The Problem—Pulling it all together
Horizon Blue Cross Blue Shield needed to find the most efficient way to manage its data—just as it has found efficient ways to manage health insurance.  With 750,000 managed care customers, Horizon BCBSNJ has been one of the fastest growing HMO providers in New Jersey for the last five years.

Horizon’s CIO has had positive experiences with data warehousing in the health care business, according to Richard Angelillo, Horizon’s data warehouse project manager.  “We believed that developing a data warehouse would greatly contribute to Horizon’s ability to successfully manage the health of our customers.”

In response, the first data warehouse that Horizon BCBSNJ initiated focused on managed care data. Specifically, medical management decision support, and provider profiling, because they were seen as having the biggest return on investment.  Roughly half a terabyte of data in DB2 and flat source files is now populating a DB2 Universal Database data warehouse on an RS6000 Unix platform.  The four major sources of data are membership, pharmaceutical, and two major managed care systems.  The number of source tables ranges between 200 and 250.  The number of tables in the data warehouse is approximately 115.

The Process
“While we didn’t have a lot of formal system architecture requirements, we did a lot of shopping,” says Angelillo.  “We had various data marts—actually just relational databases—but no existing data warehouse to start from.”  At first, Horizon BCBSCNJ looked for a packaged product to avoid building a healthcare data model from scratch.  The insurer contacted other Blue Cross Blue Shield plans but found that most are not in the business of selling their models.  Horizon BCBSNJ surmised that a lot of customization would be required so it chose a model that could be implemented easily in stages.

Commercial suppliers, on the other hand, presented models that were provider-oriented rather than payer-oriented.  Horizon reasoned that totally reorienting staff to a new way of thinking about the data would significantly increase preparation time. 

For data conversion, Horizon BCBSNJ looked at vendors such as Informatica and ETI.

Criteria
Horizon BCBSNJ wanted to capture 100% of its complex metadata in the most efficient way.  Ultimately, the critical mass of its HMO population convinced Horizon BCBSNJ to investigate its existing strengths. 

The Decision
The insurer narrowed the options to products that could: (1) Handle the sheer volume of legacy data that needed to be cleansed, transformed, moved and loaded into a Unix-based client-server system; and (2) move data across varying platforms from and to various data formats.  ETI Solution products immediately stood out.

“After a thorough due diligence effort, ETI looked like a better heavy-hitter than any of the others,” Angelillo said.  ETI Solution easily loads high quality data into IBM’s Visual Warehouse.  Horizon BCBSNJ found that the combination of ETI’s and IBM’s products provides a powerful solution for repeatable—not one time only—operations.