Speaker: Nancy Cartwright (Department of Philosophy, University of Durham and University of California, San Diego)
Title: What are pragmatic trials in medicine good for?
Abstract: There is widespread call for increasing use of pragmatic trials in both medicine and social science nowadays. These are randomised controlled trials (RCTs) that are administered in 'more realistic' circumstances than standard, i.e. with more realistic treatment/programme delivery (e.g. busier, less well-trained doctors/social workers) and a wider range of recipients (e.g. ones that self select into treatment or have 'co-morbidities' or are already subject to a number of other interventions that might interfere with the treatment). Pragmatic trial results are supposed to be more readily 'generalisable' than results from those with more rigid protocols.
We argue that this is a mistake. Trials, pragmatic or otherwise, can only provide results about those individuals enrolled in the trial. Anything else requires assumptions from elsewhere, and generally strong ones. Based on a common understanding of what causal principles look like in these domains, this talk explains what results can be well warranted by an RCT and warns against the common advice to take the criteria for admission to a trial to be indicative of where else its results may be expected to hold.
