Determining the appropriate sample size is an important part of good clinical trial design. When there is uncertainty about some of the design parameters (e.g. variability, control rate, model parameters), it can be challenging to determine up front the number of subjects required for robust evaluation of the study objectives. The aim of this PSI one day meeting is to present an overview of available methods for sample size re-estimation together with several case studies where such methods have been used in late phase clinical trials. There will be plenty of opportunity for discussion and interaction with other statisticians working in this area. Registration is now closed. Please contact the secretariat if you wish to inquire.
There are two distinct reasons for sample size re-estimation in clinical trials: the first is to maintain power when trial data indicate the response variance has been under-estimated; the second is to adapt to interim estimates of the treatment effect. I shall explain how a combination test can be used to ensure rigorous protection of the type I error rate when sample size is adapted in the light of observed data. I shall describe methods for increasing sample size to maintain power when response variance is higher than expected, based on either blinded or unblended variance estimates. I shall discuss the relationship between trial designs that adjust sample size in response to the estimated treatment effect and group sequential designs, which start with a higher maximum sample size but stop early when the data support such a decision. In particular, I shall describe the “promising zone” approach of Mehta and Pocock (Statistics in Medicine, 2011) and show how to modify this procedure in order reduce the average sample size and achieve similar performance to an efficient group sequential design.
Clinical Trials Consulting &Training Ltd
Sample Size Re-Estimation – Some random observations
I was an author on one of the very early papers on sample size re-estimation (Birkett and Day, Stats in Med, 1994; 13: 2455–2463) and have since followed the field with much interest, some despair, and more than a little exasperation.This talk will illustrate some of these facets – mostly based around such methods used in a regulatory context. Several personal experiences will be included (particularly the ones that went wrong) as well as some of the approaches and myths I see in my regular consulting work. What’s “allowed” and what’s not? What makes sense and what doesn’t?
Blinded sample-size re-estimation in multiple sclerosis clinical trials
Multiple sclerosis (MS) is a progressive, degenerative disease. MS is the most common disorder of the CNS in adults, affecting up to 2.5 million people worldwide. Clinical trials in MS use count, recurrent event and time-to-event primary endpoints. Methodology for blinded sample size re-estimation with such endpoints is briefly reviewed. A case study illustrates how to implement blinded sample size re-estimation in a confirmatory MS trial.
Mike Greenwood, AstraZeneca
Do we need more patients? Your statistics should be correct. Make sure you communicate effectively!
We performed a blinded estimation of the pooled exacerbation rate and shape parameter from a negative binomial model in a COPD exacerbation study. This talk will briefly cover the statistics, the practical aspects and focus on the importance of clear (and understandable to non-statisticians) communication of the results and their implications
Nikhil Chauhan, BTG International Ltd
An experience in implementing the Promising Zone sample size re-estimation methodology (Mehta and Pocock, 2011) in a phase 3 oncology study
I will share my experience in implementing the Promising Zone sample size re-estimation methodology (Mehta and Pocock, 2011) in a phase 3 oncology study for the purposes of obtaining a US FDA marketing approval. I will outline how we decided on using this study design, how the sample size calculation was performed, and how the promising zone boundaries were set. I will also share our experience in demonstrating the acceptability of the study design to FDA, in terms of showing control of Type I error. Reference: Mehta, CR and Pocock, SJ (2011), Adaptive increase in sample size when interim results are promising: A practical guide with examples. Statist. Med., 30: 3267–3284. doi: 10.1002/sim.4102
Blinded sample size re-estimation in a Phase III study investigating Progression Free Survival
The CLARINET study investigated the effect of lanreotide compared to placebo in the treatment of metastatic enteropancreatic neuroendocrine tumours with progression free survival as primary endpoint. We will describe the design of the study, the justification for the blinded sample-size re-estimation and some practical aspects of carrying out that decision, including communication within the company, with the DSMB and with regulatory authorities.
Sample Size Re-estimation : « De-risking » a crucial stage of clinical development
Developing cancer treatments is a high-stakes endeavor – especially for emerging biotechs and specialty pharmas with limited portfolios. Conventional phase 3 trial designs are “all or nothing” propositions and well over 50% of these pivotal studies end in failure. Unlike conventional studies, adaptive approaches allow beneficial design changes following interim analysis (IA). A Sample Size Re-estimation design allows selection of the strategy most likely to succeed. We present a case study of such an adaptive approach used to both effectively “de-risk” the final clinical stage as well as be accepted by FDA reviewers based on the concept of the “Promising Zone”. Rather than committing to a larger sample size up front, the decision is deferred until the clinical evidence justifies cost of added subjects. The strategy provided the confidence company leaders – and investors – needed to launch the final development effort toward approval.
Early Bird Rate (until 14th October 2016)
After 14th October 2016
£120 + VAT
£160 + VAT
£160 + VAT
£220 + VAT
£60 + VAT
£90 + VAT
Registration closes on 26th October 2016
Please contact the PSI secretariat on firstname.lastname@example.org if you have any queries.
Registration is now closed. Please contact the secretariat if you wish to inquire.
Joint PSI, EFSPI & ASA BIOP Webinar: Estimands
PSI, the European Federation of Statisticians in the Pharmaceutical Industry (EFSPI) and the Biopharmaceutical Section of the American Statistical Association (ASA) are jointly organising a webinar on Estimands in Practice. Speakers from regulatory authorities (FDA and EMA) and industry will present on their experience on this topic to date.
Discover your potential by working as a Covance Senior Principal / Principal Biostatistician. You'll enjoy a varied role working for different sponsors across several therapeutic areas, leading projects and mentoring junior members of the department.
Biometrics at AstraZeneca provides the data that influences decisions on how we roll back the frontiers of science to bring life-changing medicines to the world. We are at the heart of design, analytics and interpretation of AstraZeneca’s portfolio.