I am trying to estimate life-year gained using stci commands in STATA. However, other causes of death and baseline characteristics may overestimate the life-year gained calculated.Does any one know how can I adjust for competing risk and patient characteristics using STATA.
Although I cannot be sure given the brevity of your question, it sounds like you should look at in STATA. There is a good introduction to the concept and the syntax in Cleves M, Gutierrez RG, W G, Marchenko YV. An Introduction to Survival Analysis Using Stata. 3rd Edition ed. College Station, Texas: Stata Press; 2010. You'll see that the method used (based on Fine & Gray) is a proportional hazards method (like the Cox model), and that it is a little more fiddly to check these assumptions using in STATA than with . That said, you'll get a sub-hazard ratio estimate, and using the base model curve (by using ) you should be able to estimate the life years gained (or lost) proportionally due to the variable of interest.
Thanks Paul.Thanks for your help. Do you mean that I can estimate the life years-gained using the area under the curve ?
Yes, the difference in the area under the curve for those with the variable of interest minus the AUC for those without the variable of interest using . While I'm not aware of an existing simple STATA command that does this (besides , which works out if the AUCs are different, not by how much) I feel confident you'll find an answer to this step, perhaps by checking for previous answers or asking for advise on Stata-List.
Is it possible to get stats.. relating to how much you die per day?
That's what I call fun & interesting.
You can do it using stplost, user-written stata command recently published that calculates pseudovalues (after that you must use glm with id link function), The bibliographic reference is:
Overgaard M, Andersen PK, Parner ET Regression analysis of censored data using pseudo-observations: An update. The Stata Journal (2015) Vol 15 Number 3 PP 809-821.
Thanks, immunity! Pathological; nutritional, ecocide, nano genetic, nero pathos... immunity, how much do we really control ? Exectively .
- Badges
- Science topic
- Similar topics
- Mathematics
- Statistics