Placebo-adjusted effect (revision 12)
Old revision·12:32, 4 Apr 2025·AxisLabelAgda
| Placebo-adjusted effectTrial methodology | |
|---|---|
| Also called | Between-group difference, treatment difference |
| Computed as | Active-arm change minus placebo-arm change |
| Why it matters | The active-arm change alone is not the drug effect |
| Topic infobox · conventions | |
The placebo-adjusted effect is the difference between the change observed in a treatment arm and the change observed in the comparator arm. It is the quantity a randomised trial is designed to estimate, and it is what can be attributed to the treatment.[1]
The active-arm change alone conflates the treatment effect with everything else operating during the trial: regression to the mean, the natural course of the condition, the lifestyle intervention both arms received, and the effects of being observed.[2]
In obesity trials the distinction is substantial. STEP 1 reported −14.9% on semaglutide and −2.4% on placebo, so the placebo-adjusted effect is about 12.4 percentage points — not 14.9.[3]
What the comparator arm captures
[edit]A placebo arm in a modern obesity trial is not a no-treatment arm. Participants receive the same lifestyle counselling, the same visit schedule and the same measurement, and the difference between the arms is the treatment alone.[2][4]
Several distinct phenomena contribute to the change in a placebo arm. Regression to the mean operates because participants are enrolled at a moment when their measurement was high enough to qualify. The trial's own intervention — diet advice, activity advice, regular weighing — has an effect. Observation changes behaviour. And the natural course of a condition may be towards improvement or deterioration.[2]
The placebo response proper — improvement attributable to expectation — is only one of these, and in objective endpoints such as body weight it is generally small relative to the others.[2]
How figures are misreported
[edit]| Reported as | Actually | Effect on the impression |
|---|---|---|
| "Lost 15% of body weight" | Active-arm change | Overstates the drug effect |
| "12 percentage points more than placebo" | Placebo-adjusted | Correct |
| "Six times the placebo effect" | Ratio of arms | Sensitive to a small denominator |
| "Up to 24%" | Highest dose, best case | Not the trial's headline result |
The third row deserves attention: a ratio between arms is unstable when the placebo change is near zero, and a trial with a placebo change of −0.5% will produce a spectacular ratio from an unremarkable difference.[1]
Cross-trial comparison compounds all of this. Two trials with different lifestyle interventions have different placebo arms, so their active-arm figures are not comparable even when the drug effects are.[3]
See also
References
- ^ a b International Council for Harmonisation, E9(R1): Estimands and Sensitivity Analysis in Clinical Trials (2019).
- ^ a b c d Hróbjartsson A, Gøtzsche PC. "Placebo interventions for all clinical conditions." Cochrane Database of Systematic Reviews (1):CD003974 (2010). PMID 20091554.
- ^ a b Wilding JPH, Batterham RL, Calanna S, et al. "Once-weekly semaglutide in adults with overweight or obesity." New England Journal of Medicine 384(11):989–1002 (2021). PMID 33567185.
- ^ International Council for Harmonisation, E10: Choice of Control Group and Related Issues in Clinical Trials (2000).