The following short courses will be held on Sunday, 14th of September 2025 (the day before the start of the main conference).
| Time | Course | Instructors |
|---|---|---|
| Full-day | Adjusting for Covariates in RCTs: Translating Guidance into Practice | Dominic Magirr, Mark Baillie and Alexander Przybylski |
| Half-day (morning) | Multiple imputation for nested data: From theory to practice | Shahab Jolani |
| Half-day (afternoon) | Applied modeling in drug development | Sebastian Weber and Lukas Widmer |
Keywords: covariate adjustment; estimands; robust
estimation
Duration: full-day (09:00-17:00)
Course type: hands-on course
Instructors:
The FDA’s 2023 guidance document on Adjusting for Covariates in Randomized Clinical Trials for Drugs and Biological Products emphasizes key principles while introducing new techniques. This course aims to enhance participants’ understanding and implementation of these techniques in study protocols and statistical analysis plans. Focus will be placed on defining the target estimand and its impact on estimation procedures, including the calculation of treatment effect estimator variance. Hands-on exercises and case studies will provide practical experience in executing covariate adjustment analyses across different scenarios. The target audience for this course is statisticians involved in the design and analysis of randomized clinical trials with continuous, binary, and time-to-event endpoints. Throughout the agenda, participants will explore the guidance, examine practical implications, gain hands-on experience, and learn how to apply these techniques effectively to their own trials.
Learning objectives:
Participants will be able to identify in which contexts covariate adjustment is useful, and be able to engage in discussions with different stakeholders on the pros and cons of covariate adjustment. Participants will also learn how to specify covariate adjusted analyses in study protocols and statistical analysis plans, as well as how to execute covariate adjusted analyses for continuous, binary and time-to-event endpoints.
Prerequisites:
Participants are required to have some familiarity with clinical trial design, including an awareness of the ICH E9 Addendums on Estimands. Some basic knowledge of R software is also assumed. Participants need to bring their own laptop. Practical sessions will be run in Posit cloud.
Keywords: missing data; multiple imputation;
multilevel models; IPD meta-analysis
Duration: half-day (morning; 09:00-12:30)
Course type: hands-on course
Instructor:
Multiple imputation (MI) is widely used to address the issue of missing data in practice. However, standard implementations of MI often assume independent data, making them unsuitable for nested or clustered data, such as multicentre studies and individual participant data meta-analysis. Recent developments in methodology now enable imputation of multilevel data that effectively preserve the hierarchical structure of the data.
This course describes the difficulties in handling missing data in multilevel settings, notably the challenge of accounting for the multilevel structure of the data and addressing the coexistence of systematically missing data (where a variable is missing for all individuals in a study) and sporadically missing data (where a variable is missing only for some individuals in a study).
Participants will gain insight into two main families of imputation methodologies – joint modelling and fully conditional specification (FCS, or chained equations) – along with their respective strengths and limitations. Focusing on the FCS framework, we provide some theoretical background and demonstrate how the imputation model must be tailored to the intended form of analysis.
The course concludes with a hands-on practical session using the MICE package in R, where participants will work through a provided real example. A step-by-step guide will be provided, enabling participants to confidently specify and perform the imputation task.
By the end of the course, participants will - Understand the unique challenges of imputing missing data in multilevel settings. - Identify the strengths and limitations of FCS imputation. - Apply multilevel imputation methods to their own datasets using R.
Prerequisites:
Participants should have a general understanding of multilevel modelling, such as familiarity with concepts like random intercept and random slope models. Additionally, a basic working knowledge of R is required. Prior experience with multiple imputation is an advantage but not mandatory. Participants need to bring their own laptop.
Keywords: statistical modelling; Bayesian; R;
biostatistics
Duration: half-day (afternoon; 13:30-17:00)
Course type: hands-on course
Instructors:
Interpreting clinical data with applied statistical models is crucial to inform drug development. However, since clinical data comes in diverse forms and presents various statistical challenges, building models requires a lot of flexibility regarding the statistical model being applied. The course introduces the R package brms, which addresses the needs of applied modeling. brms is short for “Bayesian regression models using Stan” and uses as backend the state-of-the art MCMC sampler Stan. Participants will learn to apply brms through a range of case studies from drug development, such as the use of historical control data in clinical trials, dose-finding, longitudinal continuous endpoint analysis with MMRM, and time-to-event analysis. These examples will be addressed within a Bayesian framework, including a brief introduction to setting up priors. Hands-on exercises will complement the case studies, allowing participants to apply their learning. The course is designed for statisticians interested in applied modeling. Additional in-depth material for self-study is available on the course website.
Learning objectives:
Course participants will learn to apply the brms R package to a broad range of applied modeling problems. The goal is to demonstrate how brms can be used as a general modeling tool, capable of developing complex models for various applied statistical problems in drug development. Participants will gain skills in building hierarchical random-effects models, fitting non-linear models, handling missing data, and analyzing continuous endpoints and time-to-event data using MMRM. These techniques will be taught in the context of real-world drug development problems and will be supported by model checking, comparison, and evaluation practices. Participants will also have the opportunity to explore further details and more advanced topics, such as high-performance computing with brms on large computer clusters, through the accompanying online resources. By the end of the course, participants will have practical, hands-on experience with brms, enhancing their capability to address future modeling challenges with this versatile tool.
Prerequisites:
While familiarity with R is recommended, no prior knowledge of Bayesian statistics is required for this half-day course. Participants need to bring their own laptop with R and cmdstan pre-installed (detailed installation instructions will be provided).