Effects of Clinician- or AI-Delivered Guideline-Concordant Advice on Opioid Use in Acute Low Back Pain: A 2×2 Factorial Randomised Clinical Trial

Intervention Participants will be randomly allocated in a 1:1:1:1 ratio to one of four intervention groups. Each participant will receive only one intervention and will view one approximately one minute educational video. Before viewing the video, participants will be asked to read a brief hypothetical scenario describing that they have acute low back pain and are consulting a healthcare professional about pain management. They will then watch the assigned video and complete a post intervention questionnaire assessing their beliefs and intended use of opioids. The four intervention groups are: Arm 1: Clinician delivered, "avoid opioids" advice: Participants view a video featuring a real clinician delivering evidence based advice that opioids are not recommended and should be avoided for acute low back pain. Arm 2: Clinician delivered, "last resort" advice: Participants view a video featuring a real clinician delivering evidence based advice that opioids should only be considered as a last resort for acute low back pain after other recommended treatments have been tried. Arm 3: AI delivered, "avoid opioids" advice: Participants view a video in which the advice is delivered by an AI generated human avatar with high realism in facial appearance, voice, and movements. The content is identical to Arm 1, recommending that opioids are not recommended and should be avoided for acute low back pain. Arm 4: AI delivered, "last resort" advice: Participants view a video in which the advice is delivered by an AI generated human avatar with high realism in facial appearance, voice, and movements. The content is identical to Arm 2, recommending that opioids should only be considered as a last resort for acute low back pain. Intervention content: All four videos were developed specifically for this study and are approximately one minute in duration. The videos communicate guideline concordant advice regarding opioid use for acute low back pain. The advice provided to participants in the video (to avoid opioids where possible, or use them only as a last resort for acute low back pain) is evidence-based, reflecting current international clinical practice guideline recommendations for the management of non-specific low back pain. Participants are not simply told this is "evidence-based" without basis — the advice itself is derived directly from published guideline recommendations. This is supported by the following reference, which provides an overview of global clinical practice guidelines for non-specific low back pain: Oliveira CB, Koes BW, Pinto RZ, Traeger AC, Machado GC, Lin CC, et al. Towards global clinical practice guidelines for the management of non-specific low back pain in primary care: a review of current guideline recommendations and how they have changed over the last 30 years. Lancet Rheumatol. 2026;8(6):e470-e85. doi: 10.1016/S2665-9913(26)00077-9. PubMed PMID: 42203385. The only differences between intervention groups are: 1. Mode of delivery: real clinician versus AI generated human avatar. 2. Advice framing: opioids should be avoided versus opioids should only be used as a last resort. Apart from these two factors, the script, duration, and overall structure of the videos are standardised across all groups. Total intervention duration: The intervention consists of reading one hypothetical clinical scenario and viewing one approximately one minute educational video. Including the scenario, the intervention takes approximately 2 to 3 minutes to complete. Participants then complete the study questionnaire. Intervention adherence: Intervention delivery is fully automated within REDCap. Participants must view their allocated video before proceeding to the post intervention questionnaire. REDCap records questionnaire completion and study completion status, allowing monitoring of intervention completion and overall participant adherence.