Cox regression

[Irene Boateng]

Hello I am doing a cox regression analysis with ADR as my dependent variable and age, gender, ethnicity, previous hospitalisation, previous falls and expdose1 as my independent variable. Expdose1 is a categorical data with values of 1 to 10. The rest are indicator variables with values of 0 or 1. I am getting an error message coded as r(430) and it also says there is a flat region resulting in a missing likelihood. How do I rectify this?

[Carlo Lazzaro]

Irene:
- it's a good habit to post what you've typed and what Stata gave you back (as per FAQ);
- in all likelihood (pun intended) the MLE cannot proceed any further with the maiximization. Usually, it means that you have to reconsider your model specification, by adding one predictor at time and see when Stata starts choking.

[Marcos Almeida]

Please follow the FAQ advice. The best approach is informing exactly what you typed, as well as the output. A short example with an abridged data (or a toy example) with do just fine. You can use - dataex - and code delimiters for the matter.