Highlights from the PSI Conference

Posted by Cytel

May 26, 2016 11:37:20 AM

A number of the Cytel team were in Berlin 22nd- 25th May  for the PSI Annual conference.  The PSI Conference is an important forum for statisticians and we found this year's event to be energetic, technically interesting and well attended.  In this blog, we’ll summarise some of the particular highlights from the sessions that our team members attended.

Plenary Session:  Dr Paul Wicks,  VP Innovation at PatientsLikeMe.

The plenary session on day one was very thought provoking, bringing an important perspective on how the landscape of clinical development is transforming and patients are becoming more central and influential. Paul Wicks set the scene by describing how patients have traditionally been viewed as passive subjects or datapoints in clinical trials.  That relationship is now changing, and there are many ways in which patient input can be solicited, including social media listening, and patient reported outcomes.  Wicks discussed the continuum of patient involvement from being treated as passive subjects, through consultation right through to patient power.  At the top of this scale, Wicks highlighted the #WeAreNotWaiting movement in which Type 1 diabetes sufferers are developing platforms, and ‘hacking’ existing products with the aim of improving outcomes.    Clearly, issues emerge from the interaction between this changing profile of the patient, and the regulatory and clinical trial environment which need to be addressed. 

Modelling and Simulation Practice and Best Practice

This session was intended to expose the audience to upcoming guidance on what is good practice with regard to modelling and simulation.  Aspects covered during this session included how to prepare simulations, what scenaria to explore, how to check the simulations and report the outcomes of the models.  There was a discussion where the audience ranked the most important factors, and  unsurprisingly these included:

  • Stating objectives
  • Justifying the scenaria you are selecting
  • Being able to communicate the data and results in graphical manner

There will be a document/ paper on this topic upcoming in the Pharmaceutical Statistics journal which we’ll look forward to seeing.

To Proceed or Not to Proceed? Decision Criteria in Clinical development

This session was of particular interest for Cytel- the development of a consistent approach to quantitative decision-making means that studies are designed from the outset with the decision in mind.  In one of the presentations in the session, Paul Frewer presented on the early phase decision-making framework created within the early phase group of Astrazeneca for which Cytel is supporting a custom software interface.  20160523_152426_resized-1.jpg

In brief, the approach, (based on a methodology proposed by Lalonde et al) was designed to enable the organisation to take forward the best compounds within a candidate rich, early phase portfolio.  We describe a previous talk by Paul at the East User Group Meeting more fully in another blog here.

Breaking new ( Statistical Ground) RSS/PSI Prize Winner 2015

In the first talk, Mark Whitlock of Pfizer presented Katrina Gore’s work on the Assay Capability Tool (ACT) which was designed to guide the development of drug discovery assays and to address issues of robustness and reproducibility.  The team have implemented this standard process across the group, bringing statistical principles into this early stage research.

Of particular interest to Yannis Jemiai, VP at Cytel was the GSK work presented by Nicky Best on prior elicitation in which they have been able to take qualitative information about the chances of success and turn them into formal prior distributions.

 Creative Designs: Time for Something a Little Bit Different

Cyrus MehtaCyrusPSIMAMS.jpg, Cytel co-founder and President participated in one of these sessions, sharing his insights about multi-arm multi-stage designs and their implementation in the new MAMS module in East 6.4.  In the presentation he discusses two approaches- one the generalization of the group sequential approach to more than two arms, and the other alternative to the combination function approach of Posch et. al. (SiM, 2005). To view a video of the presentation which Cyrus gave on multi-arm multi-stage designs at the East User Group meeting, click here

Poster session

For Pantelis Vlachos, Director of Consulting at Cytel the Poster session was a highlight, and he found the below posters of particular interest.

Designing Multi-Arm Multi-Stage Trials with the R Package ‘MAMS’ Pallman, Jaki, Magirr ( Lancaster University). 

Estimating Marginal Distributions at Endpoint when only Baseline and Relative Improvement is Available: A 2 dimensional probability model for the psoriasis area and severity index. Petto, Schacht (Eli Lilly)

There were of course many other sessions which our team found very engaging including presentations on estimands and subgroup analyses, and of course we couldn't attend each and every session.  All in all it was an interesting and successful conference which addressed many of the key issues which statisticians face working in the pharmaceutical industry today. 


Topics: Clinical Data, biostatistics, adaptive designs, adaptive trials, PSI Conference

The Cytel blog keeps you up to speed with the latest developments in biostatistics and clinical biometrics.  Sign up for updates direct to your inbox. You can unsubscribe at any time.


Posts by Topic

see all

Recent Posts