Gerhard Hausmann presented on Barmenia and their use of business rules to improve customer experience. Barmenia is a private health / life insurance company in Germany with more than 2M contracts and 1.5Bn Euros in premiums. Been in business since 1904 and still have contracts that date back to the last century. In Germany there is a mix of public and private health insurance – 70M people have public insurance and 8.5M are privately insured with another 20M having some additional private health insurance. Barmenia, like many companies, is facing challenges from an ageing population. For health insurance this is a particular challenge as older people are more expensive.
Barmenia is graded by a German quality organization and consistently gets A+ (6th time in a row) in part thanks to the system supporting 48 hour claims processing that is accurate and competent. Barmenia has been adding decision support in claims processing starting 2003 with a database of reimbursement information and moving to using ILOG rules in 2007 for auditing complex dental invoices, naturopath invoices and case assignment. They took a decision-support approach because they feel they have good experts. They decided on a custom rules approach to get short logic release cycles and to avoid the need for the large integration projects that come with out of the box implementations.
They use a few thousand rules in some decision tables, a few hundred regular rules to and many patterns to manage the rules for dental invoices – pricing, eligibility, allocation etc. They scan invoices and route them through an ESB. Simple ones are verified and decisions are automated using ILOG. More complex ones are routed to the case management system (again using ILOG rules-based assignment) with embedded instructions (generated from more ILOG rules). These are displayed to experts using an environment that combines the invoices, master data and decision support. Because they are still using a mainframe application they use a mixed interface that displays a mainframe screen with claims information related to several invoices, a display of an invoice and a decision support screen.
Managing complexity in rules is a focus area for them. Rules become complex when multiple objects with multiple properties must be considered. An example rule with 9 conditions requires 500 test cases. The handling of exceptional cases is a particular indicator of complexity. To manage this they use ruleflows to put default rules (simpler, cover lots of cases) first then create a series of exception steps with each exception step handling a few rules for a group of exceptions. This prevents the complexity from overwhelming rule editing. They use the Business Action Language to make the rules accessible to business users and they put business experts into the QA process, having them create and run tests for their rules. Regression tests are handled by IT with experts conducting random tests on the new version.
- BRMS uses lots of facts, more than a person would, and so the system decides slightly better than the experts used to.
- Supports those who handle cases so that less experienced handlers (trainees for instance) helping them close out potentially complex claims. More claims can be hanlded by the usual teams.
- Increased agility – they were able to add the rules for detecting simple cases and QA the rules in just 4 weeks.
- Resources are allocated more efficiently with claims being routed to case handlers or to experts appropriately. Fewer are routed to experts (less than 40%), improving response time.
They plan to extend the system to handle underwriting as well as invoices from more kinds of providers.