OBJECTIVES: Assess the quality of evidence on experimental and observational clinical research through the same approach named GRADE (Grading of Recommendations Assessment, Development and Evaluation) METHODS: We evaluated the primary endpoint of a random sample from efficacy and effectiveness studies included in a systematic review on the treatment of rheumatoid arthritis with anti-TNFs. The quality assessment was conducted in accordance with the recommendations of the GRADE Working Group available at: www.gradeworkinggroup.org/toolbox/index.htm. It assigns at first high quality for trials and low quality for observational studies. RESULTS: The assessment of 8 efficacy and 8 effectiveness studies showed respectively that the quality of evidence were high in 5 and 0; moderate in 3 and 2; low in 0 and 2; and very low in 0 and 4. The risk of bias was present in 3 and 5; imprecision results in 0 and 5; elevated magnitude of effect in 6 and 4; controlled confounding bias in 0 and 3; presence of dose-response gradient in 0 and 1. Indirect evidence and inconsistencies were not found in any of the studies. CONCLUSIONS: The quality of evidence of 3 trials and 4 observational studies were downgraded, while 2 observational studies had the quality assessment increased.