Avaliação da qualidade de estudos clínicos e seu impacto nas metanálises

Abstract

OBJECTIVE: To evaluate whether different quality assessment tools applied to a group of clinical trials could be correlated, and what would be their impact on meta-analysis results. METHODS: Thirty-eight randomized controlled clinical trials were analyzed. These had been selected for a systematic review of the therapeutic efficacy of alpha interferon for treating chronic hepatitis B. The following tools were utilized: Maastricht (M), Delphi (D), Jadad (J) and the Cochrane Collaboration (CC) method (gold standard). The Spearman correlation coefficient was used to compare the results from the three methods. The Kappa test was used to assess the concordance between the reviewers in applying the tools, and the weighted Kappa test was applied to compare the quality ranking determined by the tools. The outcomes assessed in the meta-analyses were clearance of HBV-DNA and HBeAg. RESULTS: The studies presented regular to low quality. The concordance between reviewers varied according to the instrument utilized: D=0.12; J=0.29; M=0.33; and CC=0.53. The correlation was moderate and homogeneous (D/J=0.51; D/M=0.53; and J/M=0.52). The meta-analysis result relating to HBV-DNA ranged from RR=0.71 (95% CI: 0.66-0.77) to RR=0.67 (95% CI: 0.58-0.79). For HBeAg, the results ranged from RR=0.85 (95% CI: 0.80-0.90) to RR=0.85 ( 95% CI: 0.77-0.93). These results depended on the quality of the studies included. CONCLUSIONS: The quality assessment tools presented good correlation. In systematic reviews with the same direction of effect, the quality assessment may not significantly change the results. The Cochrane Collaboration method was the most reproducible method and easiest to apply.

Cite this work

No public copy is stored for this record, so only the metadata is cited.

Stored only when both papers are in the corpus — 24 citations exist upstream.

Papers that cite the same references — the relational fallback while embeddings are still backfilling.