STATISTICAL ANALYSIS The analytical approach and power calculations have been published previously, 14 and the com- plete prespecified statistical analysis plan is avail- able with the protocol at NEJM.org. In brief, a formal interim analysis was originally planned when 75% of primary outcome events had oc- curred. However, during the conduct of the trial it became apparent that the interim analysis would occur close to the end of the trial. The executive committee therefore decided to remove the in- terim analysis from the protocol. We estimated that 681 primary outcome events would be needed to detect a 22% lower relative risk in the dapagliflozin group than in the placebo group (hazard ratio of 0.78), with 90% power using a two-sided alpha level of 0.05 and assuming an annual event rate for the primary outcome of 7.5% in the placebo group. A closed testing pro- cedure that included a prespecified hierarchical order of the primary and secondary outcomes was used to ensure control of type I error at a two-sided 0.05 level.
I need help dissecting this (attached) in layman's terms.
Statistical Analysis-
The analytical approach and power calculations
have been published previously,14 and the complete
prespecified statistical analysis plan is available
with the protocol at NEJM.org. In brief, a
formal interim analysis was originally planned
when 75% of primary outcome events had occurred.
However, during the conduct of the trial it
became apparent that the interim analysis would
occur close to the end of the trial. The executive
committee therefore decided to remove the interim
analysis from the protocol. We estimated
that 681 primary outcome events would be
needed to detect a 22% lower relative risk in the
dapagliflozin group than in the placebo group
(hazard ratio of 0.78), with 90% power using a
two-sided alpha level of 0.05 and assuming an
annual
7.5% in the placebo group. A closed testing procedure
that included a prespecified hierarchical
order of the primary and secondary outcomes
was used to ensure control of type I error at a
two-sided 0.05 level.
Step by step
Solved in 2 steps
