BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//talks.cam.ac.uk//v3//EN
BEGIN:VTIMEZONE
TZID:Europe/London
BEGIN:DAYLIGHT
TZOFFSETFROM:+0000
TZOFFSETTO:+0100
TZNAME:BST
DTSTART:19700329T010000
RRULE:FREQ=YEARLY;BYMONTH=3;BYDAY=-1SU
END:DAYLIGHT
BEGIN:STANDARD
TZOFFSETFROM:+0100
TZOFFSETTO:+0000
TZNAME:GMT
DTSTART:19701025T020000
RRULE:FREQ=YEARLY;BYMONTH=10;BYDAY=-1SU
END:STANDARD
END:VTIMEZONE
BEGIN:VEVENT
CATEGORIES:Cambridge Statistics Discussion Group (CSDG)
SUMMARY:Bayesian and structural integration of background 
 evidence in the design\, analysis and interpretati
 on of clinical trial data - Fabio Rigat (GSK) &amp
 \; Nicky Best (Imperial College)
DTSTART;TZID=Europe/London:20181120T191500
DTEND;TZID=Europe/London:20181120T213000
UID:TALK95869AThttp://talks.cam.ac.uk
URL:http://talks.cam.ac.uk/talk/index/95869
DESCRIPTION:Although background evidence always informs clinic
 al study design\, individual trial data are often 
 first analysed in isolation and then meta-analysis
  is used to synthesise evidence accrued across mul
 tiple studies. However\, in “small population” set
 tings such as rare diseases\, paediatric populatio
 ns\, or personalised/stratified medicine\, the fea
 sible sample size that can be recruited in each in
 dividual study typically falls short of that neede
 d for a conventionally-powered trial. To address t
 hese limitations\, we assess two avenues for integ
 rated analysis of clinical data. First\, we discus
 s the use of Bayesian informative priors based on 
 relevant external data to increase the power and p
 recision of such trials and propose a range of ope
 rating characteristics that can be useful to evalu
 ate and compare designs.  Here we present various 
 prior and posterior summaries of the available his
 torical\, current and total evidence which can be 
 used to help sponsors and regulators assess the st
 rength of the prior assumptions and the extent to 
 which the current trial data can influence the fin
 al posterior inference. We illustrate these method
 s through some recent case studies covering both p
 aediatric settings borrowing efficacy data from ad
 ults\, and in confirmatory trials in adults borrow
 ing historical controls. Second\, we explore the u
 se of genetic annotation for design and analysis o
 f biomarker-rich “large p\, small n” early phase o
 ncology clinical trials. While the primary goal of
  these trials is to assess safety and preliminary 
 clinical efficacy in a possibly heterogeneous pati
 ent population\, a key secondary objective is to i
 dentify prognostic and predictive markers for desi
 gn adaptation and for developing companion diagnos
 tics. Here we focus on the analysis of association
 s between baseline gene expression and response to
  therapy\, comparing the operational characteristi
 cs of bottom-up and top-down methods. While bottom
 -up methods use gene annotations for interpretatio
 n of the analysis results\, top-down methods incor
 porate these background data within the structure 
 of a hierarchical model likelihood for estimating 
 pathway-level associations to clinical response. W
 e show that\, while top-down methods benefit from 
 the advantages of parsimonious parametric modellin
 g\, they are not less exposed than bottom-up metho
 ds to a reliance on a knowledge basis that is both
  incomplete and in constant evolution.
LOCATION:Statistical Laboratory\, Centre for Mathematical S
 ciences\, Wilberforce Road\, Cambridge\, CB3 0WB
CONTACT:Peter Watson
END:VEVENT
END:VCALENDAR