Review: Analysis of the Sample Size Definitions in the Randomized Clinical Trials Included in Systematic Reviews of the Hepato-biliary Group of the Cochrane Collaboration

The investigators aim to answers to these questions

Questions

• How large a proportion of recent hepato-biliary RCTs does report an adequate sample size estimation?
• Primary outcome is clearly defined ? and clinical meaningful ?
• Is the sample size calculation based on the primary outcome ?
• Does the predicted control event rate coincide with the actual one?
• What is the predicted intervention effect?
• Are confidence intervals of primary outcome results reported? Is there evidence of difference between the two arms?
• Is the effect size so large (> 50%) to suggest post hoc definition ?
• What is the accepted error alpha and beta?
• How many study do not comply with the usual rule of alpha = 5% beta = 20%?
• Do trials with planned sample size do better than those without ? Is the overlap between CI results less frequent ? Are their results more frequently "statistical" significant ?
• Do trials with planned sample more frequently agree with pooled results in meta-analysis ?
• Has the protocol been published in a public register before inclusion of participant?
• How often is intention-to-treat analyses employed as primary analysis?
• How often is the trial results resting upon subgroup analysis?

Methods:

All the randomized clinical trials included in a systematic review of the Cochrane Hepato-biliary group (according to my search 272) were detected and retrieved. Only two arm, parallel group superiority randomised controlled trials with a single primary outcome.