home
***
CD-ROM
|
disk
|
FTP
|
other
***
search
/
CD Concept 6
/
CD Concept 06.iso
/
mac
/
UTILITAIRE
/
Survival 6.0.2
/
Macros
/
TimeMacro 1
< prev
Wrap
Text File
|
1995-03-18
|
3KB
|
49 lines
MACRO 'TimeDependent';
{this macro assumes that you selected VAR1 as time dependent covariate
and VAR2 as fixed covariate.Notice that rVar[] addresses to data in
the selected vars array while rData[] addresses to the data in the
active data array loaded with the Get Data option of the File Menu.
If the value of the variable in columm 2 of the data matrix is less
than 650 the selected VAR1 takes the value of zero (0);otherwise its
value is substracted from the average value for that variable. This
latter operation should be done for both, time dependent and fixed
covariate, to avoid internal overflow errors.}
NOTICE that the index in the rVar[] token is the order of selection
of the variable NOT the column number of that variable in the global
data array (rData[]).
This macro MUST be inserted in the Macros Menu before performing a
Time Dependent analysis. Otherwise you'll get an error message}
begin
{rVar[1] is selected as time dependent covariate (age at diagnosis)
survival times are in column 2 of the global data array (rData[2])
If survival times are equal or greater that 650 days, age takes its
value otherwise is set to zero. This is the general rule to follow
with time dependent analysis}
FUNCTION
If rData[2] >= 650 then rVar[1]:= (rVar[1] - rMean[1]) else rVar[1]:= 0;
rVar[2]:= (rVar[2] - rMean[2]); {Fixed covariate = grade}
end;
end;
{This is the same but with a different approach. Lets suppose you selected
Vars in columns 3,5 and 7 from your database as time dependent covariates
The order they are located in the selected variables array (rVar[])
are as follows: 1 for Var 3; 2 for Var 5; and 3 for Var 7
FUNCTION (we invoque the FUNCTION macro command) Then we start by setting all time dependent variables to zero:
rVar[1] := 0; rVar[2] := 0; rVar[3]:= 0;
Now we make the following conditional test
if rData[2] >=650 then (is the condition met at each failure time ?) begin
rVar[1]:= rData[3] - rMean[1];
rVar[2]:= rData[5] - rMean[2];
rVar[3]:= rData[7] - rMean[3];
end;
NOTICE that we substract the value of the variable from its average
value (rMean[]) to avoid owerflow problems.
END; (end of FUNCTION macro command) }
IF YOU UNDERSTAND THIS YOU'RE READY TO PERFORM TIME DEPENDENT ANALYSIS }