Carlo asked:
I beg Your pardon for the possible trivialism of what follows. I am intended
to compare two different health care programmes (A and B) which perform the
same in terms of overall average per patient survival (13 years each) but
give different results in terms of years patients (50,000 patients enrolled
in programme A and 50,000 in programme B) spend throughout different stages
(I^ II^ III^ and IV^) of the disease. Particularly, I would kindly ask You
about some hints to check for any statistical significance within each
stage. I tried tabi, chi2 exact with Stata 9, but I am not sure this the
right way to tackle this issue.
-----------------------------------------------------
This is far from trivial, and this type of question has lead health
economists, philosophers, epidemiologists, and biostatisticians to go war
against each other.
It looks like a QALY (Quality Adjusted Life Years) problem. To my knowledge
(and -findit-'s knowledge) no official or unoffical procedures deal with that.
Now, you ask about checking for statistical significance within each stage.
This requires data at the individual level with number of terminating events (stage shift, death) and time at risk. Look at survival analysis
(-st...- commands; the [ST] manual) and incidence rate ratios
(-ir-; also in [ST] under -epitab-).
Hope this helps (a bit)
Svend
