AI, Healthcare, Insurance

Understanding Healthcare Claims & Fraud: An Unfortunate Certainty

Sierra Seabolt

Healthcare claims are an inviting and alluring target for fraudsters due to their huge monetary value. In 2017 alone, HHS and DoJ recovered around 2.5 Billion Dollars in claims fraud. And that is just one year.

The challenges facing healthcare payers are manifold as they try to tackle the rampant and growing problem of claims fraud. From high dollar impact, billing compliances to the high cost of manual investigations, the health care providers are sweating. So what exactly is Healthcare Fraud? And how can we try to mitigate them using technology?

Health care fraud is a crime, where one or more parties (doctors, patients, practitioners, etc) provides misleading or outright false information while determining the cost of health care services provided, and thus payable by a third party – typically an insurance company or the government (collectively, the “Payers”). Certainly, only a small percentage of practitioners or individuals provide misleading information, but this small percentage still translates to huge dollar amounts, because of the sheer volume of healthcare-related services.

Over the years, the Payers have taken many steps to try and mitigate fraud by establishing detailed rules and protocols for payment, establishing diagnosis and procedure codes and standard payment amounts, standardized groups of services for episodes of care, etc.  In spite of all these steps, the unethical and dishonest claimants still figure out loopholes that the fraudsters can take advantage of. And they are getting smarter, they know the system very well, and work their way around it.

Some of the top fraud challenges in health care claims are –

Billing for Services Not Rendered

This is probably the easiest form of fraud performed by practitioners. It requires very little brainpower to submit extra services that have not been provided to patients to be paid extra bucks.  Most of the time there is no detailed documentation supporting these claims. It’s always not just a document that is needed to justify the fraud here, witnesses are also required. Most of the time there are none, making it extremely difficult to call out an anomaly here.

Billing for Non-Covered as Covered Service

Some of the experimental treatments and non-covered care services are often coded as a covered service. The practitioner codes it in the name of other covered services, and get paid conveniently. Keep in mind that the patient only cares for two things, how to get better and how much he/she is paying out of pocket. One might also hear the practitioners justify this in the name of better patient care. But at the end of the day, the healthcare payer needs to pay for services that they had not agreed to pay for.

False or Unnecessary Prescription For (Expensive) Drugs

False or unnecessary drug prescription is an increasing concern not just for the healthcare payer but for the nation. The issue here is not just the practitioner, even patients and pharmacists are involved in forging prescriptions or ‘doctor shopping’ to get more pills. The doctors are unaware that the patient has been visiting other doctors to get more pills. In some cases, pharmacists are involved in stealing painkillers by issuing them in the name of patients who are completely unaware. All this causes huge bills for the healthcare payers.

Way Forward

The manual, rule-based investigations are too slow and cannot keep up with the speed at which fraudulent transactions are being submitted for payment.  With the limited resources, it is impossible to find out inappropriate links between different claims. This leads to fraudsters forming schemes that stay undetected. Thanks to recent advances in computer technology – both in terms of faster and more powerful computers, as well as advanced software tools that can use deep learning, artificial intelligence and machine learning capabilities to proactively apply predictive analytics to develop insights into potentially fraudulent healthcare claims by rapidly processing vast amounts of both structured as well as unstructured data.

This presents an exciting opportunity to finally get ahead of the fraudsters… NEMESIS’s data science team is actively working in this area.  Please contact us if you would like to know about our work in this area.

Sierra Seabolt