A sample real dataset with time varying covariates. It contains multiple records
of measurements of risk factors at multiple time points from 10 patients
with primary biliary cirrhosis (PBC), constructed from
pbcseq in the R package survival. The data
structure has been reformatted with left-truncated right-censored
pseudo-subject observations based on the Andersen-Gill reformulation method.
a data frame with 57 rows and 9 variables:
Patient id
the left truncation time point for the corresponding pseudo-subject observation
the right censoring time point for the corresponding pseudo-subject observation
a binary value, with 1 indicating the event occurring at
the corresponding Stop, 0 indicating right-censored
patient's age at entry, in years
alkaline phosphotase (U/liter)
aspartate aminotransferase (U/ml)
serum cholesterol (mg/dl)
0--no edema, 0.5--untreated or successfully treated,
1--edema despite diuretic therapy
Andersen, P. and Gill, R. (1982). Cox's regression model for counting processes, a large sample study. Annals of Statistics, 10, 1100-1120.