Stephen Brobst, CTO at Teradata, added that the model being used to build the repository is designed to capture detailed data for analysis, not summary data. "If you start summarizing data, you're not able to see the relationships in the way you need to -- that total picture of all the different health care delivery (methods)," Brobst said.

In addition to making sure it is using detailed data, the project requires that new categories of data be checked for quality, Brobst said. In the past, data categories required for payment such as diagnosis and procedure codes have been the focus of cleansing efforts, he said. But other categories like demographic information need to be cleaned to ensure that the BI analysis is accurate, he added.

Evangelist said that business users will determine how the project will advance after the initial data is loaded. As a result, he could not provide details about its cost or the potential for existing legacy warehouse systems to be phased out.