Strong Heterogeneity of Outcome Reporting in Systematic Reviews
Résumé
OBJECTIVES: The Core Outcome Measures in Effectiveness Trial initiative aims at developing core outcome set (COS) for research. Consensus on a COS for renoprotection research is lacking. We performed a systematic review (SR) of SRs of renoprotection to identify outcomes used in renoprotection research and to assess how frequently these outcomes could be meta-analyzed in the SRs. STUDY DESIGN AND SETTING: We searched for SRs with meta-analyses of renoprotection treatments in the Database of Abstracts of Reviews of Effects and in MEDLINE and collected outcomes that were meta-analyzed. For each outcome and SR, we assessed the proportion of trials/patients that could be meta-analyzed. RESULTS: We retrieved 66 SRs. A total of 609 outcomes were extracted for 20 distinct renoprotection outcomes. The median (interquartile range) proportion of SRs in which these outcomes had been meta-analyzed was 8% (2%, 27%) (range 2-50%). The proportion of trials that had been aggregated was >75% for only 36% of the outcomes. CONCLUSION: The outcomes meta-analyzed in renoprotection trials are heterogeneous, with a low proportion of trials or patients pooled for each outcome and each SR. A COS is needed in renoprotection research to limit this waste of research. The outcomes identified in the present work could be a basis for this COS.