Tell me more ×
Cross Validated is a question and answer site for statisticians, data analysts, data miners and data visualization experts. It's 100% free, no registration required.
up vote 0 down vote favorite

I am trying to group procedure codes so that I do not need to hard code thousands of procedure codes and diagnosis codes. I am trying to build a regression to see the effect of percentage of dollars for a provider on total billed amount for a provider on several parameters so that I can find the outliers. The parameters are procedure codes, diagnosis codes, place of service etc.

Any suggestions on grouping procedure codes and diagnosis to see the effect on total billed amount for a provider would be highly appreciated.

share|improve this question
Can you give some more detail? In what ways do the procedure and diagnosis codes vary? How many procedure and diagnosis code are there? If you have coded them in the past, how many categories did you code them into? Are those categories OK or are you looking for other groupings? Also, what are the predictors? You mentioned both procedure and diagnosis codes and percentage of dollars for a provider as predictors. Is percentage of dollars part of your question or did you mention it just to provide some context? You also mentioned predicting "several parameters". What are those parameters? – Joel W. Aug 7 '12 at 15:19
@ Joel Thank you .proc sql;CREATE TABLE PROCCD AS select provid,proccd, sum(reimbamt) as total,when proccd between '90801' and '90804' then 'proc61'when proccd between '90806' and '90809' then total is the total reimbursement amount for a provider. I am determining what percentage of total for a provider is on a certain procedure code group. My groupings did not show any pattern. I was thinking as the percentage (predictor) increases the total (response) would also increase. This way I can find the outliers. I think the grouping should be based on the expensive and cheap procedure codes. – Kuusum Aug 7 '12 at 16:22
My independent variables are percentage of business dollars for a procedure code group, for diagnosis code group, for place of service, gender of recipient, age above 65 and below 65, count of recipients for a provider, count of procedures for a provider. Log of totalcharge is my dependent variable. – Kuusum Aug 7 '12 at 16:39
How many procedure and diagnosis codes do you have? On what basis do you want to group them? Based on cost? Based on interrelationships of the procedures and diagnoses? Other? Are there logical relationships between procedures and diagnoses? If so, do you have a way of describing the relationships? Can you describe your data a bit more? – Joel W. Aug 7 '12 at 17:22
I have around 14000 diagnosis codes and 13000 procedure codes. I want to group them based on how expensive each procedure is but I have no way to find them. I was trying to find any literature that will let me classify them based on how expensive or cheap they are. There are logical relationship between diagnosis codes and procedure codes but thre is no way to define them based on costs. – Kuusum Aug 7 '12 at 18:06
show 5 more comments

Know someone who can answer? Share a link to this question via email, Google+, Twitter, or Facebook.

Your Answer

 
discard

By posting your answer, you agree to the privacy policy and terms of service.

Browse other questions tagged or ask your own question.