Towards more credible shams for physical interventions: A Delphi survey

Abstract

Background/aims: In clinical trials of physical interventions, participant blinding is often poorly addressed and therapist blinding routinely omitted. This situation presents a substantial barrier to moving the field forward. Improving the success of blinding will be a vital step towards determining the true mechanisms of physical interventions. We used a Delphi approach to identify important elements of shams for physical interventions to maximise the likelihood of participant and therapist blinding in clinical trials.

Methods: Two expert groups were recruited: (1) experts in research methodology and (2) experts in deceptive and/or hypnotic techniques including magic. Magicians were included because they were considered a potentially rich source of innovation for developing credible shams due to their unique skills in altering perceptions and beliefs. Three rounds of survey were conducted, commencing with an open-ended question. Responses were converted to single 'items', which participants rated in the following two rounds using a 9-point Likert scale, categorised as 'Not important' (0-3), 'Depends' (4-6) and 'Essential' (7-9). Consensus was pre-defined as ≥80% agreement within a 3-point category.

Results: Thirty-eight experts agreed to participate (research methodology: n = 22; deceptive and/or hypnotic techniques: n = 16), and 30 experts responded to at least one round (research methodology: n = 19; deceptive and/or hypnotic techniques: n = 11). Of 79 items, five reached consensus in the 'Essential' category in both groups, which related to beliefs of participants (n = 3 items), interactions with researchers (n = 1 item) and standardisation of clinical assessments (n = 1 item). Thirteen additional items reached consensus in the 'Essential' category in one group. Experts in research methodology had one additional item reach consensus, related to authentic delivery of study information. The remaining 12 additional items that reached consensus in the deceptive and/or hypnotic techniques group related mainly to therapist attitude and behaviour and the clinical interaction.

Conclusion: Experts agreed that, for shams to be believable, consideration of cognitive influences is essential. Contrary to the focus of previous shams for physical interventions, replicating the tactile sensation of the active treatment was not considered an essential part of sham development. Therefore, when designing sham-controlled clinical trials, researchers should carefully consider the cognitive credibility of the entire intervention experience, and not just the indistinguishability of the sham intervention itself. The findings provide new guidance to researchers on important contributors to blinding in physical intervention trials.

Keywords: Delphi; blinding; physical interventions; placebo; sham.