For further information on any CSM events please e-mail:

Centre for Statistical Methodology Seminar
Thursday 28 Nov 2019, 12:45-2:00pm
Bradford Hill Room K/LG09, Keppel Street
Verena Zuber (Imperial College London)

Title: Selecting causal risk factors from high-throughput experiments using multivariable Mendelian randomization.


Modern high-throughput experiments provide a rich resource to investigate causal determinants of disease risk. Mendelian randomization (MR) is the use of genetic variants as instrumental variables to infer the causal effect of a specific risk factor on an outcome. Multivariable MR is an extension of the standard MR framework to consider multiple potential risk factors in a single model. However, current implementations of multivariable MR use standard linear regression and hence perform poorly with many risk factors.

Here, we propose a novel approach to multivariable MR based on Bayesian model averaging (MR-BMA) that scales to high-throughput experiments and can select biomarker as causal risk factors for disease. In a realistic simulation study we show that MR-BMA can detect true causal risk factors even when the candidate risk factors are highly correlated. We illustrate MR-BMA by analysing publicly-available summarized data on metabolites to prioritise likely causal biomarkers for cardiovascular disease.

Centre for Statistical Methodology Seminar
Thursday 12 Dec 2019, 12:45-2:00pm
Curtis Room K/LG08, Keppel Street
Fabrizia Mealli (University of Florence)

Title: Assessing causal effects in the presence of treatment switching through principal stratification.


Consider clinical trials focusing on survival outcomes for patients suffering from Acquired Immune Deficiency Syndrome (AIDS)-related illnesses or particularly painful cancers in advanced stages. These trials often allow patients in the control arm to switch to the treatment arm if their physical conditions are worse than certain tolerance levels. The Intention-To-Treat analysis compares groups formed by randomization regardless of the treatment actually received. Although it provides valid causal estimates of the effect of assignment, it does not measure the effect of the actual receipt of the treatment and ignores the information of treatment switching in the control group. Other existing methods propose to reconstruct the outcome a unit would have had if s/he had not switched. But these methods usually rely on strong assumptions, for example, there exists no relation between patient’s prognosis and switching behavior, or the treatment effect is constant. Clearly, the switching status of the units in the control group contains important post-treatment information, which is useful to characterize the treatment effect heterogeneity. We propose to re-define the problem of treatment switching using principal stratification and introduce new causal estimands, principal causal effects for patients belonging to subpopulations defined by the switching behavior under control. For statistical inference, we use a Bayesian approach to take into account that (i) switching happens in continuous time generating infinitely many principal strata; (ii) switching time is not defined for units who never switch in a particular experiment; and (iii) survival time and switching time are subject to censoring. We illustrate our framework using a synthetic dataset based on the Concorde study, a randomized controlled trial aimed to assess causal effects on time-to-disease progression or death of immediate versus deferred treatment with zidovudine among patients with asymptomatic HIV infection. Joint work with Alessandra Mattei and Peng Ding.