Abstract

Background: Well designed clinical trials yield the strongest evidence for the effect of health care interventions. Aim: To assess the methodological quality of the design and report of randomized clinical trials in a sample, published in biomedical Chilean journals between 1980 and 2002. Material and methods: All trials identified by hand search by the Unit of Evaluation of Technologies in Health, were assessed for quality of randomization, blinding, analysis of results and other characteristics of trial design, along with the application of Jadad's Score, that assesses the methodological quality of clinical trials in a scale that ranges from 0 to 5. Results: Twenty eight trials were found and assessed, 75% (n=21) specified the method used for randomization, 29% (n=8) described a correct allocation concealment and 39% (n=11) were double blinded. Withdrawals and dropouts were correctly reported in 21% (n=6) of the articles, whereas intention to treat analysis was done only in one. Thirteen trials had a Jadad score equal or higher than 3 points. Conclusions: Several design deficiencies were found in the trials assessed. It is difficult to know if methodological weaknesses are due to incomplete reports or to methodologically poor designs. Adopting initiatives like the CONSORT can help improve the quality of randomized clinical trials published in Biomedical Chilean journals.