TITLE: Targeted Maximum Likelihood Estimation (TMLE) for Machine Learning: A Gentle Introduction
SPEAKER:Professor Mark J. van der Laan, University of California at Berkeley
MODERATOR: Bill Wang
The standard method for evaluating the efficacy of a new treatment is via conducting a placebo controlled superiority trial. However, a non-inferiority (NI) trial is considered when the use of a placebo arm is not ethnical or not feasible. The design, conduct, analysis and interpretation of a NI trial are more challenging than that of a superiority trial. With its increased applications in more and more disease areas in the drug development, a clear understanding of all aspects of a NI trial is crucial to the success of the trial. This short course is designed to give the audience a comprehensive understanding of the NI concept, the methods and considerations of its design, and its applications in various disease areas. The fixed margin approach, the synthesis method and the Bayesian method will be discussed for designing a NI trial. Real world examples are used to demonstrate the selection and calculation of the M1 and M2 margins when the fixed margin approach is used. Case studies involving Bayesian approaches are also included.
Many factors may impact the design, the analysis, and the interpretations of the results, such as assay sensitivity, constancy assumptions, and quality of current and historical trials. These will be discussed thoroughly in the presentation.
Dr. Susan Wang is the Asia Head of Biostatistics and Data Science of Boehringer Ingelheim (BI) Pharmaceuticals China. Before moved to China, Susan had over 20-year of experience working on global drug registrations in multiple disease areas in BI US as a lead statistician. She designed the non-inferiority RE-LY trial for BI that has led the approval of the indication of Pradaxa for stroke preventions in patients with AF. Susan received her Ph.D. in statistics from State University of New York at Stony Brook. She enjoys applying statistical innovations in drug development. She is an associate editor for J. Biopharmaceutical Statistics, a core team member of DIA statistics committee, and an associate chair for distant leaning in Biopharmaceutical Statistics
Gang Cheng (Ph.D. in Biostatistics from The University of Iowa) is a senior manager of Biostatistics in the Department of Biostatistics & Data Sciences Asia at Boehringer Ingelheim Pharmaceutical in China. Gang majorly focuses on therapeutic area of metabolism. His experience includes clinical trial design and data analysis, statistical methodology and applications in clinical development.