BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:Interim Monitoring of Clinical Trials: Decision Theory\, Dynamic P
 rogramming and Optimal Stopping - Chris Jennison\, University of Bath
DTSTART:20121116T160000Z
DTEND:20121116T170000Z
UID:TALK40392@talks.cam.ac.uk
CONTACT:Richard Samworth
DESCRIPTION:It is standard practice to monitor clinical trials with a view
  to\nstopping early if results are sufficiently positive\, or negative\,\n
 at an interim stage. We shall explain how properties of stopping\nboundari
 es can be calculated and how boundaries can be optimised\nto minimise expe
 cted sample size while controlling type I and II\nerror probabilities.\n\n
 Constraints on error probabilities complicate this optimisation\nproblem. 
 However\, a solution is possible through consideration of\nunconstrained B
 ayes decision problems which are conveniently solved\nby dynamic programmi
 ng. This conversion to an unconstrained problem\nis equivalent to using La
 grange multipliers. We shall present details\nof numerical computation for
  group sequential tests and their\noptimisation for particular criteria. W
 e shall discuss a variety of\napplications in clinical trial design includ
 ing the derivation of\noptimal adaptive designs in which future group size
 s are allowed to\ndepend on previously observed responses\; designs which 
 test both for\nsuperiority and non-inferiority\; and group sequential test
 s which\nallow for a delay between treatment and response.\n\nSince optima
 lity in the unconstrained problem can be expressed as a\nsample path prope
 rty\, this is an "optimal stopping" problem in the\nlanguage of probabilit
 y theory. The computational methods we describe\nare\, therefore\, applica
 ble to such problems and\, in particular\, to\noptimal stopping problems a
 rising in financial mathematics.
LOCATION:MR12\, CMS\, Wilberforce Road\, Cambridge\, CB3 0WB
END:VEVENT
END:VCALENDAR