Hello all,

This isn't necessarily a Stata-specific question, but more of a general statistical question.

Suppose I want to study the effects of a drug (drug A) compared to another (drug B) on kidney function (eGFR) over time. In the ideal world, I would have participants randomized, and then measure the eGFR at pre-specified intervals after initiating treatment. The analysis is then straightforward with multilevel mixed effects models.

However, realistically some patients may die before all measurements were obtained -- in such case, mortality represents sort of a 'competing event' and prevents further measurements from being obtained. This is intuitively important as those with worse eGFR may be expected to die more quickly / easily, so those with meausrements later on in the study will invariably be those who are less sick -- skewing the study population therefore.

My question is then obvious -- is there any statistical method that can account for mortality / any 'competing event' in this case? I have some knowledge of competing risk analysis in survival analysis, but do not know of any model that can deal with the above scenario. And if such models do exist, I would be grateful for any Stata implementation as well.

Thank you very much!