EU project develops digital detective to take on healthcare fraudsters
As healthcare fraud across Europe reaches a staggering estimated cost of EUR 30 billion per year, an EU-funded project is developing a digital detective to investigate healthcare cheats. The iWebcare project, which kicked off in January 2006, is a €2.3 million venture to design an information technology (IT) response to track healthcare swindles. According to the Counter Fraud Service (CFS) of the UK's National Health System, the problem is a huge one. Fraud accounts for almost 3% of public healthcare expenditures in the UK, and 3% to 10% of annual expenditures - both private and public - in the US. With the healthcare fraud across Europe estimated at €30 billion and healthcare costs set to surge as Europe's population ages, fraud rates could increase to the levels found in the US. Tackling fraud has therefore become a priority for Europe, since the money lost could be better spent on actually providing more health care to those in need. According to the UK experience, once the CFS was set up, losses due to fraudulent cases fell by 45%. The iWebcare project hopes to have an even greater impact on healthcare fraud across Europe. Whereas current detection methods rely on the relatively ineffective experience of a human fraud detective combined with computer data mining tools, the iWebcare project will develop a web platform with advanced information and communication technologies (ICT) tools, capable of detecting up to 90% of fraud. The partners in the project are working on a series of elements to create self-learning software programs to data-mine, or analyse, healthcare transaction records. The mammoth task involves creating metadata - information that is meaningful to computers - to identify a variety of data, such as names, addresses, drugs and procedures. The researchers are developing special algorithms, allowing the system to learn from past experience. In the first instance, this will involve human detective work in order to 'teach' the application to recognise patterns that deserve closer inspection. As time goes by and the application learns, it will soon become adept at spotting the patterns that point to potential frauds. The project is currently developing its first working prototype, which will be tested in real life conditions through its deployment in two Member States, the UK and Greece. At that point, fraudsters everywhere will come to fear the long arm of the digital detective.