Medication Discontinuation

[macro_2349]

Hello,
This is the data I have:

id	study_strt	study_end	drugname	drugtype	rx_date	days_sply	drug_start	rx_end
1	2008-09-26	2010-09-26	DrugA	2	2008-10-13	30	17	2008-11-12
1	2008-09-26	2010-09-26	DrugA	2	2008-10-23	30	27	2008-11-22
2	2012-08-23	2014-08-23	DrugB	1	2012-08-25	45	2	2012-10-09
2	2012-08-23	2014-08-23	DrugA	2	2012-08-25	45	2	2012-10-09

Data I want would look like this:

id	study_strt	study_end	drugname	drugtype	rx_date	days_sply	drug_start	rx_end	discontinue
1	2008-09-26	2010-09-26	DrugA	2	2008-10-13	30	17	2008-11-12	0
1	2008-09-26	2010-09-26	DrugA	2	2008-10-23	30	27	2008-11-22	0
2	2012-08-23	2014-08-23	DrugB	1	2012-08-25	45	2	2012-10-09	0
2	2012-08-23	2014-08-23	DrugA	2	2012-08-25	45	2	2012-10-09	1

Ideally I would like to keep my data in long form while identifying different medication treatment patterns. Medication discontinuation is defined as a gap of 30 days or more within the study period where there is no record of prescription. Any tips on how I can create an dichotomous variable called discontinue, where 0 = no discontinuation and 1 = discontinuation.

rx_date = date prescription is filled
drug_start = day in study period in which prescription was filled

Many thanks in advance!

[ballardw]

Please state explicitly the rule(s) involved in creating that dichotomous variable Discontinue.

It helps greatly to be consistent in code and data description. Your "have" data does not have a variable named "drugname" but that is used in your code.

Examples are nice but do not necessarily provide rules to handle cases other than what is shown.

Are you assuming the rx_date is the date a prescription is filled? What exactly does "drug_start" represent?

I am not clear exactly what you mean by " modify to create dichotomous indicator variable (0/1) rather than output dataset?".

[macro_2349]

Thank you. I have modified my original post to answer your queries. Please let me know if you require further clarification.