mets (version 1.2.9)

recreg: Recurrent events regression with terminal event

Description

Fits Ghosh-Lin IPCW cox type model

Usage

recreg(
  formula,
  data = data,
  cause = 1,
  death.code = c(2),
  cens.code = 1,
  cens.model = ~1,
  weights = NULL,
  offset = NULL,
  Gc = NULL,
  ...
)

Arguments

formula

formula with 'EventCens' outcome

data

data frame

cause

of interest

death.code

codes for death (terminating event)

cens.code

code of censoring (1 default)

cens.model

for stratified Cox model without covariates

weights

weights for score equations

offset

offsets for model

Gc

censoring weights for time argument, default is to calculate these with a Kaplan-Meier estimator, should then give G_c(T_i-)

...

Additional arguments to lower level funtions

Author

Thomas Scheike

Details

For Cox type model : $$ E(dN_1(t)|X) = \mu_0(t)ddtt exp(X^T \beta) $$ by solving Cox IPCW weighted score equations $$ \int (Z - E(t)) w(t) dN_1(t) $$ where $$w(t) = G(t) (I(T_i \wedge t < C_i)/G_c(T_i \wedge t))$$ and $$E(t) = S_1(t)/S_0(t)$$ and $$S_j(t) = \sum X_i^j w_i(t) \exp(X_i^T \beta)$$.

The iid decomposition of the beta's are on the form $$ \int (Z - E ) w(t) dM_1 + \int q(s)/p(s) dM_c $$ and returned as iid.

Events, deaths and censorings are specified via stop start structure and the EventCens call, that via a status vector and cause (code) and death.code(s) indentifies these, a second censoring indicator specifies where the censorings. See example.

Examples

## data with no ties
data(base1cumhaz)
data(base4cumhaz)
data(drcumhaz)
Lam1 <- base1cumhaz;  Lam2 <- base4cumhaz;  LamD <- drcumhaz
## simulates recurrent events of types 1 and 2 and with terminal event D and censoring
rr <- simRecurrentII(1000,Lam1,cumhaz2=Lam2,death.cumhaz=LamD,cens=3/5000)
rr <- count.history(rr)
rr$cens <- 0
dsort(rr) <- ~id-start
nid <- max(rr$id)
rr$revnr <- cumsumstrata(rep(1,nrow(rr)),rr$id-1,nid)
dsort(rr) <- ~id+start
rr$x <- rnorm(nid)[rr$id]
rr <- dtransform(rr,cens=1,revnr==1 & death==0)
dcut(rr) <- gx~x

rr <- dtransform(rr,statusG=status)
rr <- dtransform(rr,statusG=0,status==2)
rr <- dtransform(rr,statusG=2,death==1)

ll <- recreg(EventCens(start,stop,statusG,cens)~x+cluster(id),data=rr,cause=1)
summary(ll)

## censoring stratified after quartiles of x
lls <- recreg(EventCens(start, stop, statusG, cens) ~ x+cluster(id), data=rr,
              cause=1, cens.model=~strata(gx))
summary(lls)