Hi,
I am looking at a longitudinal study of two ethnic populations. they were first seen/recruited at time T1. The total number of participants was 30k. out of these 15k were then followed up at T2. however this T2 is not same for all participants, like some participants were seen after 5 years, some after 6, some after 7 and so on. My aim is to look at the progression of chronic disease in these two populations and compare the results. I want to conduct survival anaylsis but because of the variable follow-up time I don't know how to proceed.
Can you please help me understand how to apply the tools for survival analysis and cox regression in this case?/
Kind regards
Saima
you may not need survival analysis if you can score disease severity.
A very simple initial model would be to assume the chronic disease severity score , (including death as highest score) increases( or decreases) with time.
then a simple regression (initially assume linear) of disease score on time and a binary (0-1) variable for ethnicity would estimate difference in rate of progession between ethnic groups.
eg Disease score = beta0 + beta1 x ethnicity (1/0) + beta2 x time of follow up.
Hi,
Please take a look at the MedCalc Help and the SPSS, are well developed, have examples.
Look also here:
Article Survival analysis: Part 15 of a series on evaluation of scie...
https://www.medcalc.org/manual/survival-analysis.php
Success
M. Kaneva
If your outcome is a dichotomous time-to-event, then this sounds to me like a life table analysis. If the exact date of the outcomes were known, then survival analysis could be use. However, if the exact dates are not known and only measured at specific times (e.g., once every year), then life table analysis would be the choice. As noted by James (above), if your outcome is not dichotomous, then other approaches should be explored. Be careful, though -- simple regression approaches do not account for censoring.
Paul
Thank you very much all for you help.
@James so basically I am looking at kidney disease (by showing the decline in GFR from baseline). Disease score will then be based on the GFR at follow-up or the difference in GFR between the baseline and follow-up?
@Paul. I am looking at GFR change (continous) as well as the CKD stage (dichotomous). I don't have the exact date of outcome. so I will explore the Life tables, thank you so much for the suggestion.
Regards
Saima