epilepsy: CBZ versus LTG in Epilepsy Patients
In JSM: Semiparametric Joint Modeling of Survival and Longitudinal Data

Description Format Source References Examples

A randomised control trial, the SANAD (standard and new antiepileptic drugs) study, in which both survival and longitudinal data were collected to investigate the effect of drug titration on the relative effects of two antiepileptic drugs, namely CBZ (carbamazepine, a standard drug) and LTG (lamotrigine, a new drug), on treatment failure. Treatment failure, i.e. withdrawal of the randomized drug, is the event of interest. Two main reasons for withdrawal are unacceptable adverse effects (UAE) and inadequate seizure control (ISC).

A data frame with 2797 observations on the following 16 variables.

ID: patient ID, there are 605 patients in total.
Time: survival time, i.e. time to withdrawal or censoring.
withdraw: withdrawal indicator: 0 denotes censoring; 1 denotes withdrawal.
withdrawUAE: withdrawal due to UAE indicator: 1 denotes withdrawal due to UAE; 0 otherwise.
withdrawISC: withdrawal due to ISC indicator: 1 denotes withdrawal due to ISC; 0 otherwise.
obstime: time points at which the longitudinal measurements, i.e. the dose, are recorded.
dose: calibrated dose measured at obstime.
drug: drug indicator with two levels: CBZ and LTG.
age: age of patient at study entry.
gender: gender indicator with two levels: male and female.
disab: learning disability indicator with two levels: No and Yes.
start: same with obstime, starting time of the interval which contains the time of the dose measurement.
stop: ending time of the interval which contains the time of the dose measurement.
event: event indicator suggesting whether the event-of-interest, i.e. withdrawal, happens in the interval given by start and stop.
eventUAE: event indicator suggesting whether the event-of-interest, i.e. withdrawal due to UAE, happens in the interval given by start and stop.
eventISC: event indicator suggesting whether the event-of-interest, i.e. withdrawal due to ISC, happens in the interval given by start and stop.

Marson, A. G., AI-Kharusi, A. M., Alwaidh, M., Appleton, R., Baker, G. A., Chadwick, D. W., Cramp, C., Cockerell, O. C. Cooper, P. N., Doughty, J. et al. (2007) The SANAD study of effectiveness of carbamazepine, gabapentin, lamotrigine, oxcarbazepine, or topiramate for treatment of partial epilepsy: an unblinded randomised controlled trial. The Lancet 369, 1000–1015.

Williamson P. R., Kolamunnage-Dona R., Philipson P. and Marson A. G. (2008) Joint modelling of longitudinal and competing risks data. Statistics in Medicine 27, 6426–6438.

1	head(epilepsy)

Loading required package: nlme
Loading required package: splines
Loading required package: statmod
Loading required package: survival

Attaching package: 'JSM'

The following object is masked from 'package:survival':

    pbc

  ID Time withdraw withdrawUAE withdrawISC obstime dose drug   age gender disab
1  1 2400        0           0           0      86 2.00  CBZ 75.67   male    No
2  1 2400        0           0           0     119 2.00  CBZ 75.67   male    No
3  1 2400        0           0           0     268 2.00  CBZ 75.67   male    No
4  1 2400        0           0           0     451 2.67  CBZ 75.67   male    No
5  1 2400        0           0           0     535 2.67  CBZ 75.67   male    No
6  1 2400        0           0           0     770 2.67  CBZ 75.67   male    No
  start stop event eventUAE eventISC
1    86  119     0        0        0
2   119  268     0        0        0
3   268  451     0        0        0
4   451  535     0        0        0
5   535  770     0        0        0
6   770 1515     0        0        0