The “efficacy paradox”

Abstract

The “efficacy paradox” occurs when the efficacy of a treatment in a randomised controlled trial (RCT), or the recommendation in an evidence-based guideline, differs from the treatment benefits seen in clinical practice (1–3). This occurs because in RCT reporting and guidelines, treatment efficacy is judged by the difference in outcome between treatment and placebo (i.e., the specific effect of the treatment). This difference is used to estimate the “strength” or effect size (ES) of a treatment, and only treatments that separate from placebo by a minimum clinically important difference (MCID) may be recommended for use (e.g., by NICE). However, RCT participants who receive placebo often do very well, and in the treatment group the overall treatment ES (which includes both specific and contextual responses) is often in the good-strong range of ES. When a patient receives a treatment in practice it is the overall treatment effect, not just the specific treatment effect that they experience. This difference in focus can cause discordance between guidelines and clinical practice and ignores the impact of contextual response in clinical care (1). An alternative way of reporting RCTs is to firstly highlight the overall treatment effect (as would be expected in practice), and then explain how much of this is attributable to specific treatment effect and how much to placebo/contextual effects (the “proportional contextual effect (PCE)” (2)). This interpretation of all the results would diminish the paradox, alter the hierarchy of “strong treatments” for individual conditions (based on overall treatment rather than just specific treatment effects), and might lead to changes in guideline recommendations. For many chronic painful distressing conditions (e.g., OA, fibromyalgia) the majority of treatment benefit results from contextual effects, and optimising such effects in clinical practice could lead to substantial improvements in patient-centred outcomes. 1. Walach H. J Altern Complement Med 2001;7(3):213–218. 2. Zhang et al. Ann Intern Med 2015;163:392–393. 3. Zhang W, Doherty M. Clin Immunol 2017. https://doi.org/10.1016/j.clim.2017.07.018.