Better Notifications for Better Health (Audio Described Version)

Smartphones, wearables, and medical devices use alerts to inspire users to take healthy actions, such as going for a walk. However, alert fatigue can set in when these alerts occur too frequently and/or at bad times.

The National Library of Medicine is funding a study on counteracting alert fatigue by timing alerts when users are more likely to need an alert, be receptive to the technology’s support, and be able to act on the alert. These more useful alerts can come from an encouraging digital health coach, and they have shown promise for inspiring positive changes in users.

https://reporter.nih.gov/search/GMGCF...

Non-AD version -    • Better Notifications for Better Health  

Video transcript:

[Eric Hekler] So this project, funded by National Library of Medicine, this is all about trying to figure out: How do we find the exact right time and right place to give people support to help them walk more. So we see this impacting public health through basically trying to build a really smart, caring,
loving digital health coach in your pocket for increasing physical activity, changing health behaviors,
just improving your lives.

[Mohamed El Mistiri] The study involved around 50 participants and two intervention methods that were used in unison with one another, one of them to provide notifications within the day or walking notifications, and the other one is to provide goals that tell the individual how many steps they should walk based on their behavior for the last cycle of 26 days.

[Hekler] This is a collaboration. We have three, multiple, PIs, principal investigators, on this project, and the idea was we made the assumption that people are different, context matters, and things are constantly changing. We also made the assumption, though, that that happened predictably for each person, which would mean we can actually discover meaningful patterns for each person that we can use to help find only those moments when we can provide support, when they would have the opportunity to act, when they would be receptive to receiving support from the technology, and where they would actually need it.

[El Mistiri] So alert fatigue is when people get tired from notifications, and that reduces the effectiveness of behavioral interventions that are supposed to help people be healthier. And we try to study ways to dynamically model the system to know when to intervene, when to send a notification that is going to be most lucrative in the short term and in the long term.

[Hekler] What we discovered is that we could identify at least one meaningful moment, meaning a
moment where we could confidently predict that you would increase your steps, for 91% of our sample. So we basically went from I think like 18% all the way up to 91%. Another way to say this: We could actually discover, there's always been this idea of a teachable moment, right? We actually can now carefully identify everybody's teachable moment is what this is suggesting. And we can do that in a way that we can use that information to actually only provide support when somebody's in that sort of teachable moment or in that just-in-time state, right? That's a pretty novel idea. And I can
imagine a lot of people basically saying that's impossible. You're never going to be able to do something like that, but NLM took a shot at us, and I think it paid off.

[El Mistiri] Simply stated, this project would have not been possible without the NLM because the NLM provided us with the funding that allowed us to design the experiments, build the infrastructure to deliver the experiments, recruit the individuals into the study. And none of this would have been possible without them.

[Hekler] And we're only just getting started, you know. This was such a unique study and experiment that we're pretty confident this will lead to a variety of other insights as people engage with using
different types of data-analytic approaches. This is a really valuable data set. So that data set wouldn't exist if NLM didn't provide us with this support.
#humanhealth #bioengineering #behavioralhealthcare #audiodescription #audiodescriptions