Open rpietro opened 12 years ago
the issue is that in SEER-Medicare, the ethnicity (hispanic - non hispanic) is coded independent of race (white, black,...) while in NIS, hispanic is included in the variable race as an independent "outcome". I think we have to recode ethnicity in SEER-Medicare together with race (if hispanic==YES, then hispanic, if hispanic==NO and race==white - then white,...)
does this make sense?
we don't have the variable for hispanics and without that i can't align the two datasets: single variable race in NIS contains whether people are or not hispanic, while for SEER there are two variables. is that something you could talk with Igor?