Thoughts on Approach -- Length of Stay Hospital Claims
3 posters
Page 1 of 1
Thoughts on Approach -- Length of Stay Hospital Claims
Design Question:
In the HealthCare Industry, multiple claims can make up one Length of Stay (LOS) or Episode Definition.
That calculation is easy. Created from FactPaid fact table
The harder part is the "best" way to capture information about said LOS.
For each LOS you need to capture info about pre and post services, insurance eligibility etc. (For those of you in the industry, HEDIS measures, etc)
Also info about Admission and Discharge (providers, diagnoses) along with in-between information.
One approach is an all encompassing factLOS table with all of the additional measures noted above.
Another is a bridge back to FactPaid to get some of the detail info.
And a few options in between with multiple fact tables, LOS pointer in FactPaid, Bridge tables, etc.
Let me say, I think all would work to some degree or another, but we are trying to prevent political pressure by following the "book/industry standard" on this.
We just do no know what that is.
Any thoughts on how to approach this or pointers to guidelines on this would be appreciated.
In the HealthCare Industry, multiple claims can make up one Length of Stay (LOS) or Episode Definition.
That calculation is easy. Created from FactPaid fact table
The harder part is the "best" way to capture information about said LOS.
For each LOS you need to capture info about pre and post services, insurance eligibility etc. (For those of you in the industry, HEDIS measures, etc)
Also info about Admission and Discharge (providers, diagnoses) along with in-between information.
One approach is an all encompassing factLOS table with all of the additional measures noted above.
Another is a bridge back to FactPaid to get some of the detail info.
And a few options in between with multiple fact tables, LOS pointer in FactPaid, Bridge tables, etc.
Let me say, I think all would work to some degree or another, but we are trying to prevent political pressure by following the "book/industry standard" on this.
We just do no know what that is.
Any thoughts on how to approach this or pointers to guidelines on this would be appreciated.
lgroff1961- Posts : 2
Join date : 2010-08-12
Re: Thoughts on Approach -- Length of Stay Hospital Claims
Length of stay should not be that difficult to do... it's the time between admission and discharge.
"Episode" or "Instance of Care" is a whole other matter. You may wan't to see if there is a HEDIS definition for this. But, quite frankly, unless a standard definition exists (I am not aware of one), be prepared for the politics.
"Episode" or "Instance of Care" is a whole other matter. You may wan't to see if there is a HEDIS definition for this. But, quite frankly, unless a standard definition exists (I am not aware of one), be prepared for the politics.
Round Two
Okay, maybe I was not clear.
The definitions of LOS are agreed upon (as much as they ever will be)
The calculation, while not trivial, is easily accomplished.
The "basic" FactLOS is easily modeled.
The real issue is around how to model the extra factors that are related to a LOS. Pre and Post-care, HEDIS measures, Diagnoses etc.
It is quite possible for 3+ claims to make up on episode/LOS and information needs to be maintained from each claim (especially the Admission and Discharge claims) along with the above Pre/Post/HEDIS.
Just looking for some general direction on an accepted or successful approach to this.
The definitions of LOS are agreed upon (as much as they ever will be)
The calculation, while not trivial, is easily accomplished.
The "basic" FactLOS is easily modeled.
The real issue is around how to model the extra factors that are related to a LOS. Pre and Post-care, HEDIS measures, Diagnoses etc.
It is quite possible for 3+ claims to make up on episode/LOS and information needs to be maintained from each claim (especially the Admission and Discharge claims) along with the above Pre/Post/HEDIS.
Just looking for some general direction on an accepted or successful approach to this.
lgroff1961- Posts : 2
Join date : 2010-08-12
Re: Thoughts on Approach -- Length of Stay Hospital Claims
That's what I mean by instance of care...
It gets really complicated, really fast.
When I had to deal with it in the past (about 15 years ago) there were very lengthy discussions to develop a method. They decided to use DRG as the focal point. Given an occurance of a particular DRG, we would look at activities prior to and after that event involving the same DRG or similar diagnosis. The activities would be bound by a timeframe before or after another event considered part of that instance of care. The timeframe would expand as events are added to the IOC.
The issues became defining "similar diagnosis" and the timeframe. The clinical group went about building a list of diagnosis to consider for a particular DRG (this took a while), and, at the time, settled on a fixed number of days between events for the timeframe. There were valid arguments that the timeframe should vary based on DRG, but it was felt it added cost and complexity and would probably not have a significant impact on the results.
It gets really complicated, really fast.
When I had to deal with it in the past (about 15 years ago) there were very lengthy discussions to develop a method. They decided to use DRG as the focal point. Given an occurance of a particular DRG, we would look at activities prior to and after that event involving the same DRG or similar diagnosis. The activities would be bound by a timeframe before or after another event considered part of that instance of care. The timeframe would expand as events are added to the IOC.
The issues became defining "similar diagnosis" and the timeframe. The clinical group went about building a list of diagnosis to consider for a particular DRG (this took a while), and, at the time, settled on a fixed number of days between events for the timeframe. There were valid arguments that the timeframe should vary based on DRG, but it was felt it added cost and complexity and would probably not have a significant impact on the results.
LOS
This response will be based upon not knowing what you are doing but knowing alot about healtcare claims and EDW. It appears that you have a paid claims fact table i'm assuming at the line level.... i will also assume you have a method to tie claims together by episode.... using a grouper or whatever.... so it also makes sense the you will have a claim low and claim high date and an episode low date and episode high date and the difference is teh LOS..... Make sense....
kjsdad- Posts : 1
Join date : 2010-09-02
Similar topics
» Healthcare Hospital Stay fact table - transaction fact table - a row for In and a row for Out?
» Data Vault v's Dimensional Model
» Too many dates on fact - Is there such a thing as Junk Date dimension
» Insurance Claims
» Monthly Claims Snapshot
» Data Vault v's Dimensional Model
» Too many dates on fact - Is there such a thing as Junk Date dimension
» Insurance Claims
» Monthly Claims Snapshot
Page 1 of 1
Permissions in this forum:
You cannot reply to topics in this forum
|
|