Assessing the INSIDE assumption in Mendelian randomization
18th January - 12th February 2022
Biography
Dr Pires Hartwig received his MSc and Doctorate degrees in Epidemiology at Federal University of Pelotas (Brazil), where he currently is an assistant professor in Epidemiology and a permanent researcher at the Postgraduate Program in Epidemiology. Dr Pires Hartwig also holds an honorary research fellow position at Bristol Medical School (University of Bristol, UK). His research interests are related to the field of causal inference in epidemiology, with emphasis on empirical and methodological work on instrumental variable analysis. He is also interested in translation of statistical methods between different fields of epidemiology.
Summary
The proposed project is related to a statistical approach called Mendelian randomization (MR), which uses genes as tools to assess whether modifiable disease risk factors (e.g., smoking) are causally related to disease risk (e.g., cancer). The goal of MR is to provide reliable answers to causal questions, a knowledge that is essential to identify targets for effective interventions aimed at improving population health. For this, MR requires that certain assumptions hold. Much of the methodological work in the field is developing new statistical methods that require fewer assumptions. One of the first examples of this in the MR literature was the MR-Egger method, which only requires a weaker (that is, more plausible) version of the assumption required by standard method. The main assumption required by MR-Egger is known as INSIDE. Previous attempts to develop methods to test the INSIDE assumption have been shown to be fundamentally flawed. Therefore, while MR-Egger is more likely to yield correct results than standard methods, it still requires an assumption that might not be satisfied in (and cannot be tested using) the data. This project aims at developing a method assessing the plausibility of INSIDE. As preliminary results indicate, unless very specific (and rather contrived) conditions are in place, a simple statistical test can be used to test the validity of this assumption. A reliable test of INSIDE would be of great significance to applied researchers as a tool to assess the plausibility of MR-Egger assumptions, thus contributing to obtain more robust results using MR.
Dr Pires Hartwig is hosted by Professor George Davey Smith, Population Health Sciences in the Bristol Medical School
Planned events include:
Postgraduate Lecture
25th January, 1-2 pm
Exploiting genetic data to answer questions of public health relevance
Zoom link on this page
Open Seminar
1st February, 1-2 pm
Contributions of the Pelotas birth cohorts to studying the relationship between early-life poverty and health
Zoom link on this page
IEU Seminar
8th February, 1-2 pm
The NO Simultaneous Heterogeneity assumption for average causal effect estimation via instrumental variables
Zoom link on this page
