Swift writing skills were needed from IT managers keen on understanding just how Health Benefits implemented its online claims system.
Health Benefits (HBL) gave an account of the rather novel methods used to implement General Transaction Processing System (GTPS), but these were given under conditions of semi-secrecy.
“Unfortunately, due to the confidential nature of this presentation, we cannot provide hard-copy slides,” says a note in the Software Developers Conference programme. All the other presenters provided such slides, often in great detail, though a swag of notepaper was provided with the HBL presentation.
HBL, which used to be called the Health Funding Authority, distributes $1.2 billion in primary health care subsidies a year and needs to process 72 million transactions. GTPS is designed primarily to manage the payout of the government subsidies allowed to health professionals on medications which government buying agency Pharmac decides to fund. Health professionals send in claims by phone, fax, email and from a variety of different computer-based practice management systems.
GTPS claims have to be massaged into a consistent format, which at the same time is broad enough to allow for future additions, removals and changes consequent on government policy, said HBL presenter John Anderson.
Then the claims are put through a complex set of business rules to establish what subsidy is justified in each case, bearing in mind not only the current status of government policy, but the history of the patient and his/her treatment.
The most original feature of the application was its organisation as a message-passing system. The “generic” claim message, capable of accommodating different types of claim in the one format, is put down a message pipe using SOAP from the client — health professional’s — system to the WebSphere applications server and MQ to queue the messages internally.
Any extra piece of information that the claim needs is simply picked up and appended to the end of the message, rather than the conventional approach of a transaction being performed between the claim and the information. When the message reaches the front of the queue with all its appendages, it is fed into the rules engine. This ensures that the high volume of transactions are speedily dealt with.
Security is naturally a big issue with health records of any kind and there was initial difficulty combining SSL (secure sockets layer) with SOAP; but the later versions of SOAP devised for Apache SOAP by axis.org solved that.
As much off-the-shelf software as possible has been used, Anderson says. The rules engine is XpertRule from UK-based Attar Software.
Development was done in Java and XML using the Rational Rose design suite and initially Visual Age for Java but later transferring to the open-source Eclipse suite of tools. Visual Age proved “klutzy” (clumsy) in the way it handles fragments of Java code, Anderson says.
HBL began building the system two years ago, and it was completed in September last year.