Article,
Subgroup analysis in MS trials.
Multiple sclerosis (Houndmills, Basingstoke, England), 23 (1): 34-35 (January 2017)Comparacions múltiples; Anàlisi de subgrups; MS; Introductori; Exemple; Errors<br/><br/>Exemple d'error en la presentació e interpretació de dades.
DOI: 10.1177/1352458515625808
Abstract
Subgroup analysis is often conducted as a post-hoc evaluation of clinical trials. The aim of a subgroup analysis is the evaluation of the treatment effect that was tested in the trial, in a specific subgroups of patients. It can be run both on positive trials (to provide information about patients receiving the highest benefit from the treatment) and on negative trials (to test whether the treatment that had no effect on the overall population can be of any benefit in a specific subset of patients). A subgroup analysis is aimed at generating hypotheses for future research. Subgroup analyses have statistical challenges involving multiple testing and unplanned and low powered analyses; however the main issue, at least in subgroup analysis conducted so far in MS studies, seems to be related to the reporting and interpretation of results. In this viewpoint I will try to show the misleading ways of reporting subgroup analysis in MS trials, along with the correct approach based on an interaction test.
