Hi - sounds like you need a "counting process" dataset as input for PHREG.  Are you still looking for a way to create this?  I have a macro for it - let me know. ... View more

This is even better than I hoped, thank you so much! ... View more

I would argue you actually don't want your data in your binary column arrangement. Add a new drug and you have to add a new column.      I would argue more normalized data like:   ID Drug Date or dosage or whatever 1 Ceftazidme   1 Ciprofloxacin   2 Ceftazidme   2 Colistin   2 Vancomycin   3 Cefazolin   3 Ceftazidme   3 Meropenem   4 Minocycline   4 Colistin     which is essentially the format of the TEMP table in the @PeterClemmensen code would you do you more favors in the long run in terms of querying.    proc sql; select drug, count(id) as "Number of IDs Taking Drug"n from temp group by drug; quit; drug Number of IDs Taking Drug Cefazolin 1 Cefepime 1 Ceftazidme 3 Ciprofloxacin 1 Colistin 2 Meropenem 1 Minocycline 1 Vancomycin 1 ... View more

Have you attempted looping a lag function over the id? ... View more

proc genmod data=allthree order=formatted; class hospital prepost cohort/ desc param=ref; model cost=prepost cohort prepost*cohort covariate1 covariate2/link=log dist=gamma;  repeated subject=hospital/type=ind; lsmeans prepost cohort prepost*cohort/ilink diff; slice prepost*cohort/nof ilink sliceby=cohort diff; /* This will give a test of pre vs post for each level of cohort */ run;   Using the code above, How do you determine whether the 1) pre mean for all 4 cohorts values are significantly different (Test  whether  pre lsMean  in cohort1= pre lsMean  in cohort 2= pre lsMean  in cohort 3 = pre lsMean  cohort 4), post lsmean values  for all cohorts are significantly different  (Test  whether  post lsMean  in cohort1= post lsMean  in cohort 2= post lsMean  in cohort 3 = post lsMean  cohort 4and the overall differences between co pre vs post for each level of cohort  (Test  whether Mean diff  in cohort1=mean difference in cohort 2=mean difference in cohort 3 =mean difference cohort 4 ) ... View more

Glad to hear you solved your problem.  However, for others facing a similar situation, both Tom's and my suggested code were designed to concatenate, not merge, the 3 files.  I really liked Tom's suggested method. My guess, as to why it didn't initially work, was that the files weren't all first sorted by id. However, the method you selected definitely accomplishes the task, too, and only required the one sort (I presume that the idfile was already sorted). ... View more