Calculation for days of drug usage for one year, but not exceed 120 days between two prescription

[elvics]

Dear all,

I want to calculate the regular use of drug. The defination is that regular users was the regular use of drug for "more than a year" and "the interval between successive prescription drug records cannot exceed 120 days".

I wish the the data was presented as "Category (Correct)".

When the date between two prescriptions were within 120 days (Dif_Days), they belonged the same group.

If the date long than 120 days (Dif_Days), the next prescriptions was categorized as new group.

At one of steps, I cannot run the the correct category. How cou;d I modify the syntax?

DATA Final;
SET data;
by ID Drug_start
if first.ID then Category=1;
if Dif_Days<=120 then Category=1;
else Category+1;
run;

The data was shown as below.

Drug_start: the first date for this prescription

Drug_end:  the end date for this prescription

Drug_day: how long for this prescription

Dif_Days: the difference between the Drug_start and last Drug_end

ID	Drug_start	Drug_end	Drug_day	Dif_Days	Category	Category (Correct)
AAA	20091011	20091109	30		1	1
AAA	20091109	20091208	30	0	1	1
AAA	20110317	20110415	30	464	2	2
AAA	20110523	20110621	30	38	1	2
AAA	20110810	20110908	30	50	1	2
AAA	20110909	20111008	30	1	1	2
AAA	20111005	20111103	30	-3	1	2
AAA	20130217	20130318	30	472	2	3
AAA	20130313	20130411	30	-5	1	3
BBB					1	1
BBB					1	1
CCC	20080605	20080618	14		1	1
CCC	20100311	20100317	7	631	2	2
CCC	20100318	20100331	14	1	1	2
CCC	20110616	20110713	28	442	2	3
CCC	20111124	20111221	28	134	3	4
CCC	20121115	20121212	28	330	4	5
CCC	20121207	20130103	28	-5	1	5
CCC	20130110	20130206	28	7	1	5
CCC	20130418	20130515	28	71	1	5
CCC	20130530	20130626	28	15	1	5
CCC	20130706	20130802	28	10	1	5
CCC	20130727	20130823	28	-6	1	5

[ballardw]

First thing: Are your date variables SAS date valued and currently assigned a format of yymmdd8., character or numeric as shown? If they are not SAS you will want to create SAS date values. The approach differs depending on the variable type of your current values.

Then the value of Dif_days can be calculated in a datasetp as:

Dif_days = SASstartdate - lag(SASenddate);  (or do the conversion on the fly with the Lag value of date_end)

You will want to NOT calculate the dif_days this way for the first.ID

[Ksharp · Accepted solution]

I don't understand.

CODE NOT TEST.

DATA Final;
SET data;
by ID ;
if first.ID then Category=0;
if Dif_Days>120 or first.ID then Category+1;
run;

[elvics]

Thanks at lot.

It works.

But it seems need to be sort by ID and Drug_start.

DATA Final;
SET data;
by ID Drug_start;
if first.ID then Category=0;
if Dif_Days>120 or first.ID then Category+1;
run;