BEGIN:VCALENDAR VERSION:2.0 METHOD:PUBLISH PRODID:-//Telerik Inc.//Sitefinity CMS 15.4//EN BEGIN:VTIMEZONE TZID:GMT Standard Time BEGIN:STANDARD DTSTART:20251002T020000 RRULE:FREQ=YEARLY;BYDAY=-1SU;BYHOUR=2;BYMINUTE=0;BYMONTH=10 TZNAME:GMT Standard Time TZOFFSETFROM:+0100 TZOFFSETTO:+0000 END:STANDARD BEGIN:DAYLIGHT DTSTART:20250301T010000 RRULE:FREQ=YEARLY;BYDAY=-1SU;BYHOUR=1;BYMINUTE=0;BYMONTH=3 TZNAME:GMT Daylight Time TZOFFSETFROM:+0000 TZOFFSETTO:+0100 END:DAYLIGHT END:VTIMEZONE BEGIN:VEVENT DESCRIPTION:Date: Tuesday 30th June 2026Time: 13:00 - 15:00 BST | 15:00 - 1 7:00 CESTSpeakers: \;Dominic Magirr (Novartis Pharma AG)\, \;Sanne Roels (Johnson &\; Johnson)\, \;Kelly Van Lancker (Ghent Universit y and Vrije Universiteit Brussel) and \;Jurgen Hummel (Cytel).Who is t his event intended for? Statisticians active in clinical trials.What is th e benefit of attending? Increased understanding and insights in methodolog y\, regulatory landscape\, and use for covariate adjustment in clinical tr ials.OverviewThis webinar provides a comprehensive overview of covariate a djustment in clinical trials\, covering both regulatory foundations and re cent methodological developments. The session opens with an introduction t o the potential gains from covariate adjustment and discuss key recommenda tions from the 2023 FDA guidance. This includes considerations for both li near and non-linear models\, as well as areas where further research may h elp refine best practices for registrational trials.Building on this found ation\, Dominic Magirr (Novartis) will review well-accepted methods for co variate adjustment\, including standardization (g-computation) using gener alized linear models\, and covariate-adjusted extensions of the log-rank t est with corresponding hazard ratio estimation. The presentation will also highlight the distinction between marginal and conditional estimands and discuss the potential role of prognostic risk scores or &ldquo\;supercovar iates.&rdquo\;The webinar then moves to recent methodological developments beyond current standard practice. Sanne Roels (Johnson &\; Johnson) wi ll explore extensions such as covariate adjustment in group sequential des igns\, with particular attention to type I error control\, and discuss the move toward data-adaptive approaches\, including pre-specified strategies such as Targeted Minimum Loss-Based Estimation (TMLE) across common endpo int types.Looking ahead\, Kelly Van Lancker (Ghent University and Vrije Un iversiteit Brussel) will discuss promising future directions\, including t he use of data-adaptive and machine learning&ndash\;based estimators such as TMLE and related doubly robust methods. The talk will highlight key cha llenges around interpretability\, pre-specification\, and regulatory accep tability\, with particular attention to small sample settings and complex data structures such as clustered or multi-center trials. Practical consid erations for balancing innovation with robustness\, transparency\, and tru st in confirmatory analyses will also be discussed.The session concludes w ith a panel discussion led by Jü\;rgen Hummel (Cytel)\, bringing toget her regulatory\, industry\, and academic perspectives to reflect on curren t practice and future directions in covariate adjustment. \;Registrati onThis event is free to attend for both Members of PSI and Non-Members. To register your place\, please click here. \;Speaker DetailsSpeakerBiog raphyAbstractDominic Magirr\, NovartisDominic is part of the Advanced Meth odology and Data Science group at Novartis\, where he provides methodologi cal and technical support to clinical trial teams across a wide variety of statistical topics. \;In this presentation\, I will discuss well acce pted methods for covariate adjustment\, including standardization (or g-co mputation) using generalized linear models\, as well as a covariate-adjust ed version of the log-rank test with a corresponding method for hazard rat io estimation. I will cover the distinction between marginal and condition al estimands and discuss the potential role of risk scores or &ldquo\;supe rcovariates&rdquo\;. \;Sanne Roels\, Johnson &\; Johnson \;Sann e is part of the statistical modelling\, methodology and consulting group at J&\;J. The group supports teams through implementation of statistica l innovation and impactful methodology\, leveraging modelling and simulati on.Sanne founded and continues to co-lead PSI/EFSPI Working group on Causa l Inference. \;In this talk\, I will discuss methodological developmen ts that go beyond what is currently generally accepted. I will discuss cov ariate adjustment in group sequential designs and related concerns around type I error control. Next\, I will discuss the considerations of moving t oward data‑adaptive methods\, including pre‑specified data‑adaptive strate gies such as TMLE across common endpoint types. \; \; \;Kelly Van Lancker\, \;Ghent University and Vrije Universiteit Brussel \; Kelly Van Lancker is an assistant professor in biostatistics at Ghent Univ ersity and Vrije Universiteit Brussels. She received both her master degre e in mathematics and her PhD degree in Statistical Data Analysis from Ghen t University. \; Previously\, Kelly was a postdoctoral researcher at t he Johns Hopkins Bloomberg School of Public Health. Her goal is to develop innovative designs and analytical techniques \;for drawing causal inf erences in health sciences. A big part of her research focuses on more acc urate and faster decision-making in randomized clinical trials by making o ptimal use of the available data. \; \;This talk will discuss prom ising future directions and highlight key pitfalls and open problems. Thes e include the use of pre‑specified data‑adaptive and machine‑learning&ndas h\;based estimators such as TMLE and related doubly robust methods. While these approaches offer efficiency gains\, they raise practical challenges around interpretability\, pre‑specification\, and regulatory acceptability . Particular attention will be paid to small‑sample settings\, where asymp totic guarantees may be unreliable. The talk will also address clustered a nd correlated data structures\, common in multi‑center trials\, and their implications for covariate adjustment. The session will conclude with prac tical considerations on \;balancing methodological innovation with rob ustness\, transparency\, and trust in confirmatory analyses. \;Jurgen Hummel\, Cytel \;Jü\;rgen Hummel is Vice President\, Innovative St atistics at Cytel\, and in that role he provides statistical consultancy t o integrate advanced statistical approaches into development programs.&nbs p\; He has been working in Biostatistics in the CRO\, pharmaceutical and h ealth care industry for more than 30 years in various project related\, te chnical and managerial positions. \; Prior to joining Cytel\, Jü\; rgen led the Statistical Methodology groups at PPD (now Thermo Fisher Scie ntific) and at Novo Nordisk. \;Jü\;rgen is a member of the EFSPI S tatistical Methods Leaders Group\, led the PSI/EFSPI Regulatory Special In terest Group for 5 years and served on the PSI Board of Directors. \; He earned the German equivalent of an MSc in mathematics and economics at Augsburg University\, and he is a Chartered Statistician with the Royal St atistical Society.\n  \;Panel Discussion Lead DTEND:20260630T140000Z DTSTAMP:20260521T010005Z DTSTART:20260630T120000Z LOCATION: SEQUENCE:0 SUMMARY:Causal Inference SIG and EFSPI Methods Leaders: Modern Covariate Ad justment in Clinical Trials UID:RFCALITEM639149220053181547 X-ALT-DESC;FMTTYPE=text/html:

Date: Tue sday 30th June 2026

Time: 13:00 - 15:00 BST | 15:00 - 17:00 CEST

Speakers: \;Domi nic Magirr (Novartis Pharma AG)\, \;Sanne Roels (Johnson &\; Johnso n)\, \;Kelly Van Lancker (Ghent University and Vrije Universiteit Brus sel) and \;Jurgen Hummel (Cytel).

Who is this event inte nded for? Statisticians active in clinical trials.

What is the benefit of attending? Increased understanding a nd insights in methodology\, regulatory landscape\, and use for covariate adjustment in clinical trials.

Overview

This webinar provide s a comprehensive overview of covariate adjustment in clinical trials\, co vering both regulatory foundations and recent methodological developments. The session opens with an introduction to the potential gains from covari ate adjustment and discuss key recommendations from the 2023 FDA guidance. This includes considerations for both linear and non-linear models\, as w ell as areas where further research may help refine best practices for reg istrational trials.

Building on this foundation\, Dominic Magir r (Novartis) will review well-accepted methods for covariate adjustment\, including standardization (g-computation) using generalized linear models\ , and covariate-adjusted extensions of the log-rank test with correspondin g hazard ratio estimation. The presentation will also highlight the distin ction between marginal and conditional estimands and discuss the potential role of prognostic risk scores or &ldquo\;supercovariates.&rdquo\;
< br />The webinar then moves to recent methodological developments beyond c urrent standard practice. Sanne Roels (Johnson &\; Johnson) will explor e extensions such as covariate adjustment in group sequential designs\, wi th particular attention to type I error control\, and discuss the move tow ard data-adaptive approaches\, including pre-specified strategies such as Targeted Minimum Loss-Based Estimation (TMLE) across common endpoint types .

Looking ahead\, Kelly Van Lancker (Ghent University and Vrije Universiteit Brussel) will discuss promising future directions\, includin g the use of data-adaptive and machine learning&ndash\;based estimators su ch as TMLE and related doubly robust methods. The talk will highlight key challenges around interpretability\, pre-specification\, and regulatory ac ceptability\, with particular attention to small sample settings and compl ex data structures such as clustered or multi-center trials. Practical con siderations for balancing innovation with robustness\, transparency\, and trust in confirmatory analyses will also be discussed.

The sess ion concludes with a panel discussion led by Jü\;rgen Hummel (Cytel)\, bringing together regulatory\, industry\, and academic perspectives to re flect on current practice and future directions in covariate adjustment.

 \;

Registration

This event is free to attend for b oth Members of PSI and Non-Members. To register your place\, please click here.

 \;

Speaker Details

Speaker

Biography

Abstract

Dominic Magirr\, Novartis

Dominic is part of the Advanced Met hodology and Data Science group at Novartis\, where he provides methodolog ical and technical support to clinical trial teams across a wide variety o f statistical topics. \;

In this presentation\, I will discuss well accepted methods for cova riate adjustment\, including standardization (or g-computation) using gene ralized linear models\, as well as a covariate-adjusted version of the log -rank test with a corresponding method for hazard ratio estimation. I will cover the distinction between marginal and conditional estimands and disc uss the potential role of risk scores or &ldquo\;supercovariates&rdquo\;.< /td>

 \;

< img src="https://psiweb.org/images/default-source/default-album/sanne-roel s.tmb-thumbnail.png?Culture=en&\;sfvrsn=bf87a9db_1&\;sf_site_temp=tr ue&\;sf_site=00000000-0000-0000-0000-000000000000" style="max-width:100 %\;height:auto\;" width="120" height="120" sf-image-responsive="true" alt= "" title="Sanne Roels" />

Sanne Roels\, Johnson &\; Johnson

 \;

San ne is part of the statistical modelling\, methodology and consulting group at J&\;J. The group supports teams through implementation of statistic al innovation and impactful methodology\, leveraging modelling and simulat ion.

Sanne founded and continues to co-lead PSI/EFSPI Working group on Causal Inference.

 \;

In this talk\, I will discuss methodological developments th at go beyond what is currently generally accepted. I will discuss covariat e adjustment in group sequential designs and related concerns around type I error control. Next\, I will discuss the considerations of moving toward data‑adaptive methods\, including pre‑specified data‑adaptive strategies such as TMLE across common endpoint types. \;

 \;

 \;

Kelly Van Lancker\, \; Ghent University and Vrije Universiteit Brussel

 \;

Kelly Van Lancker i s an assistant professor in biostatistics at Ghent University and Vrije Un iversiteit Brussels. She received both her master degree in mathematics an d her PhD degree in Statistical Data Analysis from Ghent University. \ ; Previously\, Kelly was a postdoctoral researcher at the Johns Hopkins Bl oomberg School of Public Health. Her goal is to develop innovative designs and analytical techniques \;for drawing causal inferences in health s ciences. A big part of her research focuses on more accurate and faster de cision-making in randomized clinical trials by making optimal use of the a vailable data. \;

 \;

This talk will discuss promising future directions a nd highlight key pitfalls and open problems. These include the use of pre‑ specified data‑adaptive and machine‑learning&ndash\;based estimators such as TMLE and related doubly robust methods. While these approaches offer ef ficiency gains\, they raise practical challenges around interpretability\, pre‑specification\, and regulatory acceptability. Particular attention wi ll be paid to small‑sample settings\, where asymptotic guarantees may be u nreliable. The talk will also address clustered and correlated data struct ures\, common in multi‑center trials\, and their implications for covariat e adjustment. The session will conclude with practical considerations on&n bsp\;balancing methodological innovation with robustness\, transparency\, and trust in confirmatory analyses.

 \;

< p>Jurgen Hummel\, Cytel

 \;

Jü\;rgen Hummel is Vice President\, Innovati ve Statistics at Cytel\, and in that role he provides statistical consulta ncy to integrate advanced statistical approaches into development programs . \; He has been working in Biostatistics in the CRO\, pharmaceutical and health care industry for more than 30 years in various project related \, technical and managerial positions. \; Prior to joining Cytel\, J&u uml\;rgen led the Statistical Methodology groups at PPD (now Thermo Fisher Scientific) and at Novo Nordisk. \;

Jü\;rgen is a member of th e EFSPI Statistical Methods Leaders Group\, led the PSI/EFSPI Regulatory S pecial Interest Group for 5 years and served on the PSI Board of Directors . \; He earned the German equivalent of an MSc in mathematics and econ omics at Augsburg University\, and he is a Chartered Statistician with the Royal Statistical Society.\n

 \;

Panel Discussion Lead


END:VEVENT END:VCALENDAR