sampling

["Ngai, Christopher" <Christopher.Ngai@moh.gov.on.ca>]

Hi all,

 

I'm working with a large drug utilization database (>600M records). For a specific drug (dosage, form), I have one entry for each prescription filled by a patient. The basic layout of the database is: drug_id, user_id, date_filled, days_to_next_refill, quantity. Quantity is in the smallest unit, e.g. tablet. Days_to_next_refill is the number of days between the current and next prescription filled date.

 

Now, I want to know how many patients take a specific drug 1 tablet a day, 2 tablets a day, etc. I do not want to go through the whole database. I would like to utilize repeated sampling to get a distribution for percent of patients taking 1tab/day, 2tab/day, etc.

 

Can anyone suggest an approach using S-Plus? Thanks.

- - - - + - - - - + - - - - + - - - - + - - - - + - - - - + - - - - +
Christopher Ngai