Duration: 23 minutes
Presented by Eric Hekler, PhD
Speakers have kindly provided responses to questions submitted by conference participants during the Discussion session that did not get an opportunity to be discussed.
- Question from Anne Berman – You mentioned that this is easier than you think. Tell us how you get started?
The short summary is to find a control systems engineer at your university. They will most likely be in areas like robotics, mechanical engineering, supply chain management, or electrical engineering. If you find an engineer working in one of those areas and digital health, they will likely have some background in control engineering. I really emphasize this “easier than you think” in this longer talk for more details: https://prevention.nih.gov/education-training/methods-mind-gap/using-control-systems-engineering-optimize-adaptive-mobile-health-interventions
- Question from Linda Carlson – In the continuous tuning example, just wondering how the intervention components (step goals and reward points) were chosen? I can see how the levels are tuned, but are the components themselves just based on previous research showing what kind of intervention components are useful?
Yes, they were chosen based on prior research.
- Question from Anne Berman – Has system ID experimentation been used to compare automated ”course” tuning to human health coaching? Why or why not?
Great question and, as of yet, no. The reason is that the first System ID study conducted on humans for behavior change (as far as we can tell so far) was conducted by use a few years ago. This is still a very new concept that we are exploring. This is a great question to eventually explore once we feel we’ve developed a robust enough tool. I don’t think we are there yet though with the technology.
- Question from Anne Berman – Say I would want to do such a trial, how difficult – or easy – is it to get funding for the technical aspects of the trial? What are your hot tips on funding?
This is an important question. In brief, the key is to find a study section that would “get” it and then, of course, to seek out pilot data to get some preliminary data. My experience is that you can’t jump straight to large-scale funding (e.g., R01-equivalent in the US). You need to do the pilot work to demonstrate what you will do; this makes the promise concrete for reviewers.
- Question from Reyhaneh Yousafi – How do you capture participants’ feedback regarding the trajectory of changes in their performance, or how they feel about the goals and general process, any reliable or valid method?
Awesome question! that’s an area Pedja and I are spending a lot more time on; building more checks on our goals and algorithms so that people know what we’re doing and why and sign up. In my view, I want these tools to eventually be a bit more like glasses. We know exactly what they are doing and why and we can always choose to take them on or off whenever we don’t want that. I want what this tech does to be as transparent as that and built so that people can turn it on and off, as they need greater “clarity” t9 use the vision metaphor.