Hi guys, I would like to calculate p trends for my main exposures: cig, pipe and nass. They were originally numeric variables. I tried to use the numeric version in the Cox model but the result is either blank or "0". Then I tried to categorize these exposures into 4 levels, and I used the code below to calculate the trend, but SAS gave me this feedback. The instruction I used said there is a "ptrend" option in the Phreg function but when I entered it didn't turn blue as a valid option. Did anyone know how to fix the problem or how to calculate the p trend in the multivariable Cox model? Thank you.
73 proc phreg data=liverfp;
74 class cigd(ref='0') nassd(ref='0') piped(ref='0') everopi(ref='0')
75 Aspirin(ref='2') HCVAb(ref='0')
76 HBSAg(ref='0') DLiver(ref='2') Diab(ref='2') ALC(ref='2')
77 / param=ref;
78 model length*caseco(0,2)= cigd piped nassd everopi HBSAg HCVAb
79 DLiver Aspirin Diab Alc
80 / risklimits ties=efron ptrend;
______
22
202
ERROR 22-322: Syntax error, expecting one of the following: ;, ABSFCONV, ALPHA, BEST, CORRB, COVB, DETAILS, ENTRYTIME, FAILCODE,
FCONV, FIRTH, GCONV, HIERARCHY, INCLUDE, INITIAL, ITPRINT, MAXITER, MAXSTEP, NODESIGNPRINT, NODP, NODUMMY,
NODUMMYPRINT, NOFIT, OFFSET, PLCONV, RIDGING, RISKLIMITS, ROCLABEL, SELECTION, SEQUENTIAL, SINGULAR, SLENTRY, SLSTAY,
START, STOP, STOPRES, TIES, TYPE3, XCONV.
ERROR 202-322: The option or parameter is not recognized and will be ignored.
81 strata age1 gender ethnicity;
82 run;