hi @rch3k5
when you are instructed to set the patient who didn’t receive treatment, you are suppose to use .loc to the dataframe variable, so that it selects label classified variable related to the patients who didn’t receive treatment.
next when predicting the probability of death by including the input features into model, you are using incorrect indexing as no death is numbered as 0 and death is number as 1(refer the screenshot 
), but there is another slight mistake in that code.
I am attaching the hint screenshot, read the 3rd point again, you are asked to use a patient column with its respective death outcome. for example if patient colum mention as : and death as 1, then it should [:, 1] and not just [0]
Let me know if issue still persist.
regards
DP