Using funnel plots to develop riskbased monitoring rules for binomial and Poisson outcomes in clinical trials.pptx
Sep 28, 2016 8:04 PM
Anastasia Dmitrienko (Blue Valley High School, Overland Park, KS) and Richard C. Zink (JMP Life Sciences, SAS Institute, Cary NC)
Anastasia Dmitrienko is a senior at Blue Valley North High School.She has been interested in statistics and worked on the development of clinical trial software tools such as risk-based monitoring tools (she has presented two posters on this topic at regional and national conferences).Anastasia is a member of the American Statistical Association and has been a JMP user for over a year.
Richard C. Zink is Principal Research Statistician Developer in the JMP Life Sciences division at SAS Institute, following eight years in the pharmaceutical industry. Richard is Statistics Section Editor for Therapeutic Innovation & Regulatory Science, and holds a PhD in Biostatistics from the University of North Carolina at Chapel Hill, where he serves as an adjunct faculty member. He is author of Risk-Based Monitoring and Fraud Detection in Clinical Trials Using JMP and SAS.
Risk-based monitoring (RBM) strategies are commonly used in clinical trials to detect trial sites with unusual characteristics such as excessive adverse events risks.This presentation focuses on the development of control limits and monitoring tools such as funnel plots and heat maps for binomial and Poisson outcomes. The control limits are based on asymptotic and exact confidence intervals for the corresponding parameters.Important factors that affect the expected amount of variability such as patient’s time on study are considered when computing the control limits.The methods developed in this poster are applied to define RBM rules for monitoring patient discontinuation (binary outcome) and number of adverse events per patient (count outcome) in several real clinical trials. These RBM strategies are implemented using custom JSL scripting and existing JMP software platforms.