Post authors: Dean Fergusson, Monica Taljaard and Charles Weijer [authors listed alphabetically]
Recently, members of our research team published an ethical analysis of the Flexibility in Duty Hour Requirements for Surgical Trainees (FIRST) trial in the Journal of Medical Ethics. The FIRST trial was a cluster-randomized trial that demonstrated that more flexible (and lengthier) duty hours for surgical trainees neither imperilled patient safety nor trainee wellbeing. Publication of the trial in the pages of the New England Journal of Medicine was met with controversy, and a vigorous exchange about its ethical permissibility. We argue that appeal to the Ottawa Statement on the Ethical Design and Conduct of Cluster Randomized Trials aids the ethical analysis of the FIRST trial greatly, yet that further guidance is required.
Cluster randomized trials such as the FIRST trial are commonly used to examine educational policies, public health interventions, and health service delivery interventions. In recent years, a novel type of cluster randomized trial design, the stepped wedge, has become an attractive design for such interventions. But another application of the cluster randomized design is gaining prominence: As part of a new wave of pragmatism sweeping healthcare research, cluster randomization is increasingly being used in learning health systems to evaluate the comparative effectiveness of treatments used in routine clinical practice. Such studies can, in principle, be conducted using individual randomization; yet, the use of cluster randomization can greatly enhance the pragmatic ends of the trial. By deliberately adopting more pragmatic trial designs, investigators seek to evaluate interventions in real-world conditions, and thereby inform decisions by patients, health providers and health system managers.
Foundationally, pragmatic trials raise novel ethical issues because they evaluate health interventions in real-world settings, including family practice offices and community hospitals, and this may obscure the distinction between research and practice. Thus, while the Ottawa Statement provided much needed guidance for cluster-randomized trials, new ethical guidance for pragmatic trials is required.
Pragmatic trials raise myriad ethical issues, and a number of these are well illustrated by the FIRST trial. First, do pragmatic trials require research ethics committee review? The research ethics committee that considered the FIRST trial deemed it involved no human participants and review was not required. As the study authors explain, “the institutions were the true ‘participants’.” We argue that there are good reasons to classify surgical residents as participants and that ethics review was required. For pragmatic trials generally, it remains common for trials involving policy or health services interventions to be exempted from review on grounds that it is quality improvement and not research. But is this practice ethically defensible?
Can health providers be research participants? The FIRST trial illustrates that interventions in pragmatic trials may target health providers as well as patients. When study interventions directly impact the interests of health providers, it is clear they should be considered participants. In other cases, though, the divide between provider and participant is less clear. In pragmatic trials, health providers commonly deliver the study intervention. Does that make them participants? What if an educational program for providers is part of the study intervention? Or the study intervention itself is a rearrangement of who provides care?
When, and from whom, is informed consent required? The FIRST trial sought informed consent from neither surgical residents nor patients. We argue that consent for data collection for residents was likely required, but consent from patients was not. Pragmatic trials typically seek to include a broader, more representative group of participants through simplified eligibility criteria and recruitment procedures. Some see informed consent as an impediment to these pragmatic ends. Importantly, the Ottawa Statement supports the use of a waiver of consent when cluster-level interventions pose only minimal risk. But can waivers of consent be extended legitimately to cluster randomized trials evaluating individual-level interventions or more generally, to comparative effectiveness research of routine medical treatments? Can a study that has mortality as an outcome ever be considered minimal risk?
Where to look for ethical guidance for pragmatic trials? Members of our research team recently published a scoping review of publications on the ethics of pragmatic trials that identified important gaps. Even though pragmatic trials are global in their reach, most articles appealed only to U.S. federal regulations. As these regulations were not written with pragmatic trials in view, reliance on them is akin to a driver looking for his lost car keys under a streetlamp because the light is better. Rather, like the Ottawa Statement, ethical guidance for pragmatic trials must be grounded in the internationally accepted ethical principles of respect for persons, beneficence and justice. Indeed, our research group has embarked on a four-year project to provide needed ethical guidance for pragmatic trials.
DF, MT and CW are co-Principal Investigators of the project, “Developing a framework for the ethical design and conduct of pragmatic randomized controlled trials,” funded by the Canadian Institutes of Health Research.
Competing interests: None declared.