Wendy Nilsen | Smart & Connected Health

Computing Across Disciplines – ACM FCA

Wendy Nilsen is Program Director for the Smart and Connected Health program, a partnership between the National Institutes of Health and the National Science Foundation. She shares what interdisciplinary research looks like from the funding agency perspective, and explains her role as a program officer. We also discuss the Mobile Health Training Institute, which she helped found to build interdisciplinary skills across computing and health, and the essential skill of building relationships and asking questions.

Bio

Wendy Nilsen, Ph.D. is a Program Director for the Smart and Connected Health Program in the Directorate for Computer & Information Science & Engineering at the National Science Foundation. Her work focuses on the intersection of technology and health. This includes a wide range of methods for data collection, data analytics and turning data to knowledge. Her interests span the areas of sensing, analytics, cyber-physical systems, information systems, big data and robotics, as they relate to health. More specifically, her efforts include: serving as cochair of the Health Information Technology Research and Development community of practice of the Networking and Information Technology Research and Development Program; the lead for the NSF/NIH Smart and Connected Health announcement; convening workshops to address methodology in mobile technology research; serving on numerous federal technology initiatives; and, leading training institutes. Previously, Wendy was at the NIH Office of Behavioral and Social Sciences Research (OBSSR).

Transcript

Andrew Miller: Well, so Wendy, thank you so much for talking with us today. The first question that I ask everyone on this podcast is what first got you interested in research or in science or in computing?

Wendy Nilsen: Mine is a step-wise interest. I actually was in, before I moved into science, I was in industry and I ran a large company. I was a production manager and then the vice president of manufacturing. I found that I didn’t understand people’s behaviors the way I wanted to, things were happening that didn’t make sense. I learned how to deal with them as a manager, but I really wanted to understand exactly what was going on. Was there a process, where should I be intervening, what should I be doing, what kind of …? Why were people doing thing that were clearly seemingly irrational.

I ended up leaving industry, going back to school, and getting a degree in psychology, which told me that I was never going to know all those questions unless I worked really hard. That led me into the work that I did in my early research and then, as I realized that the work I was really interested was at a higher level, at a science policy level. I came over to the NIH. That was exciting. Gave me a very different view, but it also then exposed me to computing. Once I realized that computing had so much to offer, the health world in general, but especially I think the behavioral community is it’s such an open area. So much of health, if you ever look at the NIH website and the images that are changing, the stories that are coming on every day … either physician behavior, patient behavior, environmental issues, and people’s behavior. It’s talking about opioid epidemic or it’s talking about heart disease, all of these have a behavioral component and thinking about how does computing work in with that, computing and engineering is become, and I’m really excited by.

Andrew Miller: Yeah, so you were originally in the Office of Behavioral and Social Sciences Research. Is that right?

Wendy Nilsen: Yes. In the Office of the Director at NIH.

Andrew Miller: I also believe that you were involved very early on in what started out as the Smart Health Program and has now bloomed into the Smart and Connected Health Program. What was that like? What was it like doing this delicate dance between the NIH and the NSF and health and basic science research.

Wendy Nilsen: Well, I started this because … or I didn’t start. NSF started Smart Health and Wellbeing with me, Misha Pavel and Howard Wactlar, but they were really kind of looking at how to position it, how to think about it. They had started. It was early 2012 was really their first year moving. I met Misha Pavel in 2013 thanks to one of my colleagues in computing and we decided to partner, which I didn’t know you weren’t supposed to do because it never occurred to me I wasn’t. I wasn’t encouraged to partner across government. We partnered and one of the reasons, the main reasons, that we partnered was because I found out we spoke different languages, we had different cultures. What I considered to be my science, computing instead was an application.

I found out that often I was viewed as a clinician. Often my colleagues in computing were either programmers or technicians, so there was a huge gap in understanding. There was also a huge gap in the developmental pipeline. If you don’t start from the beginning as a multi-disciplinary team, it’s very hard to once something’s built to then make it work in a new system that is completely different than the system you developed it in. I think a great example is falls, research in falls. For years, research in falls was done with college students pretending they were going to fall and predicting that. We thought that was the best way to do it. It turns out that the people that fall don’t fall anything like a college student, so all of that research had to be redone with real people who fortunately don’t fall that much, but creates all new issues.

We started this program really to build the pipeline of people who are working together from the beginning on the right problems in using really synergistic and transformative approaches to changing problems that are literally killing our country.

Andrew Miller: Yeah. One of the interesting things I remember from being a student whose work was funded by that program was that suddenly those of us on the computing side were exposed to this whole world of these other people on the other side and that was just sort of on the researcher side. I imagine that you mentioned briefly that you spoke sort of complementary or different languages. How did you negotiate that balance? How do you think about the way that that program has developed over time?

Wendy Nilsen: The language is something we’re learning. I’ve run many, many, many panels at NSF every year. Many for the Smart Health Program and I think one of the things that research review helps us do is to build that language. To have a conversation, really respectfully and excitedly explore. When I’m talking about dissemination is it the same thing you mean? When you’re talking about ground truth, is it the same thing I mean by gold standard? What are we talking about, how are we talking about them? What’s important, what’s not important? How do we calibrate things to the various disciplines? I think this is what’s exciting and I see it developing in the community in terms of conferences and meetings. I see it also happening in review and especially, the best of it in my funded projects or the NSF fund health projects.

Andrew Miller: A couple of years ago now you crossed the boundary and moved from the National Institutes of Health to the National Science Foundation. What was the impetus for that and what was that change like for you?

Wendy Nilsen: Well, the impetus for that was it was more fun at NSF, but oh, I shouldn’t probably say that, should I? No, nothing against my colleagues at NIH, but I mean, we talk about lifelong learning, I love lifelong learning. I come over to NSF, I learn new things every day. I have colleagues in computing, engineering, physics, you name it, and the sciences are so exciting when they come together. When I came here I started learning that what I knew in bringing people together crossed the boundaries at NSF, so I’m a team member or a leader on many of our projects that cross the boundaries between disciplines because that’s one of the skill sets I have is building … I’m a whatever on this, science representative on our convergence working group and because I have the expertise on how do you bring people together. It’s been really an exciting thing. I came over in 2015 … or 2014 I think originally, as a rotator, just like most of the program offices at NSF, and then when they asked me to stay, I was thrilled, so I stayed.

Andrew Miller: I’m going to ask the basic question that I’m not sure I know the answer to, which is what does a program director do?

Wendy Nilsen: A lot of things. I would say most days the number one thing I do is talk to PIs. I talk to people that are proposing, I talk to people who’ve been reviewed, I talk to people that are getting funded. Some are much more fun conversations than others. Other days, I spend the entire day or two in review. We read, as program officers, if you’re leading a program you get all the proposals that come in, you separate them into different review committees, you are a leader with your peers on getting those reviews done, you set up the protocols for across agencies. You have to set up cross agency protocols. After you get done, you get to talk to PIs, you get to get information from them and forms completed and all sorts of things updated. Later, you have … like in September of this year we have the Smart Health PI meeting, which actually is really exciting because this year we are actually bringing in a hundred researchers who are not funded researchers into the PI meeting, so we can build our community even better.

We spend time going to regular conferences, going to workshops that are funded by our programs, PI meetings with our programs, doing review, and basically learning from people, from proposals, and meetings all the time. It’s great. Come.

Andrew Miller: Yeah, so you got to one of the things that I think is interesting about the way you approached your position, which is that not only are you doing proposal review and talking to PIs, but you also are involved in a number of these interdisciplinary meetings. You talked about the Smart Health PI meeting. I had the opportunity to go to one of those a couple years ago and it was really fascinating. You also have been involved in the mHealth Training Institute. Was that something that you were involved in creating?

Wendy Nilsen: I created that. Yeah, I did, and I must say it’s been exciting. I’ve been out of it for a few years now, but it’s been a very exciting thing. I actually started the mHealth Training Institute when I was at NIH because I was actually leading NIH’s efforts in mobile. Mobile health technology, this was early. We had Palm Pilots. This was an early, early period and what I realized is that people weren’t talking. They weren’t having the right conversations. Basically, what health people thought you could do with mHealth technology was … or mobile health or mHealth, was send text messages. When I would come over and talk to colleagues at NSF, I would go, “They’re doing a whole lot more than text messaging. Why aren’t we doing this thing?”

We talk about the gaps between agencies, but one of the gaps is if you don’t ever have the conversation, you don’t know what you’re missing. If you’ve never done something, you don’t know what’s the problem, so we started the mHealth Training Institute in 2011. I came back from a meeting with the Department of Defense where I’d learned about scenario-based training where you teach experts with consultation with experts, but it’s not didactic, it’s consultative … and actually, it was in partnership with NSF and Qualcomm.

Andrew Miller: Yes. Talk us through how that training institute works. What constitutes and immersive, cross-disciplinary event like that?

Wendy Nilsen: Our goal for it was to not have people being lectured at all the time, so we created teams of five to seven people depending, with two mentors. Every team picked before they got there a problem and it was a health problem. You might say that health problem is the high rate of hypertension, high blood pressure in African Americans between the age of 30 and 50. Somebody on the health team knew what it was, so you came in with the problem. There’s this huge problem, there’s this huge disparity in these health outcomes for this group. What we would do is sit down and then you would start to learn about behavior, you would start to learn some about computing, some about engineering approaches to solving these, so maybe it’s analytics, maybe it’s HCI, maybe it’s sensor development from an engineering approach. Each stage you’ve modified the project. You grew it and you modified it.

Often when computing comes in the room, nobody’s ever thought of those things from a health perspective and they’re just like, oh, that’s so exciting. Basically, what they ended up doing was building an entire research protocol, a research project from start to finish over a one-week period. It’s a cumulative project. There’s no right or wrong, there’s no answer on this and the projects can vary as much as sensors on we’ve done asthma projects where they were kids in the emergency room using stress sensing and inhaler sensing with [GIS 00:13:38] and some parent education to one of the projects that I mentored was a sex sensor for HIV research. HIV research has always stuck by what’s ground truth for sex. It becomes a whole different question when you think in a research study could I actually be able to determine variables like that?

Those are really fun and interesting questions about how do you get to those places, so this way I think the thing that I love about interdisciplinary training is there’s no experts. Everybody’s an expert. Everybody has knowledge and what’s been so exciting and so fun was the mentors are all very senior, very established, and they end up partnering with some of the junior people we bring in because they say, “Oh, my gosh, they know things I’ve never known. They know all sorts of exciting and interesting things and are interested in things that I would never have known about until I met them.” It levels the playing field, it puts people in a position where they can really learn from each other. That’s where it’s fun.

Andrew Miller: The first step seems to be getting people in the room together and giving them patterns for how to speak to each other across these gaps. Then, the next step is sort of, okay, but how do you make it count for everyone? How do you make it, something that’s innovative in both the computing context, for example, and then the specific health context?

Wendy Nilsen: I think it’s a very complex dance to make sure that all the voices are heard. Interdisciplinary work takes somebody who’s really committed to it. Whoever’s leading, but also in the partners, it takes a lot of work to sit and listen and not open our mouths and not engage in being an expert in everything. Most of the time we’re rewarded for being experts in everything because where are domains are, you know, what we really deal in. You think about it, you go to a conference, you’re an expert, everybody looks at you and says oh my gosh, your science is so strong. You’re this and this and this. Now, all of a sudden, you walked in a room and you’re stupid.

I spent years being stupid, so if you don’t take the opportunity not to be the expert, you’re not going to learn anything. I’ve walked into rooms where I thought, “Oh my gosh.” I went to a conference once. It was a workshop right before the conference and the two people leading that workshop who are wonderful researchers were scrawling equation after equation on the board and my head dropped lower and lower and lower. Finally, somebody next to me said, “You’re not an engineer, are you?” I said, “Oh, no. I’m not.” He said, “Well, at the break I’ll explain what’s happening to you. It’s no big deal. You got it.” He spent a half hour saying to me, “Don’t worry about the equation, worry about this.” We started talking and I had things to contribute, I just didn’t know where to start. I think that’s the thing that you have to be able to do if you’re willing to be interdisciplinary.

Andrew Miller: Yeah, I mean, in one of the … this, sort of, the mHealth Training Institute is a microcosm for these experiences that people doing interdisciplinary work are having across, throughout the year, in their actual working lives. One of the things that I think confuses people and I am not fully clear on this either is the extent to which working in an interdisciplinary context means that you have to develop that secondary expertise. Do we have to actually become these hybrid experts or is there some other kind of expertise that we gain while retaining our independence or our expertise in our own discipline. Finding that balance, I don’t know if you’ve had experience in helping people through that, but that’s something that I feel like a lot of people are wondering about.

Wendy Nilsen: Well, I’m so glad you’re wondering. That balance is the hard part. Think about the old metaphors of Jack of all trades. You can’t keep up with the literature … I think we can barely keep up with the literature in our areas at this point. Now, think about the complexity of the things that you’re dealing in, so if my expertise is in machine learning, okay, well I can go to the top machine learning conferences, I can go to the … but wait. There’s a special edition published, I have to read those. I’m reviewing and then … Well, wait. I want to see some in my application area. Oh, wait, that opens up a thousand more publications. Uh-oh. Oh, and then I’m doing it in heart disease, so now do I need to know about heart disease? Do I have to know ejection fractions? You learn the words. If you’re going to deal with somebody on heart disease, you’re going to learn the words and they’re going to learn yours. They never talked about a feature in their life, but eventually if you get a good team, they’ll be able to say, “Is that the kind of features you’re looking for?” All of a sudden, you’ll have language and you’ll say, “So, I thought you were doing, what was it, the ejection fraction on this?” I’ll go, “Oh yeah, no, but we wouldn’t be doing it on that.”

I think the thing is truly, and I think of these as transdisciplinary teams where they’ve gone from working in parallel, doing some projects together to the real synergistic work, which is transdisciplinary. It really requires people to not only … they don’t know each other’s literature. You can’t know each other’s literature. I see them come in day after day, you know, proposals, whether it’s heart disease, where the application’s heart disease, autism, diabetes. Autism is a couple thousand articles every month. How do you know it? You don’t. You partner with somebody that knows it and eventually you smart partner with somebody who works in the area of autism you’re interested in because even when you get to a disease … We were talking about it this morning, somebody was saying I want to do cardiac disease and I’m working with a surgeon and we wanted to do congestive heart failure now. Congestive heart failure’s a progressive, degenerative disease in your heart. It’s not a surgical thing, so is a surgeon the perfect person to work with? Probably not.

You would only learn that over time, but once you learn that, you’ll say, “Oh, I needed this kind of person this time,” or you will have the connections made that will help you be able to figure out. I think about it as if you think about strong ties and weak ties in a network, once you know me, you know a whole community of people around me. Then, if you need somebody else, now you can connect to my network, so you have a strong tie to me, but you’ll have a weaker, weak tie, but a tie nonetheless to the rest of that community. Does that make sense?

Andrew Miller: Yeah. I mean, that’s super helpful. Navigating that and finding those partnerships is a huge part of the work. I guess that brings me to just a more parochial question. Most of the people that will be listening to this are from the computing or computer science side of the world and we’re just sort of discovering about all these other disciplines. What are some things that we do well that we can leverage when we do our interdisciplinary collaborations and what are some things that you see people from the computing or CS world do that you wish they did differently or advice that you would have for us?

Wendy Nilsen: Well, I love this question, but I would say it’s not just the CS community. The most common error, whether it be led by medical, whether it led by engineering, social behavioral, it doesn’t matter, what I see is that they don’t take the time. If you’re going to really do interesting, exciting, interdisciplinary, multi-disciplinary, transdisciplinary convergent proposals, it takes time. It’s not just the time to meet people, that takes time, but at some level that’s the easy part. You have to find somebody you want to work with because you’re going to invest a part of yourself and somebody once said, “You only have so many hours, be careful of what you invest them in.” I think that, for me, is kind of how I live my life. I’m not going to waste time on things that aren’t important, so if you meet somebody to work with and you don’t like them, don’t work with them.

You’re going to have to spend hours working together. Find out who else don’t you know? Who else should be involved with this? You’re going to have to develop that language, which they can tell you enough that you can make intelligent decisions and actually really transcend what you know. They need to spend that time with you, so that they can just transcend what their natural limits are. It’s investments. The one thing I must say is in Smart Health we call them staple proposals. They come in and they’re just, “Here’s a computing person, oh, here’s the heart disease person. Oh, wait, there’s a mathematician on this, too. Oh, look, there’s an HGI person on there, too.” They each have a section and they don’t even match with typefaces, which is where it really gets embarrassing, but they’re stapled and the whole panel sits there and says, “Can we write it’s stapled in the panel summary?” We say, “No, but you might say it would be a much better job if you actually had everybody on the team read the proposal.”

Andrew Miller: Right.

Wendy Nilsen: You have to read each other’s proposal.Rread a proposal that your colleague is putting into NIH, you don’t even know what you don’t know until you read it and go, “Oh, my gosh. This is like wow. This is exciting. I don’t know what’s happening here.” Your students, you, everybody, you get to learn all sorts of new things. It’s all about relationships at some level. I think almost everything. Everything that has a big impact is about relationships, so it’s relationships in science, it’s relationships between different knowledge fields. Those are where the important things are. That’s where the hard stuff is.

Andrew Miller: I imagine it’s challenging to put together a panel and to find the right reviewers for a proposal in general, but I imagine it’s even more challenging when you’re dealing with a variety of different health concerns and a variety of different technologies. What is your secret sauce or your approach when you try and match reviewers and panels for the kind of work that you do?

Wendy Nilsen: There is no secret sauce. It’s a lot of work and I’m very, very grateful to my colleagues here at NIH and I’m very grateful to the community for volunteering. NSF, NIH, when we do joint projects, the panels are bigger. They’re bigger, there are more people, so if your program officer asks you to come, come. I just had one the other day. There were 18 people in the room and three on the phone. It’s really exciting and interesting, but it really gets down to the fact that to understand the complexity of all of the aspects, you really need to bring in people with expertise across the areas.

What’s exciting about that is people end up like exchanging business cards, going to lunch together, actually they’re all out at dinner together. That’s fun. That’s exciting. I’m happy about that, but also, what I think is really exciting is they start to talk in a way they haven’t talked before. They start to say, “Well, I’m working, but the person that I work with hadn’t talked to me about any of that. Can we talk more about this? Can we do that? How do we come together around these questions?” Then, they start to, by the end of the second day, they’re asking the questions. This morning I was in a panel and you couldn’t tell the difference for the most part of who was who on the second day because they’d all listened to each other on the first day.

Here I have a pediatrician sitting to my right who says, “So, they didn’t give any indication of what they would use for features. Did they? There’s nothing new in the feature extraction is it?” That’s a pediatrician, which made me smile ear to ear because that was exactly what the person across the table from her had said the day before.

Andrew Miller: That’s amazing. I mean, that’s what you want. Right?

Wendy Nilsen: They had a conversation about the first time somebody uses the word features. Wait, wait, wait, what’s a feature? Well, oh, so it’s what we call factors. I got it, but it’s… Okay. You develop the language, you develop the words, and it’s really the generosity of people being able to bring in enough for them, enough of them to really kind of make some magic.

Andrew Miller: Is there anything in particular that you feel like you want to make sure that people across computing know when they’re thinking about this kind of work or things that you think that people should know about the kind of job that you do or just anything you want to plug?

Wendy Nilsen: Oh, I do want to plug. You know me. Come to panels, see what we do. I’ve sat in rooms with people that are so brilliant and I sit with them day after day. They always walk out going, “That was so cool. I had no idea.” Even when you’re in your institutions, it’s very often that a computing program in a medical school aren’t even on the same campus. I think it’s actually the exception rather than the rule that they’re together.

I was in Vegas recently in Nevada and found out that Reno has the med school, Vegas has the engineering program, so saying, “Do you have a Webex? Do you have any video conferencing to your meetings? Could I see them?” Spend some time. Get to know them. Go to a health conference. I went to a conference the other day with a colleague from the National Institute of Mental Health and she thought, “All I have to do is tell them that we have money. They’ll do stuff with computation in it.” She went there and went, “Wendy, they’re not speaking the same language. I have no idea what they’re talking about.” We were having a conversation about application areas and she said, “I don’t have any application areas.” I said, “Yes, you do. It’s your whole life.” We had a long discussion. We ended up having a great time and she learned a ton. It’s the listening and don’t be afraid to feel stupid because you aren’t. It’s only the smartest people who are willing to put themselves in that position.

Andrew Miller: Well, I can’t think of a better way to end this conversation than a call to don’t be afraid to feel stupid, so thank you so much for talking with me today.

Wendy Nilsen: Oh, thank you so much for talking with me. It was fun.

SaveSave

SaveSave

SaveSave

SaveSave[/vc_column_text][/vc_column][/vc_row]

SaveSave

The post Wendy Nilsen | Smart & Connected Health appeared first on ACM FCA.