Proc Tabulate help: multiple analysis variables

[ColeG]

I am creating a retention report to show the number of participants retained in research procedures month over month. In this report there are rows for each site and columns for each month. Within each row is a "sub row" for "Total Enrolled," "N expected," (people we expect back within each time period; changes since not everyone is enrolled in the same month) "N retained," (number who came back for their expected procedures) and "% retained."

Since "Total Enrolled" is not time-dependent, it would be cleaner to have enrolled as a stand-alone column (one more column total, one fewer row per site) but when I try to add it to the columns dimension much as I would with the ALL option I get the error "There are multiple analysis variables associated with a single table cell in the following nesting: site*f*____*enrolled*sum*s" where "_____" is  expected, retainedN, or retainedPCT (three total errors).

I cannot use the ALL option since the N expected will not add up to the total N enrolled (e.g., early censoring).

Hopefully this pastes, I'm trying to do something like this:

		N enrolled	Time 1	Time 2
Site 1	Sub Denom	100	92	67
	Sub Num		86	60
	%		93.5%	89.6%
Site 2	Sub Denom	150	104	74
	Sub Num		84	55
	%		80.8%	74.3%
Total	Sub Denom	250	196	141
	Sub Num		170	115
	%		86.7%	81.6%

Thanks for your help!

[ballardw]

If you don't want to nest variable then just place them together. If you want the same statistics for each group with parentheses.

Some example data so we have variable names and types would help. At least show your  current code.

Is "Total enrolled" supposed to equal the number records? The likely you just need N without a variable BUT you will likely get it repeated as it crosses with each row.

proc tabulate data=sashelp.class;
   class sex;
   var height weight;
   table sex,
         n='Number of obs' height*mean weight*max
         /
   ;
run;

or provide example data:

Instructions here: https://communities.sas.com/t5/SAS-Communities-Library/How-to-create-a-data-step-version-of-your-dat... will show how to turn an existing SAS data set into data step code that can be pasted into a forum code box using the {i} icon or attached as text to show exactly what you have and that we can test code against.

[ColeG]

Hi thank you for your reply.

I'm not sure what you mean by "If you don't want to nest variable then just place them together. If you want the same statistics for each group with parentheses."

Your suggestion is close to what I'm looking for but I am creating this from a summary table rather than an observation level one.

My data looks something like this, albeit with more rows:

data retainreport;
 infile datalines delimiter=',';
 input site visitcode enrolled expected retained retainedpct;
 datalines;
1,1,50,40,30,.75
1,2,50,30,20,.67
1,3,50,20,10,.5
1,4,50,10,0,0
 ;
run;

proc tabulate data=retainreport;
 class site visitcode / order=data;
	classlev site;
	classlev visitcode;
 var enrolled expected retained retainedpct;
 table (site='')*
	   (enrolled*f=best12. expected*f=best12. retained*f=best12. retainedpct*f=percent6.),
	   visitcode*(sum='');
run;