Evaluating Non –Statistically Significant Results From Trials in Practice

To the Editor Dr Young and colleagues suggested that there are occasions when clinicians may treat statistically nonsignificant results as clinically meaningful and incorporate them in important clinical decision-making, especially when considering the effect size and associated confidence intervals. This recommendation is problematic because it elevates underpowered studies in clinical decision-making. Even more problematic in this context is the assumption that inferences can be made directly from group-level data to the individual, even if the group-level results are ambiguous or nonsignificant.
Source: JAMA - Category: General Medicine Source Type: research