I tested the code and I get the following error message: 'Variable first_date is uninitialized.' Can you maybe help me with that?   ... View more

Thank you, it seems that it is working. ... View more

Thank you, I will check if it is working for scenario I. I will also try to modify it to facilitate scenario II but any ideas there are welcome too.   Thanks again! ... View more

Regarding your first question, I mean 365 days after the first date (and not calendar year). Yes, you are right it should be summarized as 1 4/5/2005 5/20/2005 C,B. I can understand the challenge in using data summarized like this but my question relates to how doing so. ... View more

Here is what was included previously in the attachment!   Hypothetical dataset:                                     ID            Date      Diagnosis             1              9/1/2000              A             1              9/15/2000            A             1              11/3/2000            B             1              3/5/2001              A             1              4/5/2005              C             1              5/20/2005            B             2              6/6/2003              A             2              1/3/2005              A             2              2/7/2005              C             2              8/10/2012            C                                                                                                             Scenario 1:          For each person, group together diagnoses within 1 year (365 days)                         Output dataset should look like:                                ID            First date             Last date              Diagnosis 1              9/1/2000              3/5/2001              A,A,B,A 1              4/5/2005                              C 2              6/6/2003                              A 2              1/3/2005              2/7/2005              A,C 2              8/10/2012                            C                                                                                                 Secnario 2:          For each person, group together diagnoses allowing for a washout period of at least 5 years between consecutive dates                            Output dataset should look like:                                ID            First date             Last date              Diagnosis 1              9/1/2000              5/20/2005            A,A,B,A,C,B 2              6/6/2003              2/7/2005              A,A,C 2              8/10/2012                            C ... View more