She didn’t leave her room without wearing makeup, even to exercise, and she had a date for every football game. At one point, she got a poor grade in advanced mathematics because she had partied all night before the final.


For Murphy, it's dedication more than raw brain power that helped her win. “If you work hard—if you persevere—you can do well,” Murphy advises. She laughs when she says, “It’s just, I had to persevere a lot harder than everyone else.”

Case in point: Murphy’s research trajectory began with seemingly odd data. She was searching for a way to advance theoretical statistics with direct applications to real problems, and she came across a problem involving children with behavioral issues. 

“It looked like little kids who got more treatment actually had more problems, which doesn’t make any sense,” she says. Eventually, she figured out that the data showed reverse causation.

“They were implementing one of these adaptive interventions where they would reassess the little kid every semester,” Murphy says, “and if the little kid wasn’t doing well, they’d give the child more treatment. So the very children who were not doing well got more and more treatment.

“Of course, at the end of the study, the kids with the most treatment did the worst. The lack of progress prompted more treatment, not more treatment prompting the lack of progress. It was causation in the other direction.” It was then that Murphy realized she wanted to help people by improving adaptive interventions like this one.

Mobilizing Interventions

The experience reinforced her belief that evaluating the effectiveness of a patient’s intervention after the fact is problematic, and the resulting evaluations are easily challenged. “I knew that I wanted to move more into a study where we would collect data experimentally,” she says.

To do so, Murphy has been developing a statistical tool that she calls SMART—Sequential, Multiple Assignment, Randomized Trials. Rather than evaluating patient outcomes after an intervention is complete, Murphy helps clinicians implement SMARTs, which allows her to isolate and empirically optimize each individual treatment decision in an intervention. SMARTs highlight the link between a treatment and the patient’s response, which is especially useful for chronic disorders that involve complex series of treatments.

The SMART method is widely applicable to any issue that a clinician can treat with an adaptive intervention; for example, behavior disorders, obesity, diabetes, adherence to medication, mental illness, and substance abuse. Similar statistical methods also can be applied to make policy decisions, and Murphy regularly is invited to speak by social scientists such as economists.

But Murphy prefers to focus on mental illness and addiction—what she calls the “wild west” and “frontier areas” of science. Soon, Murphy wants to put real-time adaptive interventions into the hands of patients themselves—that is, when they reach into their pockets and grab their smartphones. Murphy will use her MacArthur grant to develop Just-In-Time, Adaptive Interventions (JITAIs) for mobile devices; these smartphone-based therapies will provide personalized treatment whenever and wherever a patient needs help.

Ultimately, Murphy envisions JITAIs and SMARTs as having great potential, not only for improving health, but also for applications like helping former prisoners avoid recidivism and for developing better welfare policy. “To me, these are interesting people who I care about. I want to help them. It’s important morally—these are people who are really struggling in their lives. And it’s important financially for our society—these are very high-cost individuals. There are compelling justifications in all kinds of domains.”


Cover photo courtesy of the John D. & Catherine T. MacArthur Foundation. Pictured at top: Vijay Nair, Shyamala Nagarej, Susan Murphy, and Jennifer Chu. Photo courtesy of Naveen Narisetty.