Active placebo versus standard placebo control interventions in pharmacological randomised trials: a systematic review

Background: placebos are used as control interventions in randomized trials to enable blinding of participants and personnel. This can help reduce the risk of bias due to, for example, observer bias, reporting bias and bias related to placebo effects. Standard placebos in pharmacological trials imitate the appearance of the experimental intervention, without containing the active component under investigation. They can be, for instance, saline infusions or lactose tablets.

However, the experimental drug may have detectable physiological effects or adverse effects that can lead to unblinding, and thus cause bias. This is suspected to be a potential issue with antidepressants as well as methylphenidate as a treatment for attention deficit hyperactivity disorder (ADHD).

As a potential solution, some trials have employed 'active placebo' as a control intervention instead of standard placebo. The active placebo contains a substance designed to imitate some of the physiological effects of the experimental treatment under investigation, but without being therapeutically active. One example is atropine as an active placebo for tricyclic antidepressants.

In theory, the use of active placebos as control interventions can reduce the risk of bias due to unblinding, but this has yet to be investigated systematically.

Objectives: to estimate the impact on estimated treatment effects in pharmacological randomized trials of using active placebo as compared to standard placebo.

Methods: a systematic review of randomized trials with both standard and active placebo control groups. We will search for eligible randomized trials in PubMed, CENTRAL and Embase, as well as Google Scholar, reference lists and citations of relevant papers.

For included trials, we will extract basic trial data and outcome results. From each trial, we will extract one of each of the following outcome types: a patient-reported outcome (continuous preferred), an observer-reported outcome (continuous preferred), a dichotomous harms outcome. We will select two time points for each outcome: earliest post-treatment and latest follow-up. Also, we will assess attrition rate and co-intervention use.

Based on the selected outcomes above, we will convert the outcome data to standardized mean differences (SMD) for active placebo versus standard placebo. We will then summarize the individual trials’ estimates for each outcome type in random-effects inverse-variance meta-analyses. The primary analysis is the analysis for patient-reported outcomes at earliest post-treatment. The remaining analyses are secondary analyses.

Results and conclusions: we will present results and conclusions at the Colloquium.

Patient or healthcare consumer involvement: none