An expectation-management script and its effect on placebo and diclofenac responses in knee osteoarthritis (PAKOA): a randomized controlled trial

Objectives To assess whether a placebo response mitigation script (PRMS) alters placebo responses and the discrimination between active treatment and placebo in a trial of diclofenac for knee osteoarthritis pain. Design Multicentre, randomised, placebo-controlled, double-blinded trial. Setting Three research centres in Denmark, with enrolment between October 2024 and January 2026. Participants 150 adults aged 40 to 85 years with symptomatic and radiographic knee osteoarthritis. Interventions Participants were randomly assigned (1:1:1:1) to diclofenac 50 mg twice daily or matching placebo, with or without a PRMS, for four weeks. The PRMS was a standardised psychoeducational script intended to reduce the expectancy component of placebo responses. Main outcome measures The primary endpoint was change from baseline to week four in Knee Injury and Osteoarthritis Outcome Score (KOOS) pain subscale (0-100 scale, 0 indicating worst symptoms), analysed using mixed models for repeated measures under a treatment-policy estimand. Results 150 participants (82 [55%] were female; mean age 63.9 [SD 8.5] years) were randomly assigned to diclofenac plus PRMS (n=38), diclofenac alone (n=37), placebo plus PRMS (n=37), or placebo alone (n=38). Participants receiving PRMS reported lower expectations of treatment benefit than those without PRMS (mean 1.83 [SD 1.46] vs 3.12 [SD 1.35]; p<0.0001). At week 4, the difference in KOOS pain improvement between diclofenac and placebo was 4.0 points (95% CI -2.3 to 10.2; p=0.21; Cohen's d=0.31) among participants who received the PRMS and 7.4 points (0.9 to 13.8; p=0.025; Cohen's d=0.56) among those who did not. A similar pattern was observed across pain-related secondary endpoints, whereas functional outcomes showed no between-group differences. Conclusions A placebo response mitigation script reduced both placebo and diclofenac responses, lowering assay sensitivity. This challenges the assumption that placebo responses can be selectively reduced without affecting active treatment effects. Trial registration EudraCT 2024-510757-95-01.