[Critical reading of treatment trials.]

Author(s): Fauvel JP

Affiliation(s): Service de nephrologie, Hopital Edouard Herriot, Lyon (69).

Publication date & source: 2007-02-01, Presse Med., [Epub ahead of print]

Publication type:

In view of the enormous quantity of medical information published, critical reading is essential for the rapid selection of information relevant to the physician's practice. Methodological quality is essential. The study design that produces the best evidence of the effects of an intervention is the randomized controlled clinical trial. Valid and reliable metaanalysis can increase the precision of these results. Only studies using clinically relevant outcome measures, analyzed according to intention to treat, deserve attention. Trials of lower methodological quality must be considered as hypothesis-generating trials. The groups receiving the study treatment and the reference treatment must be comparable to avoid sampling bias. The simpler the statistics, the more reliable the results, as a general rule.