Dear All,
I am using prescription drug claims data and want to identify spells of 3 months of continuous drug use. The individual level panel is unbalanced; the individual shows up in the panel only whenever a prescription is filled. Here is an example:

Code:




Code:


* Example generated by -dataex-. To install: ssc install dataex
clear
input long patient_id str10 fill_date int number_of_days_supply
1 "2012-06-15"  3
1 "2012-06-21" 10
1 "2012-06-21" 10
1 "2012-06-21" 10
1 "2018-04-13" 10
2 "2011-10-02"  6
2 "2012-01-14"  3
2 "2012-02-28"  6
2 "2012-02-28"  6
2 "2012-02-28"  6
2 "2012-02-28"  6
2 "2012-02-28"  6
2 "2012-02-28"  6
2 "2012-02-28"  6
2 "2012-02-28"  6
2 "2012-02-28"  6
2 "2012-02-28"  6
2 "2012-02-28"  6
2 "2012-02-28"  6
2 "2012-02-28"  6
2 "2012-02-28"  6
2 "2012-04-17"  1
2 "2012-04-20"  4
2 "2012-04-22"  6
3 "2012-04-06"  2
3 "2012-04-16"  5
3 "2018-05-06"  7
3 "2018-05-26" 30
3 "2018-07-03" 30
3 "2018-08-14" 30
3 "2018-09-17" 30
3 "2018-09-17" 30
3 "2018-09-17" 30
3 "2018-09-17" 30
3 "2018-09-17" 30
3 "2018-09-17" 30
3 "2018-09-17" 30
3 "2018-09-17" 30
3 "2018-09-17" 30
3 "2018-09-17" 30
3 "2018-09-17" 30
3 "2018-09-17" 30
3 "2018-09-17" 30
3 "2018-09-17" 30
3 "2018-09-17" 30
3 "2018-09-17" 30
3 "2018-09-17" 30
3 "2018-09-17" 30
3 "2018-09-17" 30
3 "2018-09-17" 30
3 "2018-09-17" 30
3 "2018-09-17" 30
3 "2018-09-17" 30
3 "2018-09-17" 30
3 "2018-09-17" 30
3 "2018-09-17" 30
3 "2018-09-17" 30
3 "2018-09-17" 30
3 "2018-09-17" 30
3 "2018-09-17" 30
3 "2018-09-17" 30
3 "2018-09-17" 30
3 "2018-09-17" 30
3 "2018-09-17" 30
3 "2018-09-17" 30
3 "2018-09-17" 30
3 "2018-09-17" 30
3 "2018-09-17" 30
3 "2018-09-17" 30
3 "2018-09-17" 30
3 "2018-09-17" 30
3 "2018-09-17" 30
3 "2018-09-17" 30
3 "2018-09-17" 30
3 "2018-09-17" 30
3 "2018-09-17" 30
3 "2018-09-17" 30
3 "2018-09-17" 30
3 "2018-09-17" 30
3 "2018-09-17" 30
3 "2018-09-17" 30
3 "2018-09-17" 30
3 "2018-09-17" 30
3 "2018-09-17" 30
3 "2018-09-17" 30
3 "2018-09-17" 30
3 "2018-09-17" 30
3 "2018-09-17" 30
3 "2018-09-17" 30
3 "2018-09-17" 30
3 "2018-09-17" 30
3 "2018-09-17" 30
3 "2018-09-17" 30
3 "2018-09-17" 30
3 "2018-09-17" 30
3 "2018-09-17" 30
3 "2018-10-18" 30
3 "2018-11-20" 30
3 "2018-12-24" 30
4 "2010-09-04"  2
end



For each patient_id, using date of fill and days of supply I want to identify spells of 90 days of continous use i.e. the individual had 90+ days of supply in the 90 day window. How can I do that?

Many thanks.

Sumedha.