AGREETAB MACRO
 8 November 2001
 Teresa Mayer
 CancerCare Manitoba
 (204) 787-4054

 The agreetab macro creates two-by-two tables and the following agreement statistics:
 * sensitivity
 * specificity
 * positive predictive value (PPV)
 * negative predictive value (NPV)
 * concordance
 * Youden's index
 * Kappa statistic and p-value
 * McNemar's test statistic and p-value

 %_agreetab (byvar= ,na= ,var1= ,var2= ,outdat= ,indat= , html=, htmldir=, print=,
            addvar= ,iterat= ,sens= ,spec= ,PPV= ,NPV= ,debug= );


Call: 
   _agreetab <options> ;


Options:
   byvar  = at this point, byvar must be supplied.  This is the variable that results
            will be stratified by.

   na     = This must also be specified at this point.  This is used to remove missing
            values.  All missing values must be LESS THAN this value.

   var1   = This must be specified.
            This is the first variable in the two-by-two table.  If one variable is
            more precise than the other, then this variable is the least precise of the two.
            default = selfmam.  Valid values are 1=yes, 2=no.  (If 0=no, then this
            macro will reset var1=2)  So, valid values are actually 1=yes, 0,2=no.

   var2   = This must be specified.
            This is the second variable in the two-by-two table, and should be the most
            precise variable.
            default = clmam  Valid values are 1=yes, 2=no.  If 0=no, then this macro
            will reset var2=2).   So, really valid values are 1=yes, 0,2=no.

   outdat = This is the name of the output dataset.
            default = &byvar

   indat  = This is the name of the input dataset.
            default = last dataset used in SAS program.

   addvar = If you want to add in another variable to the output dataset, specify the variable
            name here.  Default is blank; no other variables will be put into output dataset.

   interat= Ask Teresa what this is for.  It is for doing iterative runs of this macro.  And it
            designates which iteration (or "i" value).  MUST be specified if addvar is used.

   html   = yes if you want html output.  no if you do not want html output.  default=no.

   htmldir= This is the name of the directory where the html files go.  This must be specified
            if html=yes.

   print  = Print all agree statistics on the .lst file.  The statistics printed on the .lst are
            listed above (e.g. Kappa, concordance, etc.)
            Valid values are yes or no.  default=yes.

   debug  = Set to debug to turn on notes.  Default=nodebug.

Notes:
   The byvar= variable and na= options must be set.  If you only have one set of data then
   set a variable to a single value in the datastep prior to the macro call.  Na= should be
   larger than this value (e.g. all=1  then inthe call na=9. 

Example:

data final ;
    set survey ;
    all=1 ;
    run;

agreetab
    byvar=all
    na = 2
    var1=diabete
    var2=admin_diab
    outdat=printit
    indat=final ;