The following Departmental seminars will take place in Room A, 2nd floor
Wellcome Trust Building, at the UCL Institute of Child Health,
30 Guilford Street, London WC1N 1EH
Wed 26th October, 2pm Room A
Dr Alan Kimber
School of Applied Statistics, The University of Reading
Growth and development of low birthweight preterm babies
In this talk I will discuss the results from a small randomised trial based
at the
Royal Hampshire County Hospital in Winchester, to investigate the effect of
an early
weaning strategy for low birthweight preterm babies. The babies were
followed up from birth
until age 42 months and the results of the analyses of the data are considered.
Some issues relating to growth curve models are also raised in this context.
===============================
Wed 16th November, 2pm Room A
Dr Tanyana Jackiewicz
Telethon Institute for Child Health Research, University of Western Australia
TBA
===============================
Wed 23rd November, 2pm Room A
Dr Cosetta Minelli
Genetic Epidemiology Group, Department of Health Sciences, University of
Leicester
Recommendations for the meta-analysis of genetic association studies
Genetic association studies have shown a disappointing failure to replicate
and validate postulated associations.
Evidence synthesis has been advocated to overcome the problem of inadequate
sample size and help identify and
address other important methodological issues of primary studies. However,
despite the fast growth in number
of published meta-analyses of genetic association studies, their
methodological aspects have received little
attention. The need for guidelines for the conduct of meta-analysis in this
field is reflected by the poor
quality affecting many published meta-analyses. In this talk, an overview
of the issues which require consideration
when synthesising the evidence on genetic association studies will be
presented, and practical recommendations
proposed for dealing with both general meta-analysis issues and issues
specific to meta-analysis of genetic
association studies. These recommendations are based on evidence available
in the literature and on our original work.
Particular emphasis will be given to two specific aspects; 1) the choice of
how to pool the data across genotypes
based on assumptions about the underlying genetic model; 2) the evaluation
of departures from Hardy-Weinberg
equilibrium as a proxy for poor quality of primary studies, and the
decision of how to handle studies with departures.
Methods currently used to deal with these issues will be reviewed and
discussed, and alternative approaches
will be presented.
All welcome.
Instructions for getting here can be found in:
http://www.ich.ucl.ac.uk/gosh_families/your_visit/getting_here/index.html
_______________________
Dr Mario Cortina Borja
Senior Lecturer in Statistics
[log in to unmask] Phone +44(0)20 7905 2113 Fax +44(0)20 7905 2381
Paed. Epid. & Biostats, Institute of Child Health, UCL, London WC1N 1EH, UK
|