Speaker

Biography

Abstract

Kaspar Rufibach

Kaspar Rufibach is a biostatistician who is passionate about supporting statisticians and drug developers to continuously challenge the status quo, with the aim of improving the drug development process, making it more efficient, and enabling access.
Kaspar has co-founded and co-leads the special interest group “Estimands in oncology” which has (as of August 2024) more than 100 members from 50 institutions globally. He has also co-founded and co-leads the EFSPI statistical methodology leader group which has 14 members from 14 companies. He regularly interacts with regulators globally on various joint projects.
Kaspar’s research interests are methods to optimize study designs, platform trials, advanced survival analysis, probability of success, estimands and causal inference, and estimation of treatment effects in subgroups. Kaspar received training and worked as a statistician at the Universities of Bern, Stanford, and Zurich. From 2012 until 2024 he worked at Roche before joining Merck KGaA in October 2024 to co-lead its Advanced Biostatistical Science group.

More on the oncology estimand WG: http://www.oncoestimand.org

More on the EFSPI statistical methodology leaders group: https://efspieurope.github.io/efspi/methods/methods_intro.html

More on Kaspar: http://www.kasparrufibach.ch

I will start with providing a few examples of very valid scientific questions in drug development that typically ask for causal answers, but which are routinely answered in ad-hoc ways that rarely allow for a causal interpretation. Further reasons why I believe a clinical biostatistician needs to know about causal inference will be given. I will conclude with a call to apply and develop statistical and causal inference methodology to fill the gap between valid causal questions and routine ad-hoc answers.

Susan Gruber

Susan Gruber, co-founder of TL Revolution and Founder of Putnam Data Sciences, is a biostatistician and computer scientist specializing in  causal inference and predictive modeling.   Her work focuses on improving methods and tools for generating robust real-world evidence to support biopharmaceutical and medical decision-making through Targeted Learning. Her tmle R package on CRAN has over 70,000 downloads worldwide. 

Targeted Learning is a framework that combines causal inference, statistics, and machine learning to address complex issues in analyzing data from randomized controlled trials and studies that incorporate real-world data. This talk provides a high-level introduction to the Targeted Learning Estimation Roadmap, statistical analysis using Targeted Maximum Likelihood Estimation (TMLE), and the role of sensitivity analysis to assess the level of support for drawing a substantive conclusion from the study findings.

Florian Lasch

Florian is a Biostatistician with a degree in mathematics and a PhD from Hannover Medical School. Florian works as a Biostatistics Specialist at the European Medicines Agency, providing scientific support to development and evaluation throughout all stages of marketing authorisation assessments of medicinal products, and leads the ACT EU Priority Action on Clinical Trial Methodologies and the EMA Estimands Implementation Group.

The estimands framework facilitates the application of thinking and methodology developed in the causal inference community to the design and analysis of clinical trials. This presentation will reflect on the opportunities and challenges of applying causal inference methodology to clinical trials. A case study in Alzheimer’s Disease where the intercurrent event ‘initiation of symptomatic medication’ is handled with a hypothetical strategy will illustrate the key points.