Towards more homogenous and rigorous methods in sham-controlled dry needling trials: two Delphi surveys

Sham-controlled trials of dry needling, a popular treatment for pain, use a range of methods and theoretical frameworks and most have high risk of bias. Critically, patient blinding is often unsuccessful and therapist blinding has not been attempted. The specific effects of dry needling on pain therefore remain unclear. Our objectives were to identify (1) important elements of active dry needling; (2) important elements of shams for dry needling.

Two Delphi surveys (to quantify levels of consensus) were undertaken with three expert groups: experts in (1) dry needling, (2) research methodology, and (3) deceptive/hypnotic techniques including magic. Experts in dry needling participated in Delphi 1 and all three groups participated in Delphi 2. Each survey commenced with an open-ended question. Responses were converted to single 'items' suitable for rating on 9-point Likert scales [categorised as 'Not important' (0-3), Depends (4-6), and Essential (7-9)], which participants rated in the following two rounds. Consensus was pre-defined as ≥80% agreement within a 3-point category.

In Delphi 1 (n=20 experts), of 80 items, 35 reached consensus in the 'Essential' category, which related to explanations, therapist knowledge/skills, intervention rationale, the setting, and safety. In Delphi 2 (n=53 experts), of 97 items, 15 items reached consensus in the 'Essential' category in all three groups, which related to standardisation/indistinguishability, therapist attributes, expectations/beliefs, vision, protocol, and environment.

Experts placed high importance on the entire intervention experience for active and sham protocols. Cognitive influences that extend beyond mimicking of tactile sensations should be used to create a believable simulation of active dry needling.