Table 2 Preliminary framework of ethical issues raised by pragmatic randomized controlled trials (pragmatic RCTs)

1.

Are activities in pragmatic RCTs research or practice?

• Commonly evaluate interventions used in routine clinical practice

• Seek to evaluate interventions and assess outcomes in usual care settings

• Favor unobtrusive data collection via routinely collected sources

• Difficult to clearly separate research from clinical practice

• Advocates of learning health systems have challenged the research-practice distinction, but unclear if a new system will be preferable and how it satisfies current research regulations

2.

What level of oversight is required for pragmatic RCTs?

• May involve usual care interventions or interventions posing no more than minimal risk

• Address questions that directly inform decision-making by patients and healthcare providers

• May have quality and service improvement as a central goal

• Current oversight procedures are time-consuming, costly, and overly complex

• Some advocate that low-risk pragmatic RCTs do not require more stringent oversight than clinical practice as all patients receive an intervention used in routine clinical practice

3.

Which study designs are appropriate in pragmatic RCTs?

• Aim to maximize representativeness; may require larger sample sizes to have adequate power

• May favor novel designs, including cluster randomized, stepped wedge, registry, and cohort multiple designs

• Study designs can have differing implications for trial feasibility and logistics which must be balanced against internal and external validity

• RCTs can substantially simplify trial logistics and facilitate recruitment, but ethical justification unclear

• Interventions of unknown effectiveness may be rolled out to all clusters in stepped wedge trials

4.

Who are the research participants in pragmatic RCTs?

• May involve stakeholders at multiple levels: health system; hospital; provider; patient

• May evaluate interventions targeting one group, but measure outcomes on another

• Identification of research participants influences the scope of research ethics review, benefit–harm analysis, and informed consent procedures

5.

Do patients and providers have an ethical obligation to participate in pragmatic RCTs?

• Seek to provide highly relevant evidence for patients, providers, and health systems

• High degree of flexibility in intervention delivery and data collection implies low burden of study participation

• Unclear if health system leaders and providers have a prima facie ethical duty to seek to continually improve the delivery and outcomes of healthcare

• Patients receive benefits from the health system which may engender an obligation to participate in research

6.

From whom, how, and when is informed consent required in pragmatic RCTs?

• May involve usual care interventions or interventions posing no more than minimum risk

• May use solely routinely collected data for outcome assessment

• Informed consent can pose barriers to representative recruitment, be a burden on staff, add logistical complexity, and be costly

• May expose differing groups of participants to different aspects of the research

• May use cluster-level interventions that are difficult or impossible to avoid

• Unclear if simplified (“altered”) or no consent procedures are acceptable and under which conditions

• Unclear what aspects of research must be disclosed to research participants

• Unclear if informed consent is needed from health system leaders, decision-makers, health providers

7.

Who are the gatekeepers in pragmatic RCTs and what are their responsibilities?

• May have an impact on group or institutional interests

• May involve a variety of gatekeepers, including ministry of health, hospital administrator, and data custodians

• Lack of clarity regarding the role of gatekeepers, whose interests they protect and scope of their authority

• Unclear when community consultation is appropriate

8.

How should harm–benefit analyses be conducted in pragmatic RCTs?

• Procedures administered by a wide range of providers under imperfect conditions and in a variety of institutional settings

• May evaluate policy or health delivery system interventions or usual care interventions

• May use delayed implementation of intervention (stepped wedge design)

• Unclear how to determine the boundaries of appropriate provider experience and training

• What constitutes usual care

• Benefit–harm analysis in studies of policy or healthcare delivery is unclear

• Unclear when it is appropriate to delay delivery of interventions in stepped wedge designs

9.

How ought vulnerable groups be protected in pragmatic RCTs?

• Seek to study a representative group including individuals traditionally considered vulnerable (lower socioeconomic strata, children, pregnant women, prisoners)

• Health providers/ employees may be exposed to social risks, including reputational or professional harm

• Some criticize the traditional approach to defining vulnerability and have proposed alternative definitions

• Presence of vulnerable participants may be hidden

• Unclear how to protect employees in pragmatic RCTs

• Unclear who should set the research agenda for pragmatic RCTs