Sick of fraud

In and of itself, the term advanced analytics has been somewhat abused by the majority of mainstream business intelligence (BI) vendors of late. It has become almost synonymous with trying to market and sell a 'one size fits all' solution. The focus has shifted to what the technology can do for business, rather than the technology itself, and it has become a form of dictated analysis, rather than the exploratory analysis it should be.

Advanced analytics in its truest sense is defined as using extensive statistical and mathematical models to understand customer behaviour, predict retention, predict churn, and predict fraud.

At its heart, it provides the algorithms for complex analysis of either structured or unstructured data - including sophisticated statistical models, machine learning, neural networks, text analytics, and other advanced data mining techniques.

When deployed, it can be used to find patterns in data, prediction, optimisation, forecasting, and for complex event processing/analysis, as well as being used to predict churn, identify fraud, market basket analysis, and understand Web site behaviour, etc.

While no accurate data detailing the extent of healthcare fraud in South Africa exists, working on the assumption that international trends apply, it is completely possible that fraudulent medical scheme claims account for between R3 billion and R15 billion annually.

In 2002, in America, the Inspector General of the US Department of Health and Human Services identified $12.1 billion in fraudulent claims paid by Medicare, and international data suggest that losses to healthcare fraud and abuse may account for between 3% and 15.4% of claims paid - with an average of approximately 7%.

Fraudulent medical claims are big business.

Clearly, fraudulent medical claims are big business. To put it into perspective, in 2002, the perceived 'bigger problem', credit card fraud, amounted to approximately $788 million in annual losses - healthcare fraud accounts for over 100 times that value.

That is a massive amount of money, and all that fraudulent activity is one of the driving forces of escalating medical scheme costs. Every day, it becomes more and more imperative for medical aids and their stakeholders to find reliable ways of identifying fraudulent transactions.

Identifying syndicates, fraudulent member claims and untruthful brokers or medical professionals fast and accurately is more relevant now than ever, as it is common for instances of fraud to increase exponentially in times of economic hardship or during recessionary years.

Medical scheme administrators process millions of transactions monthly - every visit, every script, every diagnosis, every tablet, every test and every admission. That is a massive volume of data, and more and more medical aid schemes are using that wealth of electronic data to develop tools that can better detect fraud.

Tools like Discovery Health's proprietary forensic software system, Informa. Informa takes the massive amounts of data generated and stored by medical aid schemes every day and uses smart and dynamic algorithms to trawl through all the claims data on a daily basis. This enables the medical aid to identify any unusual patterns and flag items for further investigations.

Well built, well designed and robust advanced analytics tools can and should be deployed to harness the power that is inherent in all of that patient and vendor data. Using algorithms, machine learning, text mining, etc, the trends and patterns that become visible and identifiable in vast volumes of data can be used in to predict and identify fraud.

And to a degree, human nature is also quantifiable, and people are nothing if not predictable. So the segmentation and analysis of a client/customer/vendor base is also imminently possible. Finding out and tracking how a particular segment of the population/patient base/vendor behaves makes it that much easier to spot the 'outlier'.

Medical aid fraud is committed by a combination of medical scheme members, healthcare professionals, service providers in the healthcare industry, and brokers - not for one moment forgetting the sophisticated syndicates, both local and international, which appear to be targeting the healthcare system.

Only by deploying an advanced analytics tool in its truest sense can one hope to combat a crime that is 100 times more attractive than credit card fraud. Big business indeed.