Skip to main content

Tianchen Qian

“We need the right intervention, for the right person, at the right place and time.”

Understanding Causal Relationships

Professor Qian aims to elucidate causal relationships through the analysis of highly complex data. “My methodology research focuses on causal inference, longitudinal data, sequential decision making, and semiparametric models,” he says. “I am interested in a wide range of scientific problems, especially those that are health-related, such as mobile health and clinical trials.”

Advancing Health in the Digital Era

Much of Professor Qian’s methodology research applies to mobile health — the use of mobile devices to promote healthy behavior change, such as physical activity, smoking cessation or stress management. Mobile health interventions are usually delivered in the form of push notifications or app interactions. “One of the big challenges is knowing how to deliver the right intervention content, at the right time, to the right person,” he says. “By answering this question, we can build mobile health interventions that are effective and not too burdensome, thus realizing the full potential of utilizing digital devices to promote health.” Answering this key question involves collaboration of multidisciplinary teams. From a statistical viewpoint, this would combine experimental design (novel study designs to collect mobile data), causal inference (understanding the effect of interventions), and sequential decision making (such as reinforcement learning methods to adapt the treatment delivery rule in real time).

Modeling Treatment Effect with Complex Data

One project Professor Qian is working on is the modeling of treatment effect using longitudinal data with functional outcomes and time-varying treatments. The project’s data comes from a mobile health study where each participant was repeatedly randomized throughout the study to receive (or not receive) a real-time intervention — a push notification suggesting physical activity. The longitudinal outcome at each time point is a minute-level step-count curve for 60 minutes following an intervention. “We are trying to answer whether/how the push notification affects the step-count curve, and whether this effect vary across different participants.”


Ph.D., Biostatistics, Johns Hopkins University, 2017

B.S., Mathematics, Tsinghua University, 2012


Stats 265: Causal Inference
Stats 120B: Introduction to Probability and Statistics II
Stats 210C: Statistical Methods III – Longitudinal Data

Research Areas

View Biostatistics


The application of statistical methods to analyze and interpret data in the fields of biology …

Recent News

See all news