I have survival data and I would like to change the structure of the dataset.
stnum=patient id
survtime=time to event stat=event 0/1
loctime = time to event lcens =0/1
axltime=time to event acens=0/1
distime =time to event dcens=0/1
maltime=time to event mcens =0/1
first_t =time to first event first_e=0,1,2,3,4,5 where 0=censoring and 1-5 define the type of event as shown in the variables above
Please see data structure below for 20 obs
----------------------- copy starting from the next line -----------------------
Code:
* Example generated by -dataex-. To install: ssc install dataex clear input int stnum float survtime byte stat float loctime byte lcens float axltime byte acens float distime byte dcens float maltime byte mcens float first_t byte first_e 1 8.26831 0 8.26831 0 8.26831 0 8.26831 0 8.26831 0 8.26831 0 2 6.173853 0 6.173853 0 6.173853 0 1.3524983 1 6.173853 0 1.3524983 4 3 7.175907 0 7.175907 0 7.175907 0 7.175907 0 7.175907 0 7.175907 0 4 9.505818 0 9.505818 0 9.505818 0 9.505818 0 9.505818 0 9.505818 0 5 9.09514 0 9.09514 0 9.09514 0 9.09514 0 9.09514 0 9.09514 0 6 8.095825 1 8.095825 0 8.095825 0 8.095825 0 8.095825 0 8.095825 1 7 .7255304 1 .7255304 0 .7255304 0 .6023272 1 .7255304 0 .6023272 4 8 4.963724 0 4.963724 0 4.963724 0 4.963724 0 4.963724 0 4.963724 0 9 8.848734 0 8.848734 0 8.848734 0 8.848734 0 8.848734 0 8.848734 0 10 5.163587 0 5.163587 0 5.163587 0 5.163587 0 5.163587 0 5.163587 0 11 8.837782 0 8.837782 0 8.837782 0 8.837782 0 8.837782 0 8.837782 0 12 8.99384 0 8.99384 0 8.99384 0 8.99384 0 8.99384 0 8.99384 0 13 7.195072 0 7.195072 0 7.195072 0 7.195072 0 7.195072 0 7.195072 0 14 9.385352 0 9.385352 0 9.385352 0 7.909651 1 .9418207 1 .9418207 5 15 9.002053 0 9.002053 0 9.002053 0 9.002053 0 9.002053 0 9.002053 0 16 1.0349076 1 1.0349076 0 1.0349076 0 .7392197 1 1.0349076 0 .7392197 4 17 .3422314 1 .3422314 0 .3422314 0 .3422314 0 .312115 1 .312115 5 18 9.475701 0 9.475701 0 9.475701 0 9.475701 0 2.625599 1 2.625599 5 19 9.54141 0 9.54141 0 9.54141 0 9.54141 0 9.54141 0 9.54141 0 20 9.308693 0 9.308693 0 9.308693 0 9.308693 0 9.308693 0 9.308693 0 end
1. To expand the data in a way that each patient appears according to the number of events she/he had and the corresponding times 1 and 2. This would comply with the Andersen-Gill and the Prentice-William-Petersen models.
For example, stnum==1 had 0 events so time1=0 and time2=8.26831 with overall status==0 and event_type==0. This patient had no event and was cencored at time 8.26831. This patient will appear only once.
Patient stnum==2 had 2 events and will have to appear 3 times
time1 time 2 status event_type stnum
0 1.3524983 1 1 2
0 1.3524983 1 4 2
1.3524983 6.173853 0 0 2
ect
2. I would like to the create a structutre which would comply with the Wei-Lin-Weissfel model where each patient has 6 observations equal to the number of all possible event types.
stnum==1
event_type time status
0 8.26831 0
1 8.26831 0
2 8.26831 0
3 8.26831 0
4 8.26831 0
5 8.26831 0
stnum==2
event_type time status
0 6.173853 0
1 1.3524983 1
2 6.173853 0
3 6.173853 0
4 1.3524983 1
5 6.173853 0
I hope this is clear. any help would welcome.
Thank you,
Nikos
0 Response to Restructure Survival Dataset
Post a Comment