Trend in randomized controlled trials
- PMID: 15526517
Trend in randomized controlled trials
Abstract
Virtually any study may achieve statistical significance if we increase the sample size indefinitely. We cannot discuss about trend without interpreting confidence intervals (CI). The sample size being computed function of the magnitude of the clinical effect is considered worth detecting, when the result is not statistically significant, it is neither clinically significant. Increasing the sample, the "trend" can disappear or even change its direction.