Summary for resmean_phreg/cif_yearslost fits

If times=NULL was used when creating the object, the summary is the full curve at all event times – exactly what plot would show – as a data frame with the estimate, its standard error, confidence interval, and (for resmean_phreg) a years-lost column (time - rmst); for cif_yearslost one such table is given per cause.

Usage

# S3 method for class 'resmean_phreg'
summary(object, contrast = NULL, cause = NULL, time = NULL, level = 0.95, ...)

Arguments

object

Object of class "resmean_phreg".

contrast

Optional contrast (matrix or function), applied via lava::estimate before summarizing; triggers the usual summary.estimate test output. Only used when times was given.

cause

Only relevant for cif_yearslost objects: restrict the summary to this cause (default: all causes).

time

Optional: restrict the summary to a single time (only relevant if several times were given in the original call).

level

Confidence level, only used for the no-times curve-summary.

...

Additional arguments passed to summary.estimate (only used together with contrast) or to conftype (when times=NULL).

Details

If times was given, the object already stores the relevant lava "estimate" object(s) (see estimate.resmean_phreg), and this just returns those stored object(s) as-is – printing one shows the usual Estimate/Std.Err/CI/P-value table. No joint null-hypothesis test is added by default; pass contrast (a matrix or function, applied via lava::estimate) if you want to test a specific linear combination, in which case the usual summary.estimate output (including the test) is shown for that contrast (see summary.resmean_estimate for how this is applied when there is more than one time/cause). For cif_yearslost objects, cause=NULL (the default) summarizes every cause (returned as a named list); cause=1 (say) restricts the summary to that cause. time similarly restricts to a single time (only relevant if length(times)>1 in the original call).