Expedite Rare Disease Drug Development Through Bayesian Analyses, External Controls, and Real-World Data

• Activity Number: 135 - Multiplicity, Missing Data and Other Topics
• Type: Contributed
• Date/Time: Monday, August 9, 2021 : 1:30 PM to 3:20 PM
• Sponsor: Biopharmaceutical Section
• Author(s): Florence Yong* and Jeffrey Palmer
• Companies: Pfizer Inc. and Pfizer Inc
• Keywords: Rare Disease; Bayesian Analyses; Real World Data; Drug Development; Data Sharing; External Controls

Abstract

Drug development for rare diseases has numerous challenges including small population size, poor understanding of the natural history, and lack of well-defined registrational endpoints for clinical trials. We discuss our experience in utilizing Bayesian analyses and machine learning to facilitate earlier decision-making. Case studies will be presented to illustrate the integrative approach from data gathering opportunities to the incorporation of external control and RWD to overcome small sample sizes. This includes the application of machine learning techniques to help understand patients’ characteristics, develop endpoints, and inform study design. Bayesian approaches to estimate the posterior probability of decision criteria for proof of concepts will be discussed. Analyses led to the early termination of the development of an investigational drug product (IMP) unlikely to reach patients’ bedside or the faster advance of the IMP can reduce patient burden and ultimately bring treatment to patients with unmet medical needs faster. The presentation should provide insights into some enabling approaches and opportunities for more efficient drug development.