To think an online transaction processing, or OLTP, database will enable that capability effectively is wrong. There are relational databases on the market today that are built specifically for deep analytics. VistA (using its new or old MUMPS OLTP database) should be feeding data into a relational database for advanced analytics. In fact, all of the VA's front-end applications should be feeding a relational database for cross-functional advanced analyses, data-mining and business intelligence.
Front-end applications/tools are good for what they are designed for. However, for thorough enterprise reporting and analytics, extract, load and transform tools along with a relational database should be used for deep analytics, data-mining and or business intelligence. Business area-specific OLTP databases, such as MUMPS, struggle with cross-business-area analytics because that is not what they are designed for.
Therefore, again the better question to ask is what does the VA want to do with all of its data? By deploying a relational database built for analytics, along with its new/old MUMPS OLTP database(s), data-mining, advanced analytics and business intelligence are possible.