The Centre for Biostatistics at The University of Manchester and the Royal Statistical Society Manchester Local Group are hosting the following seminar.
All are welcome to attend, and there is no need to register in advance.
Time: 16:05 - 17:00, Wednesday 25th February 2015
Location: Room 1.009, Roscoe Building, Brunswick Street, Manchester, M13 9PL
Speaker: Dr Stephen Burgess, University of Cambridge
Title: "Instrumental variable analysis with no valid instruments"
Abstract:
Mendelian randomization is the use of genetic variants as instrumental variables to obtain causal inferences in observational data. The number of Mendelian randomisation analyses including large numbers of genetic variants is rapidly increasing. This is due to the proliferation of genome-wide association studies, and the desire to obtain more precise estimates of causal effects. However, some genetic variants may not be valid instrumental variables, in particular due to pleiotropy.
We view show that bias caused by pleiotropy can be viewed as analogous to publication bias. Causal estimates using each instrument individually can be displayed visually by a funnel plot to assess potential asymmetry. Egger's test, a tool to detect publication bias, can be used to test for bias from pleiotropy. Under the assumption that the association of each genetic variant with the exposure is independent of the direct effect of the variant on the outcome (not via the exposure), Egger's test gives a valid test of the null causal hypothesis and a consistent causal effect estimate even when all the genetic variants are invalid instrumental variables. This provides a sensitivity analysis for the robustness of the findings from a Mendelian randomization investigation.
Refreshments: Tea and coffee will be available from 15:30.
You may leave the list at any time by sending the command
SIGNOFF allstat
to [log in to unmask], leaving the subject line blank.
|