It can be difficult to assess analgesic effects of clinical interventions due to the possibility of the placebo effect.This is because randomization and blinding are not possible in most clinical settings.As such, 30% or 50% reduction in pain intensity are often used as a criterion of efficacy based on psychometric studies showing that these ratings correspond to reports of "moderate" or "substantial" clinical benefit, respectively(Dworkin et al 2008).However, we do not know the likelihood of patients reporting such pain intensity changes following placebo interventions, which hinders our ability to estimate the possibility of the placebo effect in the clinical setting.

We systematically searched for randomized control trials and case series of facial pain treatments that included placebo intervention and measured percentage change in pain intensity.Random-effects meta-analyses were performed to estimate pooled effect sizes. We used PRISMA 2020 flow diagram for source search (PRISMA:Preferred Reporting Items for Systematic reviews and Meta-Analyses), we visualized the distribution and cumulative probability of percent change in pain ratings after placebo (absolute value).

Our literature search yielded 117 eligible studies resulting in 3,710 facial pain patients receiving placebo/sham interventions. The pooled effect size was 29.25% pain reduction with placebo/sham (95%CI:24.49% to 34.01; prediction interval=-18.78% to 77.28%). We found that reports of >30% or >50% change in pain intensity were not uncommon (~50% and ~25% of patients, respectively). However, reports of >80% change in pain intensity were very rare after a placebo intervention (~5% of patients).

In facial pain, reports of >30% or >50% improvements in pain intensity are not incompatible with a placebo effect but reports of >80% pain intensity improvements suggest that the placebo effect is very unlikely (<~5% probability).