https://communities.sas.com/t5/SAS-Health-and-Life-Sciences/grouping-patient-diagnoses/m-p/13191#M890 I need to figure out the earliest diagnosis that a patient has given that each patient has many encounters and 10 diagnosis codes for each encounter. The catch is that we only want to look at certain diagnosis codes. Does anyone have any ideas? I already sorted by PatientID and encdate. Just looking at diag1 isn't too difficult. It's after that things get sticky <span class="lia-unicode-emoji" title=":neutral_face:">😐</span> Thu, 29 Apr 2010 21:34:11 GMT marilyn934 2010-04-29T21:34:11Z https://communities.sas.com/t5/SAS-Health-and-Life-Sciences/grouping-patient-diagnoses/m-p/13191#M890 I need to figure out the earliest diagnosis that a patient has given that each patient has many encounters and 10 diagnosis codes for each encounter. The catch is that we only want to look at certain diagnosis codes. Does anyone have any ideas? I already sorted by PatientID and encdate. Just looking at diag1 isn't too difficult. It's after that things get sticky <span class="lia-unicode-emoji" title=":neutral_face:">😐</span> Thu, 29 Apr 2010 21:34:11 GMT https://communities.sas.com/t5/SAS-Health-and-Life-Sciences/grouping-patient-diagnoses/m-p/13191#M890 marilyn934 2010-04-29T21:34:11Z https://communities.sas.com/t5/SAS-Health-and-Life-Sciences/grouping-patient-diagnoses/m-p/13192#M891 use PROC TRANSPOSE to convert from one row per encounter to one row per diagnosis. then you can treat it as you would diag1. Wed, 14 Jul 2010 16:40:34 GMT https://communities.sas.com/t5/SAS-Health-and-Life-Sciences/grouping-patient-diagnoses/m-p/13192#M891 Doc_Duke 2010-07-14T16:40:34Z