Study
Opening the study…
Study
Opening the study…
Where the work did something other than what it follows, its registered plan or the claim it replicates, as its author says.
From its pre-registered plan, under plan/
Not stated
The plan specified sharp coupling bounds and attaining witnesses but not the additional optimization minimizing landmark-survival discordance under a zero individual binary late-death contrast; that optimization was added after preliminary bounds were computed and is exploratory.
Bears on C1
Not stated
Exact current Evolut early risks were not labeled on the matched curves. Conservative visually checked intervals were selected after inspecting those curves and before the interval-region LP calculations: at year 1 TAVR 0–3% and surgery 0–4%, and at year 5 on the nonsweep curve 12–15% and 14–17%. These are deterministic sensitivity regions, not confidence intervals.
Bears on C1
Not stated
After the binary-null results, an exploratory stronger construction required every always-survivor to remain alive through the later horizon. Its capacity condition permits arbitrary marginal within-window death-time patterns without digitization. Minimizing discordance within that restricted class and repeating over rounding regions were added after the initial results were known.
Bears on C1
Not stated
The shared-early-death frontier displays the fixed-stratum formula already used for validation. The additional nonnegative-association constraint for early potential mortality was selected after the initial results and the stronger construction; it is exploratory, not a measured association or a prespecified identifying assumption.
Bears on C1
Not stated
The omission sensitivity assumes aligned time origins. The source curves are indexed by procedure; randomization-to-procedure delays and index dates for participants without the assigned procedure were not reconstructed, so this remains a mixture sensitivity model rather than an actual intention-to-treat reanalysis. Additional exploratory calibrations minimize omitted-surgery mortality needed for a binary null and compare the target with overall mortality; witness ties are broken by minimizing omitted TAVR mortality. These changes do not assert that omitted participants had those outcomes.
Bears on C2