PhD in estimands/causal inference in trials (UK/EU)

If you are a UK/EU resident interested in pursuing a PhD on estimands/causal inference in clinical trials, please see the advert here. There is (rightly) increasing emphasis in clinical trials in clear specification of the scientific question and hence target estimand or parameter.

While one might think the process of choosing and specifying the estimand is usually easy, in many settings various things can happen during follow-up which complicate this. Examples include patients changing treatments, failing from competing risks, or dying before the endpoint of interest can be measured. This has led to the ICH E9 addendum on estimands, whose final version will soon be published. There remain a number of areas where deciding what the most appropriate estimand is and how one can validly estimate it from the observable data is challenging, and this PhD will seek to address some of these outstanding areas. For more background on this area, I’d recommend reading this paper.

The PhD will be based at the University of Bath, with myself as primary supervisor. The student will benefit from additional supervision from leading researchers in causal inference: Rhian Daniel (Cardiff), Jack Bowden (Bristol) and Daniel Farewell (Cardiff).

For information about funding and the application process, please see the information here. The application deadline is 25th November 2019.

Causal interpretation of the hazard ratio from RCTs when proportional hazards holds

In 2015 I wrote a post about the causal interpretation of hazard ratios estimated in randomised trials, following a paper by Aalen and colleagues. One of the arguments made in that paper was that the hazard ratio does not have a valid interpretation as a causal effect in this setting, even when the proportional hazards assumption holds:

This makes it unclear what the hazard ratio computed for a randomized survival study really means. Note, that this has nothing to do with the fit of the Cox model. The model may fit perfectly in the marginal case with X as the only covariate, but the present problem remains.

With recent discussions on estimands in light of the estimand addendum to ICH E9, I have been thinking more on the argument/claim by Aalen et al.

Read more

Why you shouldn’t use propensity score matching

I’ve just watched a highly thought provoking presentation by Gary King of Harvard, available here, on why propensity score matching should not be used to adjust for confounding in observational studies. The presentation makes great use of graphs to explain the concepts and arguments for some of the issues with propensity score matching.

Read more