Table 1 Assumptions made in conducting randomized controlled trials

Equipoise

Patient and provider do not have a preference for a treatment

Lack of knowledge

It is truly unknown which of two alternatives is "better" and there is insufficient evidence about treatment effects from other sources

Preference for specificity

Only specific effects attributable to the intervention are therapeutically valid

Context independence

There is a "true" magnitude of efficacy, or a stable effect size independent of context

Ecological and external validity knowable

The knowledge about a therapeutic effect extracted from an RCT is readily transferable into clinical practice, if exclusion and inclusion criteria of the trial match the characteristics of a given patient