Physiotherapy. 2021 Aug 25:S0031-9406(21)00082-1. doi: 10.1016/j.physio.2021.08.007. Online ahead of print.
OBJECTIVES: Dry needling is widely used in physiotherapy. Lack of blinding in clinical trials means that dry needling effects and mechanisms remain unclear, with blinding issues accepted as an unavoidable barrier to better evidence. The authors aimed to overcome this barrier by designing a protocol to double-blind dry needling trials.
DESIGN: A prospectively registered randomised experiment tested a novel blinding protocol for blinding effectiveness.
SETTING: University physiotherapy clinic.
PARTICIPANTS: Therapists (n=15) and asymptomatic volunteers (‘recipients’) (n=45) were randomly allocated to real and/or sham interventions.
INTERVENTIONS: The protocol involved custom-made needles and cognitive and multisensory blinding techniques.
MAIN OUTCOME MEASURES: The primary outcome was guesses about allocation. The a priori criterion for successful blinding was ≤50% correct guesses (random chance). Secondary analyses explored blinding patterns using blinding indices.
RESULTS: Correct guesses were not different from 50% for therapists [41% (95%CI 30 to 50), n=120 guesses] or recipients [49% (95%CI 38 to 60), n=90 guesses]. Blinding indices supported the primary result but revealed that recipients were better at detecting real dry needling than sham.
CONCLUSION: Both therapists and recipients were successfully blinded, which contrasts with the widely held assumption that double-blinding is impossible for needling interventions. The authors recommend that any future trials can, and therefore should, blind therapists. However, secondary analyses revealed that recipients receiving real dry needling were less blinded than sham recipients, which may still create bias in clinical trials and suggests further work is needed to improve recipient blinding. Nonetheless, the current findings offer an opportunity to gain better evidence concerning the effects and mechanisms of dry needling.
STUDY REGISTRATION: https://osf.io/rkzeb/.