Smarter Confidence Intervals for Percentages in Multicenter Trials: Applying the BSVI Correction to Improve Precision and Coverage
Confidence intervals are an important component of clinical trial reporting, supporting the accurate and reliable interpretation of study results.
When analyzing percentages in multicenter trials, key considerations include which confidence interval method to use and whether, and how, to account for differences between centers. For a 95% confidence interval, coverage should be approximately 95%, while avoiding intervals that are unnecessarily wide. In some applications, including vaccine studies, the width of the confidence interval can even influence whether a study meets its objectives.
In this webinar, Paul Talsma, PhD, Principal Statistician at Phastar, and Francesco Innocenti, PhD, Assistant Professor at Maastricht University, will share findings from their systematic review of key methods for constructing 95% confidence intervals for percentages.
They will also introduce the Binomial Sum Variance Inequality (BSVI) correction, a new approach designed to account for differences between centers while achieving adequate coverage with minimal interval width.
The session will provide practical guidance on selecting and applying the most appropriate methods when working with percentages in multicenter trials, helping attendees achieve reliable coverage while maintaining precision.
What You'll Learn
* Current methods for calculating a 95% confidence interval for a percentage and how they perform in practice
* Why differences between centers need to be considered in multicenter trials
* The background and rationale behind the BSVI correction
* How the BSVI correction can help achieve adequate coverage while maintaining precision
* How to use the BSVI correction in practice
Meet Our Speakers:
Paul Talsma, PhD
Principal Statistician, Phastar
Paul Talsma, PhD, has been a Principal Statistician at Phastar for seven years and has 37 years of experience working as a statistician, including 20 years in the pharmaceutical industry. His career spans academia, pharmaceutical companies, CROs, and the medical device industry, giving him extensive experience applying statistical methods across a range of clinical research environments.
Francesco Innocenti, PhD
Assistant Professor, Department of Methodology & Statistics
Maastricht University
Francesco Innocenti is an Assistant Professor in the Department of Methodology and Statistics at the Faculty of Health, Medicine and Life Sciences (FHML) of Maastricht University (NL). His research focuses on optimal design, sample size determination, and statistical methods for multilevel populations, with additional work in psychometric methods. He coordinates and teaches statistics courses at the Bachelor’s, Master’s, and PhD levels and provides statistical consultancy to researchers in fields including hepatology, gastroenterology, nutrition science, geriatrics, and clinical research. He obtained his PhD in Statistics from Maastricht University in 2021, following an MSc in Statistics and a BSc in Economics from the University of Florence (Italy).
Moderated by: Ian Wadsworth, PhD
Associate Director, Statistics, Phastar
Ian is an Associate Director of Statistics at Phastar, where he has mainly worked on late phase oncology studies since 2018 and leads the technical strategy within the statistics function, supporting staff with statistics research, training and conference submissions. Prior to this, Ian worked at Lancaster University as a Research Associate in medical statistics for six years. Whilst in this role, Ian studied part-time for a PhD in Statistics looking at extrapolation in paediatric medicine development. Before his PhD, he studied for a BSc in Mathematics and MSc in Statistics, also at Lancaster University. Outside of Ian's Phastar role, he has also been a member of the PSI Scientific Committee since 2022.
Register Below:
Stay Tuned—More Webinars Coming Soon
