Trial methodology and patient characteristics did not influence the size of placebo effects on pain.
Level 1 - systematic review of randomized trials
Meta-regression of randomized clinical trials
PubMed 18226748 · doi:10.1016/j.jclinepi.2007.03.017
What was done
Random-effects meta-regression analyzing clinical trials with pain outcomes from the Hróbjartsson and Gøtzsche dataset. Eight factors were coded and evaluated for their ability to predict placebo effect size: trial design (nonstandardized co-analgesia, co-intervention), patient factors (pain type, patient group, residual pain score), and placebo factors (placebo type, indistinguishability, structural equivalence). The pooled effect of placebo was also recalculated.
What was found
The pooled placebo effect was 3.2 points on a 100-point scale (95% CI: 1.6 to 4.7). None of the eight evaluated factors significantly predicted the size of the placebo effect: effect sizes for all factors were close to zero, all 95% confidence intervals spanned zero, and P-values ranged from 0.13 to 0.90.
Why it matters
This study confirms that average placebo effects in clinical pain trials are small and establishes that larger placebo responses are not systematically explained by specific trial designs, patient groups, or placebo characteristics.
Limits
The abstract does not report the number of trials or total participants analyzed. Findings are constrained to the subset of trials from a single existing meta-analysis, and trial-level meta-regression may lack power to detect small subgroup differences.
Cited by
- context The strength of a placebo response increases with the size and invasiveness of the intervention, with larger pills producing stronger effects than smaller pills, and sham injections producing stronger effects than pills.