Dear Allstater,

Please find below information on the forthcoming joint meeting of the
Medical Section and the East Kent RSS local group.

Members wishing to attend should register beforehand with A.P.Grieve, Pfizer
Global R&D, Ramsgate Road, Sandwich, CT13 9NJ, Kent or by email
([log in to unmask]). If you require directions please
request them when registering. On arriving at Pfizer, attendees should
report to the West Site Reception.

Andy Grieve


JOINT MEETING RSS MEDICAL SECTION/EAST KENT LOCAL GROUP


Thursday 28th November, 2pm, Auditorium, West Site, Pfizer Global R&D,
Ramsgate Road, Sandwich, CT13 9NJ (Tea will be available afterwards)

Meeting on Response Adaptive Clinical Trials

CHRIS PALMER (Center for Applied Medical Statistics, University of
Cambridge)
Data-dependent designs in clinical trials: Time to start learning-as-we-go.

After a brief review of the history and ethics of clinical trials, I trace
the recent development of data-dependent designs, meaning those employing
adaptive, Bayesian, decision-theoretic or sequential methods. All of these
are motivated by ethical considerations but possess desirable side effects
of economy, in terms of expected numbers of patients and costs involved. I
discuss whether such trials are also economical with the truth, or if they
provide a suitable framework for answering pragmatic questions of clinical
importance. Given the many present-day pressures on traditional, fixed
sample size clinical trials (recruitment difficulties, growing threat of
lawsuits, rise of patient support groups, etc.) I suggest that, in the right
circumstances, practical application of data-dependent designs may be the
key to alleviating these pressures and could point forward to how more
trials ought to be conducted in this 21st century. After all, why should
today's clinical researchers be limited to statistical methods that were
available long before the computer age?

STEVE COAD (University of Sussex)
Sequential adaptive urn designs with elimination for comparing k*3
treatments.

A fully-sequential procedure is proposed for comparing k*3 treatments with
immediate binary responses. The procedure uses an adaptive urn design to
randomise patients to the treatments and stopping rules are incorporated for
eliminating less promising treatments. Simulation is used to assess the
performance of the procedure for several adaptive urn designs, in terms of
expected numbers of treatment failures and allocation proportions, and the
effect on estimation at the end of the trial is also addressed. It is
concluded that the drop-the-loser rule is more effective than equal
allocation and all of the other designs considered. It is then shown how
the sequential elimination procedure may be used in dose-finding studies and
its performance is compared with a recently proposed method. Several
possible extensions to the work are briefly indicated. (The talk is based
on joint work with Anastasia Ivanova at the University of North Carolina.)


MARGARET JONES (Crathorne Statistical Consultants) & ANDY GRIEVE (Pfizer
Global R&D)
Is adaptive attractive, is flexible feasible? Practical issues in
implementing response adaptive designs in drug development.

Within drug development the use of response-adaptive designs has been
restricted largely to early phase I volunteer studies or to oncology studies
and there have been few attempts to utilize them in phase II or phase III
clinical trials. Why? In this paper we look at some of the issues which make
difficult the practical implementation of these designs, these include: the
need for early access to data; the ability to predict patient outcomes,
either from early data or from surrogate endpoints; the need to interact and
convince regulatory authorities unfamiliar with such designs; the need to
develop computer systems to run such trials and to validate the systems; the
need to integrate the output from the system with existing reporting
standards. We conclude that despite these difficulties such trial designs
are feasible to implement and that they potentially offer a useful
alternative to traditional clinical trial designs.





     LEGAL NOTICE

     Unless expressly stated otherwise, this message is confidential and
     may be privileged. It is intended for the addressee(s) only. Access to
     this e-mail by anyone else is unauthorised. If you are not an
     addressee, any disclosure or copying of the contents of this e-mail or
     any action taken (or not taken) in reliance on it is unauthorised and
     may be unlawful. If you are not an addressee, please inform the sender
     immediately.

     Pfizer Limited is registered in England under No. 526209 with its
     registered office at Ramsgate Road, Sandwich, Kent CT13 9NJ




----------------------------------------------------------------
This message and any attachment has been virus checked by
Pfizer Corporate Information Technology, Sandwich.
----------------------------------------------------------------