A separate, unspecified baseline hazard per stratum with one shared
coefficient vector: the strata contribute their own partial
likelihoods, which are summed.
Usage
morie_cox_stratified(
time,
event,
X,
stratum,
ties = "efron",
max_iter = 50L,
tol = 1e-09
)
Arguments
- time
follow-up times, non-negative.
- event
1 for an event, 0 for right-censored.
- X
covariate matrix, one row per subject.
- stratum
stratum label per subject.
- ties
"efron" or "breslow".
- max_iter, tol
Newton-Raphson controls.
Value
list with beta, se, z, p_value,
hazard_ratio, loglik, strata,
events_per_stratum, empty_strata.
Details
The stratifier gets no coefficient and no hazard ratio – that is the
price of not assuming proportional hazards for it. Stratify on the
nuisance variable, never on the exposure you are trying to estimate.
Strata with no events contribute nothing at all and are reported.
References
Kalbfleisch, J. D. and Prentice, R. L. (2002). The
Statistical Analysis of Failure Time Data, 2nd ed., Sec. 4.4.
Examples
set.seed(1)
X <- matrix(rnorm(80), ncol = 1)
s <- rep(1:2, each = 40)
tt <- rexp(80, exp(X * 0.5) * s)
morie_cox_stratified(tt, rep(1, 80), X, s)$beta
#> [1] 0.6277523
