I'm a bit puzzled about competing risks. I've used competing risks methods (gray's test and fine and gray regression) a while ago, but can't seem to make it work this time.

I'm using SAS macro's to calculate cumulative incidence (CI) in a small dataset.
I've created a fake dataset to test the macros, which contains a time variable, a status variable with 0 being censored observations, 1 being the event of interest and 2 being the competing event. I've got 12 censored observations and each event counts 19 observations (total dataset=50obs).

Now, using the CumInc macro (http://www.uhnres.utoronto.ca/labs/h...ros/cuminc.txt) the output says that the sum of the CI of each event is 1.000. I don't understand how this can be 1, since I already have censored observations?
When I do the same in a dataset which has a larger percentage of censored observations, it does not add up to 1. Also, I tested if the CI of both events combined in one variable is the same as the kaplan-meier of both events, which did match.

Can anyone explain this to me? Thanks for any suggestions that could make this a little more clear for me!