Healthcare, diagnosis and Dimension model

Hi, I'm current in research for doing some major improvements in our healthcare data warehouse. I have few questions regarding using a dimensional model for our healthcare data. Hopefully, they are trivial to some of you.

Here is our problem

Currently we do not have a data model in the proper meaning of the word. We have several data marts who sometimes are dependent and sometimes not. My idea is to introduce a data model and I'm currently checking if dimensional models is an approach. I can also say that we work mainly with the SAS software.

When I read about Dimensional models, it's mostly about some type of economical approach and therefore it's practical to store billing information in the fact table. In our data, we mainly have yes/no answers to questions like "Do you have Diabetes". We have only a few measurements like blood pressure, radiation dose but I don't think they are Fact-material. Is this a problem? Are they any examples on successful healtcare implementations?

My main concern is the question of how the model supports diagnosis, when a patient can have any number of diagnois. Is this Hierarchy or Level? We have several one to many answers in our data.

I understand that we can use conformed dimension to link a patient-dimensional model with a operation-dimension model. Is this correct?

I have one more question.
Is the ETL goal to create a data model and from the model create different datamarts?

With regards
Roger Svensson

I actually found an entry of this in the "The Data Warehouse Toolkit, 2nd ed., p.262-" ;o) This may help...