A Bayesian reanalysis of the Standard versus Accelerated Initiation of Renal-Replacement Therapy in Acute Kidney Injury (STARRT-AKI) trial

Table 2 Results for the primary endpoint according to different priors

Prior	Median	HDI 95%	P (Benefit)*	%ROPE**	P (effect not large)^‡	P (OR < 0.84)^†	P (diff < − 0.04)^⁋
Theoretical priors
Neutral	1.01	0.87–1.15	0.47	0.99	1.00	0.00	0.01
Optimistic	0.98	0.86–1.12	0.59	0.99	1.00	0.00	0.02
Pessimistic	1.03	0.90–1.18	0.35	0.99	1.00	0.00	0.00
Data driven priors
AKIKI	0.99	0.87–1.13	0.54	0.99	1.00	0.00	0.01
ELAIN	0.96	0.83–1.10	0.73	0.97	1.00	0.00	0.04
Meta-analysis	0.99	0.88–1.12	0.56	0.99	1.00	0.00	0.00
STARRT-AKI Pilot	1.00	0.87–1.15	0.48	0.98	1.00	0.00	0.01

*Probability OR < 1.0. **Probability effect size (OR) is within 0.83–1.19 (equivalence margin). ^‡Probability effect size is outside a large margin effect of OR between 0.77 and 1.30. ^†Probability OR is below 0.84 (which results in a 4% reduction in primary outcome). ^⁋Probability the difference is outcome is greater than 4% favoring accelerated strategy given the data and prior

ISSN: 1364-8535