Hello  I have a dataset with patients that have different therapies. In the therapies there are multiple medications. It all is in a long format.    It is in this format:    data your_dataset; input PatientID TherapyID Drug $; datalines; 1 1 DrugA 1 1 DrugB 1 1 DrugC 1 2 DrugA 1 2 DrugB 1 2 DrugC 1 3 DrugB 1 3 DrugC 1 4 DrugA 1 4 DrugC 1 4 DrugB 1 4 DrugD 2 5 DrugA 2 5 DrugD 2 7 DrugA 2 7 DrugC 3 6 DrugB 3 6 DrugE ;     I would like to identify, which drugs are newly started or finished within a patient, when they have a new therapy. And flag them accordingly.   Solution example:  1 1 DrugA baseline 1 1 DrugB baseline 1 1 DrugC baseline 1 2 DrugA last_time 1 2 DrugB no_change 1 2 DrugC no_change 1 3 DrugB no_change 1 3 DrugC no_change 1 4 DrugA restart 1 4 DrugC no_change 1 4 DrugB no_change 1 4 DrugD new_start 2 5 DrugA baseline 2 5 DrugD last_time 2 7 DrugA no_change 2 7 DrugC new_start 3 6 DrugB baseline 3 6 DrugE baseline   They could also have variables added with a binary value to it.    So far, I have used this code, but there are multiple errors in it.   data drug_changes; set your_dataset; by PatientID; retain PrevPatientID PrevTherapyID FirstTherapyDrugs; if first.PatientID then do; PrevPatientID = .; PrevTherapyID = .; FirstTherapyDrugs = ''; end; if PatientID ne PrevPatientID then do; PrevTherapyID = .; FirstTherapyDrugs = ''; end; if TherapyID ne PrevTherapyID then do; if Drug ne '' then do; if PrevTherapyID ne . then do; action = 'Finished'; output; end; if TherapyID = 1 then do; /* Check if it's the first therapy */ FirstTherapyDrugs = catx('|', FirstTherapyDrugs, Drug); /* Concatenate drugs */ action = 'First Start'; output; end; else do; action = 'Started'; output; end; end; end; PrevPatientID = PatientID; PrevTherapyID = TherapyID; if last.PatientID then do; if FirstTherapyDrugs ne '' then do; do i = 1 to countw(FirstTherapyDrugs, '|'); Drug = scan(FirstTherapyDrugs, i, '|'); action = 'First Start'; output; end; end; if Drug ne '' then do; if TherapyID = 1 then do; /* Check if it's the first therapy */ FirstTherapyDrugs = catx('|', FirstTherapyDrugs, Drug); /* Concatenate drugs */ action = 'First Start'; output; end; else do; action = 'Started'; output; end; end; end; drop PrevPatientID PrevTherapyID FirstTherapyDrugs i; run; proc print data=drug_changes; run; Thank you so much for your help in advance.  Nina ... View more

data example1; input Patient $ Testname $ Testdate MMDDYY8. claim $ Drugclaimdate MMDDYY8.; cards; patienta a 03/01/20 ' ' . patienta ' ' . druga 03/02/20 patienta ' ' . druga 03/28/20 patienta ' ' . druga 09/15/20 patientb a 03/01/20 ' ' . patientb b 06/01/20 ' ' . patientb ' ' . druga 02/28/20 patientb ' ' . drugb 03/05/20 patientb ' ' . druga 04/25/20 patientb ' ' . drugb 06/17/20 patientb ' ' . druga 07/05/20 patientb ' ' . druga 09/18/20 patientc ' ' . druga 03/01/20 patientc ' ' . drugb 04/01/20 patientc ' ' . drugc 05/01/20 ; run; Hello dear SAS community    I need some help with the following question. I have a dataset with patients and their drug claims. Moreover, in this dataset are certain laboratory tests of the patients. I would like to make a variable interval30 with the values    Interval30=1 if the drug claim is within 30 days before a test of this patient Interval30=2 if the drug claim is within 30 days after a test of this patient Interval30=0 if no drug claim is within 30 days before or after a test of this patient.   One person can have only lab test, only drug claims or both (in any number).   The resulting dataset should look like the following.    Patient Testname Testdate ddmmyy Drugclaim Drugclaimdate ddmmyy Interval30 patienta a 01/03/20       patienta     Druga 02/03/20 2 patienta     Druga 28/03/20 2 patienta     Druga 15/09/20 0 patientb a 01/03/20       patientb b 01/06/20       patientb     Druga 28/02/20 1 patientb     drugb 05/03/20 2 patientb     Druga 25/04/20 0 patientb     drugb 17/06/20 2 patientb     Druga 05/07/20 0 patientb     Druga 18/09/20 0 patientc     Druga 01/03/20 0 patientc     Drugb 01/04/20 0 patientc     drugc 01/05/20 0   Thank you for your help! ... View more