SAS

[Polly Nomial]

I know nothing about SAS but have been asked what we might use in our department.

I am wondering how helpful SAS could be for health/HMO.

1. Is it mostly good for creating good looking studies or are the procedures valuable for getting at things that would be otherwise difficult?

2. Does anyone use SAS? For what?

Any info would be appreciated!

[Dr T Non-Fan]

I've found it to be an allocation-burner on mainframe. Of course, I was using very large files and summing them and creating a large summary file. I fixed my way of doing things.
I used it mainly to extract data to download to my PC for use in Excel for reports and such.
All that fancy stat stuff ain't worth much. Not to me.

[Need_A_New_Career]

From what I understand, SAS is very good at performing statistical analyses on large data sets. However, like DTNF, I have mainly used it for gathering data.

My current company still uses mainframe SAS, but we are migrating to a data warehouse environment and have been given SQL and Crystal. I absolutely hate working in a mainframe environment, so I'm focusing on the new tools.

[semper augustus]

You can use SAS for a variety of things in the insurance world, it's really up to how much you like it (and become a SAS geek).

A couple of common uses are:
Experience/Lapse/Mortality/Morbidity studies
Report generation

SAS can handle vast amounts of data, you are able to use it's database like features along with normal APL/C++ programming functionality.

I have read in data with 15 million records (and 20+ variables each) without a problem (albeit, it's not that fast with that much data!)

The windows version is easy to understand and pick up.

On the other end of the spectrum if you really get into the language there would be nothing to stop you from programming illustration/reserve routines, with GUI dialogs boxes, etc.

[josie]

We use SAS to develop our completion/lag factors for our reserving/IBNR process. The SAS program quickly calculates the base completion factors for at least 50 buckets. We used to have this process in Excel, but as our business expanded, the Excel sheets became very cumbersome. SAS allowed us to automate this triangle process. We do it in SAS with arrays and the results spit out in very little time.

[Miguel on the Beach]

Our Actuarial Dept uses SAS quite a bit. We mainly use it to extract data, and sort, merge, filter, and summarize large volumes of claims, premium and membership data on the mainframe before downloading it to Excel. We also have some SAS data modeling programs that evaluate health plan benefit relativities.
The statistical procedures and functions are almost never used.

We are now moving to SAS in the UNIX environment, switching from running on the MF. I myself generally only SAS occasionally to do ad-hoc reports to price odd benefits where I need to get down to specific CPT or DRG codes - data that is not available on our normal experience reports. But, we have a pretty extensive IW where I can easily pull a lot of data using Clear Access, so that is what I usually use to do my work.