Zelen’s Randomized Consent Design

Limitations of Conventional Informed Consent

[Page 1838]In a conventional randomized controlled trial, participants provide informed consent before being randomly allocated to (usually) the intervention or control treatment groups. This can present a number of difficulties for researchers and for the clinicians who are oftentimes tasked with entering patients into the trial. Obtaining consent to participate in conventional trials requires full disclosure to participants about all potential benefits and risks associated with participating in the trial, including the intervention being tested and the alternative (control) treatments used for comparison. Participants may be concerned about being randomized to the control group (often seen as the inferior or less effective intervention) and/or not receiving their preferred intervention and refuse to participate. This can slow trial recruitment and may limit the feasibility of conducting research in some circumstances. It also has implications for the risk of bias. Participants who hold positive attitudes toward, and beliefs about, the intervention being studied are more likely to enroll than those who don’t. This may mean the sample ultimately recruited is not reflective of the broader clinical population. Importantly, knowledge about the intervention of interest and the alternative (control) can influence participants’ adherence to treatment as well as their self-reported response to treatment. It is widely acknowledged that patient pretreatment expectations of outcome influence treatment response on patient-reported domains such as pain, physical function, quality of life, and mental health. Put simply, participants who expect to get better with treatment at the outset tend to report better treatment outcomes than those who don’t. For participants allocated to an intervention that they perceive to be less effective than the alternative, resentful demoralization may occur. This is particularly relevant to participants allocated to control, placebo, or “usual care” trial arms. This increases the risk of increased dropouts and/or reporting of poorer outcomes and threatens the validity of the trial.

Randomized Consent

Zelen’s design can overcome some of the problems experienced with conventional consent procedures by randomizing participants before consent to participate has been obtained. Zelen proposed two types of randomized consent design termed single and double consent designs, respectively. Participants randomized to the control group (usually a standard treatment or usual care) with single consent are not asked for consent to participate in the trial. Instead, they are provided standard treatment (or usual care) for their medical disease or condition, without knowing they are participating in the trial and without knowledge of the other treatment arm. Their data are collected (usually) as part of standard treatment and used as a comparison for the intervention group. Only participants allocated to the intervention group are asked to consent, and if they decline the intervention of interest, they receive standard treatment (or usual care). With double consent designs, participants in the control group are also approached to provide consent to receive standard treatment. Those who decline receive an alternative treatment that can include the experimental intervention. With both designs, trial data are analyzed according to original random allocation, irrespective of the treatment received.

...