Detecting health care claims fraud
How to get faster results, fewer false positives and more fruitful investigations
Several years ago, the US Department of Justice announced that a Florida urologist would pay more than $3.8 million for ordering unnecessary medical tests – more than 13,000 urinalysis tests that are only applicable to certain cases related to bladder cancer. The No. 1 referring physician in the country for the fluorescence in situ hybridization (FISH) test, this doctor had received $2 million in bonus payments from the lab that performed the tests.
“Greed was the clear motive," said a health care claims fraud investigator involved with the case. “Patients' needs played no role in ordering tests that were medically unnecessary and could have endangered patient care.”
This case was uncovered by a whistleblower, but what about less flagrant fraud activities? How much fraud is being missed?
“One-off, large-scale frauds are unusual, but widespread low-value fraud is common,” said Joseph Kutzin, Coordinator of Health Financing Policy at the World Health Organization (WHO). “This makes it harder to detect more than a small proportion and means that an emphasis needs to be placed on pre-empting fraud, rather than reacting to it after it has occurred. Fraud which is visible and which has been detected is only a small element of the total cost.”
That total cost is staggering. In 2015, the accounting firm PKF Littlejohn and the Centre for Counter Fraud Studies (CCFS) in the UK set out to understand the financial cost of health care claims fraud globally, using 17 years of data from 33 government health organizations. Researchers concluded that as of 2015, fraud and error losses in any organization are at least 3 percent, probably more than 5 percent and possibly more than 10 percent.
At the current global average loss rate of 6.19 percent, that means fraud diverts US$455 billion away from patient care. According to the report, if health care organizations reduced those losses by 40 percent – which individual organizations have achieved – it would free up more than US$182 billion.
Payment integrity: Predict fraud and detect loss
How much are fraud, waste and abuse costing your organization? Costs to insurers are huge – as much as 25 percent of payments made. Learn about the role payment integrity plays in helping you make better payment decisions to protect against this risk.
Finding under-the-radar health care claims fraud
“You have a moral obligation as an insurer to see that the money of the insured person does not reach the wrong parties," says Ivo van den Berg, Team Leader for Business Intelligence and Analytics at the health care insurer DSW. Rotterdam-based DSW provides individual and business/corporate health insurance for roughly 600,000 policyholders in the Netherlands and is the fifth-largest health care insurer in the country.
“Fraud and inappropriate use of health care is a growing problem for every insurer,” says Van den Berg. “To optimize the process of identification and prevention, we needed a solution that could support us.” DSW previously worked separate, hypothesis-driven investigations. “However, with that approach, you miss things that are outside your hypothesis. It is also very difficult to identify relationships between treatments and providers. Many decisions were made only on the basis of experience while we want to do the right thing based more on data and facts.”
Based on a fellow health insurer’s experience, DSW implemented the SAS® Fraud Framework, which offers a combination of:
- Data management to consolidate historical data from internal and external sources – claims systems, watch lists, third parties, unstructured text and so on.
- Rule and analytic model management to create and manage business rules, analytic models, alerts and known bad lists.
- Detection and alert generation based on a calculated propensity for aberrant billing at first submission, with claims rescored at each processing stage as new data is captured.
- Social network analysis, also called link analysis, which reveals connections among entities to expose organized fraud rings or collusive activities.
DSW first applied the SAS Fraud Framework to examine claims from dentists. “This was an area where we already had a fair amount of knowledge about who has probably claimed incorrectly,” said Van den Berg. “So we could pretty easily test whether the framework also indicated what we already knew. This was the case, so there was a green light to expand the work to other care areas.”
Elements of models that were built to examine dental care fraud were easily applied in other domains, Van den Berg said. “For example, we can use the analysis of the deviations in cost between care providers in each discipline. This saves a lot of time. When looking at physiotherapists for instance, we can see very quickly whether they deliver more standard treatments than their peers.”
We can use the analysis of the difference in cost between care providers in each discipline. This saves a lot of time.
Ivo van den Berg • Team Leader for Business Intelligence and Analytics • DSW
A trifecta of benefits
Better use of data. "A big advantage of the SAS solution is that we work with one version of the truth,” said Van den Berg. “Through standardization, everyone sees the same information, and there are no differences in interpretation. This is essential, especially when you're talking about as serious a matter as fraud.”
Faster insights. Van den Berg is satisfied with the speed with which his team can show results: “Because we have short communication lines within DSW and we work with small, agile teams, we could rapidly present the first concrete results. We immediately saw the benefits of the solution.”
Fewer false positives. “We already had the idea that we detected more fraud with the SAS Fraud Framework, but it was nice to see it confirmed,” said Van den Berg. “Now we can better focus on fraud alerts that really are worthwhile to investigate further."
An investment that repays itself
Improving your ability to detect health care claims fraud pays off quickly, according to PKF Littlejohn and CCFS: “Where losses have been measured, and the organizations concerned have accurate information about their nature and extent, there are examples, especially in the UK and US, where losses have been substantially reduced.” Indeed, these examples include historic success in the UK’s National Health Service (the second-largest government health care payer organization in the world), “where losses were reduced by up to 60 percent [from 1999 to 2006], and by up to 40 percent over a shorter period.”
An analytics-driven health care claims fraud platform also reduces the cost of preventing those losses, said Van den Berg. “Once you have your well-appointed detection, it takes much less effort to detect fraud – partly because the process of recovering undue payments is very time-consuming and expensive, so correct detection is very important. Prevention is better than a cure, because you as an insurer [discover fraud] in time to start the conversation with the caregiver and can counteract new inefficient or improper claims.”
“Real gains have been made by reducing the cost of fraud – with up to a 40 percent reduction possible within 12 months,” said Kutzin. “This is obviously good news, and with health systems under financial pressure, cutting the cost of fraud can be a significant additional source of progress toward [universal health coverage], for example, by putting the ‘recovered’ resources toward increased coverage of services. As with the protection of health more generally, proactive, pre-emptive action has an important role to play.”
Learn more about SAS solutions for fraud.
Recommended reading
- Artículo Fraud detection and machine learning: What you need to knowMachine learning and fraud analytics are critical components of a fraud detection toolkit. Discover what you’ll need to get started defending against fraud – from integrating supervised and unsupervised machine learning in operations to maintaining customer service.
- Artículo 6 formas en que el big data analytics puede mejorar el tratamiento de datos de la indemnización del seguro¿Por qué integrar la analítica en el tratamiento de datos de las reclamaciones de seguros? Porque añadir la analítica al ciclo de vida de las reclamaciones puede proporcionar un retorno de la inversión cuantificable.
- Artículo How AI and advanced analytics are impacting the financial services industryTop SAS experts weigh in on topics that keep financial leaders up at night – like real-time payments and digital identity. See how advanced analytics and AI can help.
- Artículo Intelligent policing: Data visualization helps crack down on crimeLearn how data visualization can give police real-time views of locations enriched with other data to help them make intelligent, fact-based decisions.