2019 Challis Lecture

Susan Murphy
Harvard University
November 4th and 5th
Reitz Union, Room 2355
4:00 PM to 5:00 PM


Professor Susan Murphy is Professor of Statistics and Radcliffe Alumnae Professor at the Radcliffe Institute at Harvard University and Professor of Computer Science at the Harvard John A. Paulson School of Engineering and Applied Sciences.  She is known worldwide for her research on experimental design and causal inference to inform sequential decision making, particular with regards to sequencing treatments in health and in mobile health intervention development and for inference for high dimensional models.

She received a Ph.D.in Statistics from University of North Carolina, Chapel Hill and a B.S. in Mathematics from Louisiana State University. Before going to Harvard, she was H.E. Robbins Distinguished University Professor of Statistics in the Department of Statistics at Univ. of Michigan (as well as Professor of Psychiatry).  She began her career at Penn State University.

She has held numerous leadership positions including serving (or served) as President of the Institute of Mathematical Statistics (IMS) and the Bernoulli Society.  She has also received many prestigious honors including being awarded the Royal Statistical Society Guy Medal in Silver, being selected as a MacArthur Fellow, being elected a member of the U.S. National Academy of Sciences and the National Academy of Medicine, and being elected a fellow of both IMS and the American Statistical Association.


General Lecture

Monday, November 4, 2019

Personalized HeartSteps: A mHealth RL Algorithm for Optimizing Physical Activity
HeartSteps is a mobile health intervention for individuals who have Stage 1 Hypertension. The goal of HeartSteps is to help individuals alter their lifestyle so as to avoid taking hypertensive medications.   Mobile health interventions involve sequences of treatments delivered to individuals as they go about their everyday life.  Ideally we would design a system that learns and relearns the best treatment for an individual given their current context (location, weather, mood…).   This is reinforcement learning.  Multiple challenges confronted us in designing an online reinforcement learning  algorithm for HeartSteps including high noise in the reward, potentially strong delayed negative effects of the actions, non-stationary rewards and the need to conduct causal inferences at trial end. We discuss how we use the an initial study HeartSteps V1 to confront these challenges.
Technical Lecture

Tuesday, November 5, 2019

Stratified Micro-randomized Trials with Applications in Mobile Health
Technological advancements in the field of mobile devices and wearable sensors make it possible to deliver treatments anytime and anywhere to users like you and me. Increasingly the delivery of these treatments is triggered by detections/predictions of vulnerability and receptivity. These observations are likely to have been impacted by prior treatments. Furthermore the treatments are often designed to have an impact on users over a span of time during which subsequent treatments may be provided. Here we discuss our work on the design of a mobile health smoking cessation study in which the above two challenges arose. This work involves the use of multiple online data analysis algorithms. Online algorithms are used in the detection, for example, of physiological stress. Other algorithms are used to forecast at each vulnerable time, the remaining number of vulnerable times in the day. These algorithms are then inputs into a randomization algorithm that ensures that each user is randomized to each treatment an appropriate number of times per day. We develop the stratified micro-randomized trial which involves not only the randomization algorithm but a precise statement of the meaning of the treatment effects and the primary scientific hypotheses along with primary analyses and sample size calculations. Considerations of causal inference and potential causal bias incurred by inappropriate data analyses play a large role throughout.