L. J. Wei
Primary Faculty

L. J. Wei

Professor of Biostatistics



Other Positions

Director, Industry Partnership Program


Harvard T.H. Chan School of Public Health


L.J. Wei's research is in the area of developing statistical methods for the design and analysis of clinical trials. In 1977-78 he introduced the "urn design" for two-arm sequential clinical studies. This design has been utilized in several large-scaled multi-center trials, for example, the Diabetes Control and Complications Trial sponsored by the NIH and the Matching patients to Alcoholism Treatments sponsored by NIAAA.

In 1979, he proposed a response adaptive design, a randomized version of Marvin Zelen's play the winner rule, was used in the ECMO trial, a well-known study which evaluated extracorporeal membrane oxygenation for treating newborns with persistent pulmonary hypertension. Currently several trials sponsored by private industry are using this particular design to relax the ethical problem arising in using the conventional 50-50 randomization treatment allocation rule clinical studies. To monitor trials sequentially for economic and ethical reasons, in 1982 Wei and his colleagues presented a rather flexible monitoring scheme, which has become a classical reference for the literature in interim analysis for clinical trials.

Dr. Wei has developed numerous methods for analyzing data with multiple outcome or repeated measurements obtained from study subjects. In particular, his "multivariate Cox procedures" to handle multiple event times have become quite popular. He and his colleagues are also responsible for developing alternative models to the Cox proportional hazards model for analyzing survival observations.

A very important issue in statistical inference is to check whether the model used to fit the data is appropriate or not. Currently, Wei and his colleagues are developing graphical and numerical methods for checking the adequacy of the Cox proportional hazards model, other semi-parametric survival models, parametric models, and random effects models for repeated measurements. The new procedures are much less subjective than the conventional eye-ball methods based on ordinary residuals plots.

Since the cost of computing has been drastically reduced, some analytically intractable statistical problems can be handled numerically. Presently, Wei and his colleagues are working on various resampling methods for quantile regression, rank regression, and regression models for censored data.

Dr. Wei is also a senior statistician at the Statistical and Data Analysis Center. He works closely with the medical investigators in Pediatrics AIDS clinical trials for evaluating new treatments for HIV patients.