Dr. Jane Paik Kim is Clinical Assistant Professor in the Department of Psychiatry and Behavioral Sciences at Stanford University School of Medicine. Her professional aim is to improve public mental health through the application and development of statistical methods in mental health research. Her research interests are statistical methods for digital health interventions delivered through mobile or wearable devices, and psychiatric ethics research. Her statistical interest areas are in the robustness of regression-based inference for both clinical trials and observational studies, as well as methods development for survival data arising from non-standard biased sampling schemes.
+ Full Transcript
Brian Tarran: This is Brian Tarran with Significance Magazine here at JSM 2019 in Denver, Colorado. And I’m with Jane Paik Kim of Stanford University School of Medicine. Hi Jane.
Jane Paik Kim: Hi. Thanks for having me.
Tarran: Welcome. We’re going to be talking about mobile health today but I thought we could start with you telling listeners a little bit about yourself, your background, your areas of interest? That would be wonderful.
Paik Kim: Of course. I am a trained statistician. My Ph.D. was in statistics and right after I graduated from Columbia, I went to Stanford School of Medicine and I am now in the department of psychiatry. I’ve been in the department of psychiatry for six years and I’ve been collaborating with researchers in the mental health space.
Tarran: Okay. And so your talk is today at JSM?
Paik Kim: Yes.
Tarran: It’s on the mobile health project.
Paik Kim: Yes.
Tarran: Specifically using wearable devices to- is it to encourage people to be more active? Tell us a bit about the project and what led to it.
Paik Kim: Of course. So this is a project- so Stanford launched a center for digital health recently and as part of its launch, it opened up an Apple watch seed grant program calling for proposals to use the Apple watch in ways that would impact health. So a team of researchers at our department, which includes myself as a clinical psychologist, some other lab members, we proposed using the Apple watches to message people to improve their adherence behaviors. And so adherence behaviors that we focused on were very simple- physical activity. So, getting people to set a goal for themselves and stick to that goal, and we let them have the watch for eight weeks and we’re going to message them over that course of time.
Tarran: So the question that you’re looking to answer is do these devices- do these prompts- help us to meet our goals to achieve what it is we’ve set ourselves to?
Paik Kim: Yeah, that’s one question that we are interested in, absolutely. I think that a lot of the digital health industry, and the academia, it’s the question of “Can technology be used to actually impact health behaviors?”, and so, exactly.
Tarran: So what’s the underlying driver for this research? Why are you asking these questions? The devices now are available, but is there a particular health issue that you’re looking to address through this?
Paik Kim: Great question. So, in the area of mental health, apps have been used to provide treatment for all kinds of disorders and a lot of therapies that are available through manuals and in-person treatment are now digitized, and so the question is can these apps really work? And can we get people to use them? And I think the central question is can adherence behaviors be improved and can technology be used to target and improve adherence behaviors? And so what we’re focusing on is a very simple adherence behavior among healthy people, but I think the question of using technology for mental health is way more complicated, and it’s a question that needs to be addressed before going into that area.
Tarran: Of course. So how many people did you have on the study?
Paik Kim: Right now we’re in the very early stages, so we’re generating interest lists. But we do have 300 watches that we are intending to distribute.
Tarran: So each person gets a watch, it will be set with a goal, maybe 10,000 – 15,000 steps a day, or be active every 30 minutes, or something like that? And so the data that you’re collecting, are you getting all the information from the phones so that not only what time are the reminders going out, but then are you recording movement and activity? Is that the sort of information that you’re getting?
Paik Kim: Yes exactly. We’re getting both watch data, and phone data. So the question that I’m focusing my research on, which I’m talking about later today is does the watch actually provide useful information? Which is not very different from some of these questions from when the watch first came out, like is it going to actually be useful? So I’m talking about- I’m trying to use statistical inference or hypothesis testing to actually test whether some of these variables, some of the data that we get from the watch is useful.
Tarran: So when you say useful, what do you mean? Useful in a statistical sense or can you give us a bit more detail?
Paik Kim: Yes, exactly. So, can it be used in such a way to provide- so what we’re basically doing is we’re getting people to wear their watches and then we observe their physical activity and every time we observe their physical activity we’re going to send them a sequence of messages that are based on some kind of decision role, and then this process repeats. So we send them messages, then they read them, and we observe how much they move over the course of the day, and that serves as input for this decision role. And this decision role is updated every day in such a way to maximize their success in achieving their goal. And the question is whether that information is useful in providing the right message sequence that’s intended to have them reach their goal every day. So whether it can be useful for that decision role- is the question.
Tarran: So you say that it’s in the early stages- the project?
Paik Kim: Yes.
Tarran: So where are you at in terms of your findings so far, or early indications?
Paik Kim: We’re in early stages, as in, truly early stages. We’re completing the app development, so we are dealing with P.H.I., protected health information, so we do have to make sure that our app is HIPAA compliant that we’re protecting information. And so we’re working within our institution and making sure that everything is compliant. So actually building the app part wasn’t too difficult and it was fairly quick but it’s mostly this compliance issue, which could be a whole other topic. But something that we’ve been learning in real-time.
Tarran: Well, I guess I mean if you’re collecting watch data, phone data, that’s quite- you expect quite a big insight into people’s lives, right? So I guess the privacy issues are particularly crucial.
Paik Kim: Exactly. And we’re also asking about PHQI, so it’s an instrument that’s generally asking about their mental health information.
Tarran: It’s very sensitive information.
Paik Kim: Yes.
Tarran: So when you present later today you won’t be saying, “We’ve cracked it. This information is useful.” It’s more of an introduction to a work in progress?
Paik Kim: Right, it’s more of, “This is what can be done”, and hopefully generating some collaboration. I think I outlined some issues that are common to mobile health and also issues that are specific to this particular project that can generate some application of model statistical methods, and hopefully, that will catalyze some collaboration.
Tarran: Okay and in terms of the mobile health space more broadly, a couple of years ago there were some presentations here on that topic, but they were in quite early stages then. How far have we come in the space of three or four years in terms of figuring out a role for these devices for making people more active?
Paik Kim: Wow, that’s a loaded question. I think there’s still work to be done. I mean clearly it’s a huge industry, and there’s a lot of brilliant work that’s being done in academia that’s addressing the causal inference and there’s work to be done. But obviously the industry is still interested in it and I think we haven’t really cracked this issue of whether it’s efficacious, whether technology can really be used for that. So clearly it’s not going anywhere. The field is not going anywhere so I think it’s an exciting place to be in right now.
Tarran: The watches themselves – I guess they’ve always been kind of a slightly elite device, not everyone is going to have them, right? I mean Apple- there are non-Apple versions that are more affordable. But even in phones, you’re starting to see these fitness apps built-in. Mine tracks my steps and congratulates me if I’ve hit my 10,000 targets. So it’s everywhere, and I think statisticians like yourself want to know how to make use of that information. But I think for people as well, it’s something they’ll be looking at trying to understand and how to optimize their lifestyles based on that information, I guess.
Paik Kim: Yeah, hopefully. That’s the assumption, that people are actually interested in improving their health, so I think it’s – it seems like there are a bunch of tech companies that have figured out how to get people hooked to certain technologies and certain apps, but that’s not exactly the same question as getting people to use apps that are trying to benefit them.
Tarran: Yeah. I find my devices- I’m quite happy when it tells me I’ve met my goals but if I haven’t there’s no real punishment or downside, so.
Paik Kim: Right.
Tarran: So maybe my adherence is not quite as good as it could or should be.
Paik Kim: Yeah.
Tarran: Excellent, well I hope the rest of the talk goes well for you.
Paik Kim: Thank you.
Tarran: And the rest of your stay at JSM.
Paik Kim: Thank you.
Tarran: Thank you for talking to us today.
Paik Kim: Thank you
Tarran: My name is Brian Tarran and I’m the editor of Significance Magazine. Find us online at significancemagazine.com. For this special JSM series of podcasts we’re collaborating with Stats and Stories. Stats and Stories is a partnership between Miami University’s Departments of Statics, and Media, Journalism and Film and the American Statistical Association. Follow us on Twitter, Apple podcasts or other places where you can find podcasts. If you’d like to share your thoughts on our program send your email to Statsandstories@miamioh.edu, or check us out at Statsandstories.net. And be sure to listen for future editions of Stats and Stories, where we explore the statistics behind the stories and the stories behind the statistics.