Multicentre randomised trials in anaesthesia: an analysis using Bayesian metrics

AnaesthesiaPublished 20 Sep 2022Structured review of multicentre randomised trials with Bayesian reanalysis

Cite

Authors' conclusion

We think that inadequate power and mortality as an outcome are why few trials declared non-zero effects.

Study context

Population: Multicentre anaesthesia trials published in NEJM, The Lancet, JAMA, BJA and Anesthesiology, February 2011 to November 2021
Participants: 56 primary outcomes

Comparison: Non-zero effect declared vs No non-zero effect declared
Non-zero effect declaredn = 12
No non-zero effect declaredn = 44
Primary endpoint: Whether trials declared a non-zero effect of the intervention on the primary outcome

Methodological detailsBayes factor, post-test probability (50% prior belief in zero effect)

As reported in the abstract

Hypothesised absolute effect size, median (IQR [range]): 7% (3-13% [0-25%])
Observed absolute effect size, median (IQR [range]): 2% (1-7% [0-24%])
Non-zero effects declared: 12/56 outcomes (21%)
Bayes factor favouring zero effect, trials declaring non-zero effect: 0.000001-1.9
Post-test zero-effect probability, trials declaring non-zero effect: 0.0001-65%
Bayes factor favouring zero effect, trials not declaring non-zero effect: 1-688
Post-test zero-effect probability, trials not declaring non-zero effect: 53-99%

Study costs (US$), reported by 20 corresponding authors

Median (IQR [range]): $1,425,669 ($514,766-$2,526,807 [$120,758-$24,763,921])

Multicentre randomised trials in anaesthesia: an analysis using Bayesian metrics. Anaesthesia. Published 20 September 2022. doi:10.1111/anae.15867

Composed by AI from the published abstract and checked against it. How Briefs are made →