Grouping Data in Oracle

QUESTION

I am interested in running a DRG grouper against large (over a million records) data sets.  My Oracle data warehouse sits on a Solaris box, but I also have windows and linux servers if needed.

I simply want to assign DRGs to my data set.  So, I'm interested in information on how to do this and what product would be right.  It'd also like to get something in house to to ensure that we can actually use it.

ANSWER

If you are not a programmer, then the way to do this would be with DRGFilt:

1. Export your data as fixed-width text
2. Create a DRGFilt control file to match your export
3. Run your text file through DRGFilt
4. Import the DRGFilt output

DRGFilt outputs all the input data, plus the DRG (and MDC and whatever else you configured), so you can import the DRGFilt and do whatever you like with it.

If you are a programmer, you can get an embeddable DRG Assignment product and call it from a program. Supported environments are Windows or most Unixoid systems. Supported runtime environments are Perl, Visual BASIC, PHP and C. A typical program will do the following:

1. Select rows out of your database
2. Process each row one at a time
3. Extract the DAE inputs from the row
4. Call the DAE
5. Parse out the DAE outputs from the call
6. Either update the original row or create a new row

Finally, if you do not want to do the work yourself, or do not want to deal with the version question (ie, which version DRGs do I use for this given record?) you can use our DRG Assignment Service (DAS) to do the work. In that case, the steps are these:

1. Extract your data, without PHI
2. Send us the extract, along with your preferences for the returned data set
3. Usually in one, but sometimes in up to three business days, you get the data
4. Import or process the data set returned by us