Effectiveness of Choosing Wisely recommendations in reducing physiotherapists' intentions to refer for imaging and use electrotherapy for low back pain: a randomised controlled experiment

Abstract

QUESTION: What is the effect of Choosing Wisely recommendations on physiotherapists' intentions to refer for imaging and use electrotherapy for low back pain? DESIGN: Three-arm parallel-group online randomised controlled trial. PARTICIPANTS: Physiotherapists who treat people with low back pain. INTERVENTION: Participants were randomised to receive: (a) two original Australian Physiotherapy Association Choosing Wisely recommendations about low back pain, (b) two optimised versions of these recommendations based on previous research and (c) no recommendations. Participants were then directed to read three clinical vignettes of a person with low back pain and respond to questions regarding each vignette. OUTCOME MEASURES: Primary outcomes were physiotherapists' intentions to refer for imaging and use electrotherapy for low back pain. Secondary outcomes were physiotherapists' intentions to use other treatments for low back pain, the influence of the recommendations on decision-making in the vignettes and familiarity with the recommendations. RESULTS: 723 participants opened the survey and 473 (65%) provided complete responses. Across all vignettes, there were no statistically significant differences in intentions to refer for imaging or use electrotherapy between those who received Choosing Wisely recommendations versus no recommendation (imaging ORs ranging from 0.7 (95% CI 0.5 to 1.0) to 0.9 (0.6 to 1.4); electrotherapy ORs ranging from 0.9 (0.5 to 1.7) to 1.1 (0.7 to 2.0)). Similarly, no significant differences were observed between those who received optimised versus original recommendations for all three vignettes. CONCLUSION: Our study suggests simply presenting Choosing Wisely recommendations to physiotherapists does not influence their intentions to refer for imaging or use electrotherapy for low back pain, even if the language of the recommendations is optimised.