Zelen's Randomized Consent Design

Overview

The RCT has earned its position as the methodological gold standard for studies of effectiveness, particularly in health research. Because outcomes are being compared across groups of individuals that have been formed through the process of random allocation, if the groups are of sufficient size, any difference observed between the groups is caused by the intervention. However, this assumes the study is not biased or contaminated.

For example, as an RCT compares different interventions (most commonly, different treatment options), many trial participants will prefer one treatment over another. So-called preference effects can potentially lead to biased results, for example, participants randomized to an intervention they do not want might lead them to (consciously or unconsciously) respond differently to the treatment. A particular version of this has been termed resentful demoralization. Resentful demoralization can affect trial continuance; that is, demoralized participants might be more likely to drop out, introducing selection bias. These participants might be more likely to report adverse events or to report more negatively on outcome measures. The effect of resentful demoralization on study outcomes is acknowledged as a potential confounder in open randomized trials (that is, those without blinding to the intervention).

In addition to the potential introduction of bias, there is a longstanding practical problem of recruitment into RCTs—both at the level of engaging physicians/clinicians who will enable recruitment of patients as well as difficulties in recruiting patients themselves. Recruitment difficulty largely stems from problems clinicians might have in explaining to patients core trial concepts—especially equipoise and randomization. This, together with the acknowledgment of the uncertainty relating to patient care, can make the process of fully informed consent problematic. Recruitment difficulties can result in unrepresentative samples, therefore, having implications for external validity.

These difficulties stem from the fact that in most RCTs consent to both treatment and random allocation is sought from participants before randomization. In addressing the issue of recruitment difficulty, which also addresses the problem of resentful demoralization, Zelen proposed two modifications to the usual practice of gaining consent before randomization, known as the single-and double-consent methods.

...