Governing Reference Data in the Cloud

Insurers are progressing with governance programs, glossaries and metrics for a mix of financial and transactional data records. Fewer have moved onto comprehensively governing reference data – in a cloud solution no less - as has Pennsylvania-based health insurer Independence Blue Cross.

“You need to manage reference data in a way that is simple and automated,” said James Denyer, Director Data Warehouse Development, Independence Blue Cross (IBC). When Denyer refers to reference data, he’s talking specifically about IBC’s hundreds of code data sets, including medical code sets, administrative code sets and master data reference data. The lists within the code sets range from simple to complex. The architecture also covers hierarchies of tables and crosswalk tables, which create a relationship to other code tables through translating external values to a single dialect.

Crosswalk codes help to eliminate synonymous or redundant code sets used in different reports. “Even though the reference data might come from three different places, it all looks the same,” said Denyer.

The challenge was to integrate a cloud-based data governance framework with an existing enterprise metadata repository to create a complete solution for both reference and crosswalk governance. The solution needed to support stewardship and other activities at IBC.

Since reference data is mostly in the form of lists that are descriptive and relatively innocuous, it was well suited for cloud deployment. Additionally, cloud eliminated e-mailing, file transfer and firewall obstacles. Denyer and others worked with the vendor, Collibra, to jointly develop the solution for IBC.

“You can manage all these glossaries in the cloud and people can go there with a single logon,” Denyer said. “Everything is maintained and updated in the cloud, and users aren’t typically aware the solution is off-premise. 

IBC created a synchronization point with ansay SQL database and RESTful services that exchanges updates to the cloud and creates a workflow that informs users of new codes.

At IBC, code sets are managed for impact on health policy, business rules, and crosswalk logic across medical, drug and administrative codes repositories. As new codes are needed and requested, users propose a code through a form, which is forwarded to subject matter experts for review and approval. Once approved, it is deployed via RESTful services to the cloud and another notification is sent to all parties that the process is complete.

It’s an elegant solution but not necessarily an easy elevator pitch to management, Denyer said. “Reference data is not something that keeps people up at night usually. But if you can identify how you can prevent reputational damage from incomplete code sets or help initiatives and processes move along faster, it’s a benefit.”

Denyer delivered his presentation at the annual IDMA conference in Philadelphia.

For reprint and licensing requests for this article, click here.
Analytics Data and information management Core systems
MORE FROM DIGITAL INSURANCE