Analyze This! IBM Helps Lower Anxiety over Medical Claims
July 7, 2010, Haifa, Anyone who has ever read an explanation of benefits from his health insurance knows that the term "explanation" is loosely used. The complex system of coding used to identify medical visits, procedures, and diagnoses, may as well be hieroglyphics.
Companies select among different plans for their employees, and some may offer specialized plans, resulting in vastly different levels of coverage among policy holders. Legislative updates, changes in company demographics, and individual employee preferences, can all further affect coverage policy. Medical coding has evolved into a complex and ever-changing science, prone to confusion and misunderstandings.
One company is out to change that.
A large company that serves BlueCross BlueShield plans throughout the United States, is committed to providing the most efficient and dependable claims-processing system available. To that end, they approached IBM looking for innovative ways to streamline their system to allow for frequent changes that could at any point be clearly traced back to their origins.
The ensuing project, under the leadership of Jonathan Bnayahu, manager of software asset management at IBM Research in Haifa Israel, became an IBM "First-Of-A-Kind" (FOAK). This put the healthcare insurance provider in impressive company-IBM's FOAK projects have helped more than 150 companies by developing and testing new technologies on actual business challenges.
Translating Medical Code into Plain English
One of the company's largest challenges involves traceability-being able to identify connections within their system. "Ideally, each benefit rule could be traced back to its definition in the insurance policy, but in the real world this information is generally not saved," explains Bnayahu from his Haifa-based lab. "The system simply evolves as the requirements change," he goes on to say.
For example, if Maryland passed a law requiring companies with over 50 employees to offer 100% coverage of maternity care, benefit coders need to locate the impacted benefit rules, rewrite the code to reflect that change. Today, that process is manual and time-consuming, and is highly error prone due to the inability to verify the changes.
"A larger problem comes into play when developers would like to make a system-wide change," notes Bnayahu. If Maine was to pass a similar mandate, benefit coders would have to rewrite code from scratch, with little or no reuse from the already validated implementation.
With IBM's innovative technology for traceability, benefit coders could sweep the system to find where this code is already being used, and then use the existing information to implement the new code wherever relevant.
To help reduce error rate (and increase first-pass rate), the project also addressed the issue of auditing. Using the new technology, healthcare payers can view how the actual code running in the system matches their requirements. Up till now, this could only be done by experienced coders. Now payers are also able to audit claim information, in the form of summarization tables and flow charts. The tables, like the one shown below, enable users to isolate data among large amounts of information, and visualize and easily trace any detail.
With the successful completion of the First-of-a-Kind phase of the project, the Haifa team is now working with other IBM units to deploy this technology in the production environment, making it available to health insurance clients. In parallel, the team is working to enhance and expand traceability-based testing and auditing capabilities. "Ideally, to a certain degree at least, we can automate the auditing process-it's like giving every piece of data its own button that says 'analyze this'," jokes Bnayahu.