0

I have a little knowledge of SAS but I have no idea about the structure of claims data .I want to select participants for a cohort study based on the following criteria from three data sets:

o Received a dementia diagnosis (ICD-9-CM: 290.0x - 290.4x, 294.1x, 331.0x, 331.1x, 331.2x, 331.82, and 331.9x in any position) during 2001-2005. The date when the patient first received a dementia diagnosis is defined as the index date. o ≥ age 65 on the index date o ≥12 months enrollment prior to the index date

files: File name 1: Inpatient_claims2001.csv - Inpatient_claims2005.csv

Data fields: patID admission_date discharge_date (e.g. 9/27/2002) Primary_diagnosis diagnosis_code1 – diagnosis_code6

File name 2: outpatient_claims2001.csv - outpatient_claims2005.csv

Data fields: patID service_date (e.g. 3/27/2002) diagnosis_code1 diagnosis_code2

File name 3 : enrollment_file2001.csv - enrollment_file2005.csv

Data fields: patID date_of_birth (e.g. 10/18/1945)
enroll_mon1 – enroll_mon12 (values: 0 or 1 with 0=not enrolled and 1=enrolled for the month)

If anyone have an experience in claims data please help me on how can I approach and code in SAS to get the requested results

Thanks, Ankur.

Ankur Rao
  • 1
  • Do you have your CSVs loaded in to SAS? If not, you're putting the cart before the horse here. I would start with what you have, how to get that into SAS and then make a small example in Excel. Show a base case of what you have and what you expect as output and you can start figuring out the steps in between. It looks like you'll need to figure out how to import multiple CSV files with the same structure, appends, joins/merges, working with SAS dates for starters. Or start by taking the free SAS training class. Your question isn't particularly difficult but it is some work. – Reeza Mar 22 '21 at 21:04
  • I dont have the CSV files this is actually a programming question/scenario that I have to answer by providing the SAS code.I already have the code for importing the file in SAS . my plan is to merge enrollment data with inpatient claims first and then use if and else statement to get the desired result.Repeat same with the merged outpatient and enrollment data .However, the problem is I dont know how to extract data from multiple variables – Ankur Rao Mar 22 '21 at 22:43
  • I am not sure how to extract data into one variable say " dementia" from six diagnosis code (inpatient claims data) .Also, I am not sure which function to use to calculate age at index – Ankur Rao Mar 22 '21 at 22:48
  • You need to show how the codes are actually stored in your data. The claims data will have actual detailed codes. But your example list of codes for a disease include `x` character which is normally used to mean any of the possible subcodes. So you need to understand how to test that 290.01 and 290.02 are example of 290.0x. – Tom Mar 23 '21 at 13:16
  • Sounds like your problem is now simplified to two issues: how to calculate age and find dementia. Post each of those as new questions following the guidelines here [ask]. It's best if you follow the format of this is what I have, this is what I need as output showing examples, and preferably including what you've tried. – Reeza Mar 23 '21 at 15:39

0 Answers0