Addressing 'hard-to-fund' research questions in bowel research: Exploring the role of decision architecture...

Patients and health professionals face difficult decisions every day when choosing between standard-of-care treatments for the same disease. In bowel disease, some treatments are used interchangeably, but a lack of evidence may exist as to whether one is more effective than another. Randomised controlled trials provide robust evidence in these settings, but they are expensive, time-consuming and disruptive to usual care. The introduction of electronic health records (EHR) has created unique opportunities to develop efficient trial designs, such as Decision Architecture Randomised Trials (DART). These implement unobtrusive and sometimes indiscernible EHR 'nudges', prompting clinicians towards one treatment or another. There is no obligation for clinicians to use the assigned intervention or for the patient to accept it. However, the association between the assignment and the actual treatment received can be used to estimate unbiased treatment effects. We discuss the key features of DART studies and outline the persisting uncertainties to operationalising this study design in clinical practice.