survival analysis and length of follow up

[kende jan]

Hi,

How to compare survival time between 2 groups of treatment observed over the same period (for example between 2000 and 2005) BUT when more individuals from the group 2 were enrolled later (for example on 2003) and then have a shorter length of follow-up ?

Does Cox allow such comparison taking into account the potential bias?

Many thanks for your contribution.

SL.




====================To manage your subscription to SPSSX-L, send a message to
[hidden email] (not to SPSSX-L), with no body text except the
command. To leave the list, send the command
SIGNOFF SPSSX-L
For a list of commands to manage subscriptions, send the command
INFO REFCARD

[Marta Garcia-Granero]

kende jan escribió:
> How to compare survival time between 2 groups of treatment observed over the same period (for example between 2000 and 2005) BUT when more individuals from the group 2 were enrolled later (for example on 2003) and then have a shorter length of follow-up ?
>

First of all, you must be quite sure that there is no "late entry bias":
subjects enrolled later have similar clinical/baseline characteristics
than those enrolled earlier.

You can replace Log-rank test by Breslow, because the last one gives
less weight to those time points where there are few individuals at risk
(higher lenghts of follow-up). Breslow test focuses on short follow-up
differences.

HTH,
Marta

--
For miscellaneous statistical stuff, visit:
http://gjyp.nl/marta/

=====================
To manage your subscription to SPSSX-L, send a message to
[hidden email] (not to SPSSX-L), with no body text except the
command. To leave the list, send the command
SIGNOFF SPSSX-L
For a list of commands to manage subscriptions, send the command
INFO REFCARD