This is a utility function.
The original records with first.well, last.well and
first.ill are
expanded to multiple records; several for each interval where the
person is known to be well and one where the person is known to fail.
At the same time columns for the covariates needed to estimate rates
and the response variable are generated.
expand.data(fu, formula, breaks, data)A 3-column matrix with first.well, last.well and
first.ill in each row.
Model fromula, used to derive the model matrix.
Defines the intervals in which the baseline rate is assumed constant. All follow-up before the first and after the last break is discarded.
Datafrem in which fu and formula is interpreted.
Returns a list with three components
Dataframe of covariates for estimation of the
baseline rates --- one per interval defined by breaks.
Dataframe for estimation of the covariate effects. A
data-framed version of the designmatrix from formula.
Response vector.
B Carstensen: Regression models for interval censored survival data: application to HIV infection in Danish homosexual men. Statistics in Medicine, 15(20):2177-2189, 1996.