Skip to main content icon/video/no-internet
icon backReturn to Entries

Wennberg Design

The Wennberg design is one of a number of variations of the randomized controlled trial (RCT), in which participants’ preferences whether to receive the new treatment or the comparison treatment are taken into account. The primary purpose is to minimize refusals at the recruitment stage because of the reluctance of some participants to be randomized. It might also, however, increase compliance with the intervention and improve retention rates (reduce attrition) over time as patients are receiving their preferred treatment modality.

Rationale for the Design

Many procedures used in clinical research, such as random allocation, blinding of participants and raters, matching, stratification, and so forth, are done with one goal in mind—to reduce bias. In this context, bias means any systematic, nonrandom effect other than the intervention that can influence the results. The effect of the bias could be to weaken or distort the effect of the intervention or to spuriously overestimate its effect. However, one bias that cannot be dealt with using these techniques is the reluctance of some people to participate in a trial that simply randomizes them to the new treatment or the comparison condition, ignoring their preferences.

In some studies, as many as 80% of patients refuse to be randomized because they want either the new (and presumably better) treatment; do not want the new treatment because it is unproven; or are concerned that they might be randomized to the comparison condition, which might in some cases be a placebo group. If these people were a random subset of the population of interest, then the effect would be “merely” a greater difficulty in achieving the desired sample size. However, studies done over the past 50 years have shown that those who agree to participate in a study differ systematically from those who refuse; among other things, there is a greater tendency for them to be married, employed in professional or skilled jobs, be from the upper half of the socioeconomic spectrum, be more concerned about their health, be less likely to smoke, to have children living at home, and be either Jewish or Protestant. Consequently, participants in a trial represent a biased subset of people, which jeopardizes the external validity of the study; that is, the ability to generalize the results to the population for whom the intervention is designed. Furthermore, study participants who did not receive their preferred intervention respond less favorably to the treatment to which they were allocated—an effect called resentful demoralization—which increases the likelihood that they will be nonadherent to the study protocol and to drop out before the study ends. This jeopardizes the internal validity of the trial (i.e., the freedom from alternative explanations of the obtained results).

Patient Preference Designs

There have been several attempts to deal with preference bias through modification of the classical RCT design in clinical research. At the core, the modifications are principally around allocation of subjects to either the treatment or control conditions. As expected, all of these designs offer some degree of decision authority to participants but at different stages of the consenting process. The Zelen randomization, for example, uses a post-randomization consent model—participants are randomized before consent is sought, and then only those in the group who will receive the experimental treatment are asked to consent. In the control arm, patients receive treatment as usual (TAU). Those selected to receive the experimental treatment can either refuse to participate, and therefore receive TAU, or agree and become part of the experimental group. The refusers are given the standard treatment but are analyzed as if they were in the treatment group. The main purpose of the design, therefore, is to reduce preference bias associated with being allocated to the control condition. However, this design is feasible only if the comparison group receives TAU (as opposed to, say, a placebo), and if consent is not required to gather any additional information from them. For these reasons, it has rarely been used.

...

  • Loading...
locked icon

Sign in to access this content

Get a 30 day FREE TRIAL

  • Watch videos from a variety of sources bringing classroom topics to life
  • Read modern, diverse business cases
  • Explore hundreds of books and reference titles

Read next

More like this

Sage Recommends

We found other relevant content for you on other Sage platforms.

close

Have you created a personal profile? Login or create a profile so that you can save clips, playlists and searches

close
close
close Signed In to profile You are signed in as:
close
Logged in Institution Institution