Video-on-Demand Library


Collection
Keyword or phrase

Topic

EFSPI/PSI Causal Inference SIG Webinar: Instrumental Variable Methods

30 September 2025

In recent years, instrumental variable (IV) methods are being increasingly used by pharmaceutical companies in the process of drug development. For example, genetics-based IV methodology (a.k.a Mendelian randomisation) is used extensively in R&D departments to evaluate the promise of potential drug targets through the combined analysis of observational cohort and genome-wide association study data. At the other end of the drug development cycle, IV approaches are emerging as a key tool for quantifying treatment effects in pivotal clinical trials affected by intercurrent events, as part of the Estimand Framework. The webinar is targeted at statisticians working in the pharmaceutical industry, and the objective is to 1) provide a basic understanding of IV methodology including how it relates to causal inference, and 2) present two inspirational pharma-relevant applications.

Read more...

A natural experiment is an observational study that enables causal conclusions to be drawn. In a natural experiment, individuals are divided either at random or in a way that mimics randomization (known as quasi-randomization), but this division is performed by a natural process or an artificial distinction rather than by an investigator. The variable that divides individuals is known as an instrumental variable.  In this talk, I will provide an introduction to instrumental variables: what they are, and how they can be used to make causal claims. I will then present Mendelian randomization, the use of genetic variants as instrumental variables. Due to inherent randomness in the process of genetic inheritance, genetic variants act somewhat like randomization. As the majority of drugs influence proteins and genes encode proteins, there is a specific relevance for Mendelian randomization to the drug development pipeline. Mendelian randomization can provide evidence on target validation, re-purposing, safety signals, mechanistic relevance, and effect heterogeneity: are pathways worth drugging, what outcomes do they affect, how do they affect them, and who would most benefit from intervention? The talk will be illustrated with examples of targets for existing and emerging drugs.

Speakers: Jack Bowden, Novo Nordisk and University of Exeter and Stephen Burgess, University of Cambridge

AI agents for drug discovery and development

23 September 2025

AI agents are large language models equipped with tools that can autonomously tackle challenging tasks. This talk will explore how generative AI agents can enable biomedical discovery. I’ll first introduce the Virtual Lab—a collaborative team of AI scientist agents conducting in silico research meetings to tackle open-ended R&D projects. As an example application, the Virtual Lab designed new nanobody binders to recent Covid variants that we experimentally validated. Then I will present CellVoyager, a computational biology agent that analyzes complex genomics data to derive new insights.

Read more...

AI agents are large language models equipped with tools that can autonomously tackle challenging tasks. This talk will explore how generative AI agents can enable biomedical discovery. I’ll first introduce the Virtual Lab—a collaborative team of AI scientist agents conducting in silico research meetings to tackle open-ended R&D projects. As an example application, the Virtual Lab designed new nanobody binders to recent Covid variants that we experimentally validated. Then I will present CellVoyager, a computational biology agent that analyzes complex genomics data to derive new insights. 

Speaker: James Zou, Stanford University

PSI VisSIG Wonderful Wednesday 66: Fatty liver disease

10 September 2025

In a recent publication of a clinical trial a pair of figures has been attached, that invites a discussion on graphic design principles. This discussion is moderated by Bodo Kirsch. An interactive application to generate visualisations and the respective R code is available via the Wonderful Wednesday blog.

Read more...

The quality and effectiveness of figures in publications could easily be improved by small changes in use of colour, proximity, and space. An actual objective should be formulated for each visualisation. Always ask why! The next challenge is to explore possible subgroup effects in the current data example. See the Wonderful Wednesday homepage for more detail.

Wonderful Wednesdays are brought to you by the Visualisation SIG. The Wonderful Wednesday team includes Bodo Kirsch, Zachary Skrivanek, Lorenz Uhlmann, Steve Mallett, Rhys Warham, Mark Baillie, Paolo Eusebi, Martin Brown, Benjamin Lang, Lovemore Gakava, Aditeya Pandey

PSI VisSIG Wonderful Wednesday 65: Intercurrent events

13 August 2025

In clinical trials, intercurrent events are post-baseline occurrences that affect the interpretation or existence of outcome data. Rhys Warham is presenting visualisations that help understand the possible impact of these events on the interpretation of the study results. All visualisations are available on the Wonderful Wednesday blog.

Read more...

The side-by-side display of the treatment comparison using different approaches of handling intercurrent events is a powerful presentation of the robustness of the results. Anker the display at the intercurrent event can help understand the data, but special care should be taken on possible bias before drawing conclusions. The next challenge is to improve an existing plot from a recent publication of study results. See the Wonderful Wednesday homepage for more detail.

Wonderful Wednesdays are brought to you by the Visualisation SIG. The Wonderful Wednesday team includes Bodo Kirsch, Zachary Skrivanek, Lorenz Uhlmann, Steve Mallett, Rhys Warham, Mark Baillie, Paolo Eusebi, Martin Brown, Benjamin Lang, Lovemore Gakava, Aditeya Pandey



PSI VisSIG Wonderful Wednesday 64: Hyperkalemia Trail (continued)

09 July 2025

This month’s webinar builds on the topic from last month (improving an existing plot from a recent publication on a hyperkalemia trial), but focuses on visualising patient-level data (generated with AI) rather than group-level summaries. Rhys Warham leads the panel discussion. All visualisations are available on the Wonderful Wednesday blog.

Read more...

Several visualisations were discussed, all forming part of a wider submission. The discussion highlights how a series of data visualisations can be used to investigate and identify key relationships in the data, and then to subsequently highlight these key relationships when communicating findings. The challenge for the next webinar was set, which will be to use visualisations to explore how the handling of intercurrent events might influence the comparative efficacy of two treatments. See the Wonderful Wednesday homepage for more detail.

Wonderful Wednesdays are brought to you by the Visualisation SIG. The Wonderful Wednesday team includes Bodo Kirsch, Zachary Skrivanek, Lorenz Uhlmann, Steve Mallett, Rhys Warham, Mark Baillie, Paolo Eusebi, Martin Brown, Benjamin Lang, Lovemore Gakava and Aditeya Pandey

PSI Vaccines SIG Webinar: Challenges in Estimating Vaccine Effectiveness Against Progression to Severe Disease

07 July 2025

This work highlights methodological challenges in isolating a vaccine's effect on progression to severe disease after infection.

Read more...

This work highlights methodological challenges in isolating a vaccine's effect on progression to severe disease after infection. Vaccines can reduce an individual's risk of infection and their risk of progression to disease given infection. The latter effect is less commonly estimated but is relevant for risk communication and vaccine impact modeling. Using a motivating example from the COVID-19 literature, we note how vaccine effectiveness against progression can appear to increase over time in settings where true biological strengthening is unlikely. We use mathematical modeling to demonstrate how this phenomenon can occur when there is an underlying vulnerable subpopulation with poor vaccine response against infection and progression. We describe a modeling framework to link underlying immunology and post-vaccination outcomes that we use to further examine this problem.

Pre-Clinical SIG Webinar: Methodology and first results of the iRISE consortium

25 June 2025

The iRISE (improving Reproducibility In SciencE) project aims to deepen our understanding of the drivers contributing to poor reproducibility and to conduct a detailed evaluation – including primary research – of the effectiveness of interventions to increase reproducibility.

Read more...

The iRISE (improving Reproducibility In SciencE) project aims to deepen our understanding of the drivers contributing to poor reproducibility and to conduct a detailed evaluation – including primary research – of the effectiveness of interventions to increase reproducibility. The project’s work package (WP) on theory is dedicated to clarifying terms and underlying concepts related to reproducibility. To ensure efficient communication within the iRISE consortium, we produced a glossary with working definitions of terms including reproducibility, replicability, and replication.  A concept important in pre-clinical drug discovery, translatability, was for example defined as "the ability to apply research discoveries from experimental models to applications that directly benefit humans".  The definitions are complemented with measures to quantify different types of reproducibility, and theories on the possible factors leading to irreproducibility.

To view the presented slides, please click here.

PSI VisSIG Wonderful Wednesday 63: Hyperkalemia Trail

11 June 2025

This month it’s all about improving an existing plot from a recent publication on a hyperkalemia trial. Bodo Kirsch is leading the panel discussion on dos and don’ts of effective visualisations. All visualisations are available on the Wonderful Wednesday blog.

Read more...

The discussion touched various issues in scientific data visualisation, like messaging, axis limits, background grids and shading, interpolation, choice of legend, reference lines and explanatory text. This was followed by a demonstration of instant patient level data simulation using AI. The next challenge is to include these patient level data into the graphical display of the hyperkalemia trial results. See the Wonderful Wednesday homepage for more detail.

Wonderful Wednesdays are brought to you by the Visualisation SIG. The Wonderful Wednesday team includes Bodo Kirsch, Zachary Skrivanek, Lorenz Uhlmann, Steve Mallett, Rhys Warham, Mark Baillie, Paolo Eusebi, Martin Brown, Benjamin Lang, Lovemore Gakava and Aditeya Pandey

PSI Conference 2025: Leadership TED

11 June 2025

Chair: Kate Taylor How to be wrong (T006) - Simon Cleall Stepping into leadership: How will I manage? (T002) - Catherine Dixon Enhancing Cross-functional Partnership in Early Oncology Clinical Development: A Practical Guide for Biostatisticians (T008) - Laura Barker Trust actually: Building teams that love to work together (T007) - Zainab Walsh Building High-Performing Teams: Leadership Strategies for Navigating Change and Driving Growth(T010) - Aga Rasinska Trust: The Backbone of Leadership (T005) - Alun Bedding

Read more...

Chair: Kate Taylor
How to be wrong (T006) - Simon Cleall
Stepping into leadership: How will I manage? (T002) - Catherine Dixon
Enhancing Cross-functional Partnership in Early Oncology Clinical Development: A Practical Guide for Biostatisticians (T008) - Laura Barker
Trust actually: Building teams that love to work together (T007) - Zainab Walsh

Building High-Performing Teams: Leadership Strategies for Navigating Change and Driving Growth(T010) - Aga Rasinska
Trust: The Backbone of Leadership (T005) - Alun Bedding

PSI Conference 2025: Use of external data to improve clinical trials

11 June 2025

Chair: Jyoti Soni Steps in using healthcare systems data as outcome data in clinical trials (O024) - Sharon Love Why Accurate Time to response prediction matters? (O026) - Donia Skanji Survival of the Fittest: Digitising Survival Data for Enhanced Decision-Making in Clinical Trials (O044) - James Sykes and Nelson Kinnersley

Read more...

Chair: Jyoti Soni
Steps in using healthcare systems data as outcome data in clinical trials (O024) - Sharon Love
Why Accurate Time to response prediction matters? (O026) - Donia Skanji
Survival of the Fittest: Digitising Survival Data for Enhanced Decision-Making in Clinical Trials (O044) - James Sykes and Nelson Kinnersley

PSI Conference 2025: Bayesian Dynamic Borrowing

11 June 2025

Chair: Julia Saperia Unexpected results and challenges when using mixture priors for Bayesian borrowing (O031) - Darren Scott Non-monotonic power in Bayesian dynamic borrowing: insights and practical remedies (O035) - Gianmarco Caruso Biased borrowing or borrowing bias? Leveraging Bayesian borrowing and quantitative bias analysis for robust comparative effectiveness insights (O049) - Grace Hsu

Read more...

Chair: Julia Saperia
Unexpected results and challenges when using mixture priors for Bayesian borrowing (O031) - Darren Scott
Non-monotonic power in Bayesian dynamic borrowing: insights and practical remedies (O035) - Gianmarco Caruso
Biased borrowing or borrowing bias? Leveraging Bayesian borrowing and quantitative bias analysis for robust comparative effectiveness insights (O049) - Grace Hsu

PSI Conference 2025: Marginal Estimands and Estimation with Covariate Adjustment for TTE Endpoints

11 June 2025

Chair: Sarwar Mozumder Session Introduction - David Wright and Sarwar Mozumder Marginal hazard ratios and covariate adjustment – A causal inference perspective - Rhian Daniel Efficiency of nonparametric superiority tests based on restricted mean survival time versus the log-rank test under proportional hazards - Dominic Magirr Covariate adjustment in time-toevent data: single and doubly-robust methods - Sanne Roels Ensuring covariate adjustment methods for TTE outcomes are fit for use - Tim Morris Discussion Panel: Thoughts from A Regulator’s Perspective – What are the Expectations? Armin Koch, David Wright, Rhian Daniel, Dominic Magirr, Sanne Roels and Tim Morris

Read more...

Chair: Sarwar Mozumder
Session Introduction - David Wright and Sarwar Mozumder
Marginal hazard ratios and covariate adjustment – A causal inference perspective - Rhian Daniel
Efficiency of nonparametric superiority tests based on restricted mean survival time versus the log-rank test under proportional hazards - Dominic Magirr
Covariate adjustment in time-toevent data: single and doubly-robust methods - Sanne Roels
Ensuring covariate adjustment methods for TTE outcomes are fit for use - Tim Morris Discussion Panel: Thoughts from A Regulator’s Perspective – What are the Expectations?
Armin Koch, David Wright, Rhian Daniel, Dominic Magirr, Sanne Roels and Tim Morris
Page:

Upcoming Events

PSI Book Club - The Art of Explanation: How to Communicate with Clarity and Confidence

Develop your non-technical skills by reading The Art of Explanation by Ros Atkins and joining the Sept-Dec 2025 book club. You will be invited to join facilitated discussions of the concepts and ideas and apply skills from the book in-between sessions.

Joint PSI/EFSPI Visualisation SIG 'Wonderful Wednesday' Webinars

Our monthly webinar explores examples of innovative data visualisations relevant to our day to day work. Each month a new dataset is provided from a clinical trial or other relevant example, and participants are invited to submit a graphic that communicates interesting and relevant characteristics of the data.

Open-Source Lunch Bites

Our monthly webinar will allow attendees to gain practical knowledge and skills in Open-Source coding and tools, with a focus on applications in the pharmaceutical industry. The sessions will provide starting points in a number of areas, correct any common misconceptions and provide valuable resources for further learning.

PSI Training Course: Introduction to Machine Learning

Four sessions will include ML foundation (including an introduction, data exploration for ML and dimensionality reduction and feature selection), Supervised learning (including support vector machines and model evaluation and interpretation), model optimization and unsupervised learning (including clustering) and advanced topics (including neural networks, deep learning and large language models).

Stability Modelling: New Tools and Approaches

Webinar dedicated to the topic of stability modelling; opportunity to hear from statisticians working in the CMC field, with open question and answers. Brought to you by the CMC Statistical Network Europe (CSNE) SIG.