Reminder: the next meeting of Birmingham University Medical Statistics group
will be held next Wednesday 13th October, details below. All are welcome.
BIRMINGHAM UNIVERSITY MEDICAL STATISTICS GROUP
WORKSHOP ON BAYESIAN VERSUS FREQUENTIST METHODS
<?xml:namespace prefix = o ns = "urn:schemas-microsoft-com:office:office"
/>
Wednesday 30 October 2002
5.15-8.00pm
Rooms 223 & 224
Primary Care Clinical Sciences Building
Cindy Billingham
A Bayesian Approach to Survival Analysis: Practicalities, Issues and
Arguments
Survival data from a large phase III trial in non-small cell lung cancer
will be analysed using a Bayesian approach. The practicalities of such an
approach will be examined and the results will be compared to those obtained
from a standard frequentist analysis. The advantages and disadvantages of
the Bayesian approach will be explored, opening the forum for discussion.
Keith Wheatley
Incorporating prior expectations into clinical trial data monitoring
Differences between treatments compared in clinical trials are, in
expectation, moderate. An extreme p-value seen associated with a large
treatment difference is therefore less plausible than the same p-value seen
with a moderate difference. The implications for clinical trial data
monitoring are discussed with practical examples.
Chair: John Duffy
After 2 hours of presentation and open discussion (participants are welcome
to show up to 3 overheads) the meeting will conclude with informal
interaction over a glass of wine: all welcome
Directions: turn left out of University station and the Primary Care
building is 50 metres downhill past the statue on the left.
|