I would like to analyse the effect of multiple variable (age, sex, tumour grade etc) to overall survival of patients stratified by expression of a particular gene. Do I need to place all the variables in the covariate lists to perform multivariate cox proportional hazard analysis in SPSS? I highly appreciate your suggestions.
Hi Abdullah,
There are some good you tube videos on this.
Try this one: https://www.youtube.com/watch?v=wTLsw-Ckfvw
Good luck!
Hi, Abdullah,
I'd like to make certain that I understand your question. Multivariate designs involve more than one outcome, and multivariable designs involve more than one predictor: do you intend to simultaneously develop models for two or more genes using a single multivariable battery (I am unaware of this capability), or to separately develop a model for a single gene (per analysis) using a multivariable battery?
If your question pertains to the latter, then yes, you need to place all the predictor variables in the covariate list to perform Cox modelling.
There is a new, much more powerful method than Cox for identifying all statistically viable time-to-event models which exist for the sample, which explicitly maximize classification accuracy--and require no distributional assumptions. The following articles explain the new methodology--which can only be conducted using CTA software.
- Linden, A., & Yarnold, P.R. (2017). Modeling time-to-event (survival) data using classification tree analysis. Journal of Evaluation in Clinical Practice, 23, 1299-1308.
- Linden, A., & Yarnold, P.R. (2017). Using classification tree analysis to generate propensity score weights. Journal of Evaluation in Clinical Practice, 23, 703-712.
- Linden, A., & Yarnold, P.R. (2018). Estimating causal effects for survival (time-to-event) outcomes by combining classification tree analysis and propensity score weighting. Journal of Evaluation in Clinical Practice, 24, 380-387.
- Badges
- Science topic
- Similar topics
- Mathematical Sciences
- Statistics
- Statistical Software
- SPSS