Meet Susan A. Murphy On and Off the Hockey Rink
Amy Munice, ALM Communications
In 2014, Susan A. Murphy could give students at least 1.75 billion reasons to go into statistics. This year, she has even more …
Why 1.75 billion? That’s the number of smart phones proliferating the planet, creating a window of opportunity for statistics research she enthusiastically points out to students at the University of Michigan whenever she can. Murphy says, “This is a fantastic time to go into statistics. Every time we have figured out a new way to collect data, there is an explosion in both the need for statisticians and great excitement for what can be done with the data. This is one of those historic times! We have all these new ways to collect data via wearable sensors, the Internet, genetics, astronomy, and more, and it’s creating all kinds of new problems for data scientists to solve.”
For Murphy, the way in which smart phones are opening doors for statisticians isn’t just theoretical. Smart phones and other mobile devices are key to her research on Just in Time Adaptive Interactions (JITAI). Whether it’s finding ways to help drug addicts, smokers, heart disease patients, dieters, or people with any number of similar chronic health conditions where behavioral change is key to clinical outcomes, Murphy and her many collaborators are looking at ways to use mobile devices for both data collection and as intervention tools.
By implementing micro-randomized trials on various intervention regimens, Murphy helps JITAI clinicians get sound data on what works and what doesn’t. Getting it right is no small matter to Murphy, who says she was originally attracted to creating solutions in the addiction field because “these are people who live on the margins of society and [are] usually not viewed with much empathy. From a more scientific perspective, what’s really cool is that, in the addiction and mental health fields, we understand how difficult it is for a person to be motivated to change their behavior, like avoiding risky locations and/or taking their medication on time… . Things can happen in a person’s life that throw them off course.”
Ping the patient too many times and they might turn off or tune out the clinical intervention app. Send messages to de-stress patients whose calendar shows they are having a jam-packed day and you may be spinning your wheels again, or worse— triggering them to figuratively toss their nagging smart phones into the waste bin. These are the kinds of issues Murphy’s statistics aim to accommodate, yet still help people improve their health. See her YouTube video in which she gives a pithy summary of how JITAI works.
Listen to Murphy talk and you are immediately impressed by her ability to cut through the complexities within data and data analysis to make both the promise and nuances of JITAI plain. Perhaps that ability was a driving factor for the MacArthur Foundation when they awarded her a MacArthur Fellowship in 2013. Murphy started to transition her research into JITAI mobile health about four years ago, having the usual challenges of gaining credence and getting funding. The Genius Grant changed that. “Now, because of the MacArthur grant, I have much greater cachet with behavioral and clinical scientists, who now are more willing to conduct clinical trials and submit funding proposals related to JITAI,” says Murphy. “They know they are taking years out of their life taking on these research projects, and the external acknowledgement of the MacArthur Award has been a swaying factor.”
Along with the MacArthur, Murphy gained additional clout with scientists with her election to the Institute of Medicine in 2014. There is a downside to these awards, however. Murphy is now facing an unprecedented number of requests for her time and attention. Her dance card was already quite full, even before the MacArthur award. In addition to her many collaborative research projects involving JITAIs, her duties include teaching and other responsibilities as the H.E. Robbins Professor of Statistics, research professor at the Institute for Social Research, and professor of psychiatry—all at the University of Michigan in Ann Arbor.
Since 2000, she also runs an active lab, the Statistical Reinforcement Learning Lab, which includes another statistician; a behavioral science researcher; and several post-doc fellows, graduate students, and undergrads. The lab not only works on statistics for JITAI development, but also the field that Murphy has pursued before—Sequential Multiple Assignment Randomized Trials (SMART), a type of clinical-trial design for helping clinicians provide the best sequence of treatments, adapted to their patients.
Murphy spends a good deal of her time espousing the benefits of adaptive interventions and JITAI to statisticians, clinicians, and scientific researchers. “A few years ago, we ran a workshop for clinicians at which we coined the phrase Just In Time Adaptive Intervention,” she says. “Now we see references to this acronym appearing across mobile health. Our practice starts with how you make intervention design sufficiently precise so that you can improve the intervention by using data. We work to help identify the components of JITAIs so as to help scientists think experimentally about how to manipulate the components so as to improve effectiveness.”
She continues, “It was only one year ago that we introduced micro-randomized trials for use in constructing JITAIs, and we already are seeing behavioral scientists moving on these ideas. It’s a new world now in terms of how to get ideas out. In the old world, ideas were touted in publications. In the new world, publication remains important, but so are tutorials, doing workshops, meetings, and running brainstorming sessions that help get the ideas out.
“With micro-randomized trials, our next phase is to collaborate on a series of high-quality trials; these trials will not only provide data to help scientists develop JITAIs, but will also spur us to modify the trial design to increase the usefulness of micro-randomized trials in mobile health. The MacArthur Grant is really helping us develop micro-randomized trials and encourage their adoption.”
If you view the research Murphy and her teams are doing through the lens of the national health crisis and runaway costs, you too may start chomping at the bit for results sooner rather than later. Indeed, Murphy comments, “Too much of the morbidity in the U.S. is due to health behaviors. We eat too much. We don’t exercise. We don’t take the pills that have been prescribed to us. We smoke. We don’t manage our lives well. When someone gets a chronic disorder—for example depression or addiction—and they try to deal with it, they often develop maladaptive behaviors. The illnesses themselves are maladaptive.”
She adds, “Think of being sedentary. It’s easy to be sedentary. The question is: How can we get you some support in your life that will help you be less sedentary and won’t aggravate you so much that you just delete the app?
“Wearable devices collect large amounts of data in real time. This is a Big Data problem with challenges such as figuring out which features of this vast amount of data (your location, how busy you are, volatility in stress) we use to do the adaptation as part of the JITAI. What I really want to see in the next 10 years is the development and trialing [sic] of statistical algorithms that will continuously update Just In Time Adaptive Interventions as people experience the interventions via the mobile device in real life. I think of real situations where people’s lives change abruptly—their child starts school or their spouse starts a new job. If we are providing support by mobile devices, we will need statistical algorithms that can quickly change when or which interventions should be provided when someone’s life has changed in these ways. In a decade, we will hopefully have tested many algorithms, and the ones that enable these interventions will be identified.”
Busy as she may sound—and she is—you’d be mistaken to think Susan A. Murphy is all work and no play. Play she does—last year, three times a week on various ice hockey teams where she gives full throttle to her parallel passion for athletics. Currently on a wing position, she’s played both offense and defense. Her teammates include people from all walks of life, including psychiatrists, engineers, and graduate students. Socially, these are her main group of non-work-related friends.
Murphy explains, “Hockey is completely consuming, so you can’t think about work. It’s a totally different way of using your mind than being a statistician, but it’s similar to academic work in that it’s very difficult. You have to constantly work at getting better.”
She adds, “I try to be on the ice as much as I can… . I’m not the best player in the world because the really good players started when they were four. I used to be a runner, but I tore my knee up and it turns out that ice skating is easy on your knees. Usually, I’m the only gray-haired person on the team!”
Ice hockey might strike you as an unlikely passion for a born-and-bred Louisianan like Murphy, who suffered through the “cold weather” in North Carolina when she earned her PhD at UNC Chapel Hill. But perhaps a window on just how deep Murphy’s ice hockey passion runs comes from her husband, Terry Murphy, a former athlete who competed at the state level when they met in college. He comes to her games when he can and is eager to give her pointers. Then again, he can only take so much given Murphy’s native abilities. Maybe that helps explain why he has created a rule that she can only talk about hockey for one hour a day. Murphy quips, “If he left me tomorrow, hockey is probably all I’d ever talk about.”
Then again, get Murphy talking about mobile health and Big Data and you realize ice hockey is but one passion. Just listen to her talk to newly minted PhDs at UNC:
Where it’s SMART or JITAI, statistics-driven health care is always a passion and one reason why the MacArthur Foundation, the ASA, and all those in the know count Murphy as one of the most influential statisticians of our time.