Evaluating Non –Statistically Significant Results From Trials in Practice

To the Editor The Viewpoint by Dr Young and colleagues discussed the interpretation of non –statistically significant results from clinical trials. The topic of statistical vs clinical significance has long been debated in medical research. Although there have been attempts to dismiss P values, arguing that the criteria for statistical significance should be substituted with “discussi ng compatible effects,” most of the decisions in medicine remain dichotomous: a new drug will either be licensed or not and it will either be used for a certain condition or not. We believe P values and confidence intervals still have an important place in interpreting well-conducted clinical tria ls because of their objectiveness, although any interpretation leading to decision-making is far more complex than statistical significance.
Source: JAMA - Category: General Medicine Source Type: research