Joint Meeting between RSS East Midlands Local group and the RSS Medical
Section.
Time: Thursday 8th November 2007
Place:
Room C12, Maths & Physics Building, University of Nottingham
from 3pm-6pm with wine and cheese to follow.
Talk 1
Title: "Including covariates in evidence synthesis"
By Dr Alex Sutton, Reader in Medical Statistics, Department of Health
Sciences, University of Leicester
Talk 2
Title: Can meta-analyses be believed? * A case study with sodium
cromoglicate.
By Mike Stevens, Statistician, Markfield, Leics
Talk 3
Title:Publication bias and selection functions
John Copas, University of Warwick
Please contact Dr Rachel Williams for further information.
Abstracts
Talk 2
Meta-analyses can be very influential in decision-making. This talk will
discuss a series of published meta-analyses for an anti-asthmatic drug,
sodium cromoglicate, and the impact the conclusions have had on the
place for the drug in the treatment of asthma in children.
Talk 3
Publication bias has been called the "bane of evidence based medicine".
It is impossible to allow for it unless we make strong assumptions about
the underlying selection mechanism. But what assumptions are
"reasonable", and are they "useful" in the sense of leading to good ways
of analysing data from systematic reviews? Should we aim for a single
conclusion, or should we make do with a sensitivity analysis? We shall
look at some recent research about selection functions in meta analysis,
and one or two examples.
Dr Rachel Hilliam
B.Sc.(Hons), Ph.D., FRSS, CMath.
Medical Statistician
Research and Development Office
Derby Hospitals NHS Foundation Trust
University of Nottingham Medical School at Derby
Uttoxeter Road
Derby
DE22 3DT
Tel. 01332 724679
Email: [log in to unmask]
|