relevant sensors are the motion sensors and the Watch’s microphone, which the team recently began exploring for the potential of things like noise detection. The microphone listens for the sounds of not only running water, but also soap squishing on hands. By combining all three markers, the detection becomes robust in situations where, for instance, an eco-friendly sink causes the running water sound to stop and start.
That is not to say the algorithm is perfect – Lynch admitted that it can still be fooled occasionally by activities like washing dishes that feature similar sounds and motions to handwashing.
Once Lynch’s team perfected the automatic detection, the next challenge was crafting a user experience that would be helpful and enjoyable and not nagging, condescending or annoying to others. That last consideration is what led the team to abandon ideas of using 20-second song snippets, for instance. Instead, the handwashing feature offers an animated countdown, reminders to continue if the user ends their handwashing prematurely, and a “good job!” message at the end – inspired by the approach the team has taken to its Activity app.
The feature is optional, Lynch said, so users can easily turn it off if they find it’s not for them.
For healthcare workers, the feature might appeal as a way to improve handwashing adherence. The CDC says that on average healthcare workers wash their hands less than half the times they should.
Lynch says Apple hasn’t taken any steps to create an enterprise version of this offering, but the tools are there for a hospital or startup to do so: The feature feeds handwashing duration data into HealthKit, which means that data will be available for app developers to work with, assuming users grant permission to share it. That means the feature could be used, for example, with a dashboard that lets teams see how they’re doing with handwashing.
“It’s very much a platform approach with health where we’re very conscious of the fact that there’s a lot of smart people in the world, and there’s a lot of work to do around health,” Lynch said. “So we’re contributing where we can and where we see an opportunity to make a difference, and we’re creating a rich ecosystem for developers that is really custom-built for each health domain.”