Recently, there has been a large shift in the United States towards better support for those facing behavioral health difficulties. In 2022, the 988 hotline was established as a 24/7 resource for anyone experiencing issues with their mental health. For example, A CDC report from 2023 outlines large increases in depression and violence among teen girls in the US; for our purposes, it’s a microcosm of what’s playing out across the country for people of all ages and genders. To keep up with the staggering rise of behavioral health issues and diagnoses, our health system continues to expand access to care through updated policies, the loosening of regulations, and alternative access routes such as telehealth. Unfortunately, bad actors are taking advantage of the surge in complex behavioral health needs to commit more fraud.
Applied Behavior Analysis (ABA) treatments in particular pose an avenue for fraudsters to take advantage of kickbacks, overbilling, and more. Behavioral health claims often cover complex evaluations, interventions, and treatments - creating gray areas that bad actors can exploit. There are various levels of procedure codes, alignments with diagnoses, overlap of coverage and care, and other factors that can make it difficult to keep track how a plan member is covered or how a claim is properly billed or paid.
This is a multifaceted issue that involves high sums of money and difficult-to-follow changes in care and codes, which means health plans need to be more vigilant than ever to make sure that treatment is appropriately given and that claims are processed quickly and accurately
Policy Changes Present a "Moving Target"
One of the biggest challenges in navigating this landscape is the rate at which codes and policies surrounding behavioral health treatments change. Throughout the increase of behavioral claims, codes have been updated numerous times, even within the same calendar year, which means what was covered six months ago might not be anymore or vice versa. This challenge also makes behavioral health a great target for fraudsters because they know that investigators are swamped with valid as well as fraudulent cases and are unable to investigate each claim.
It’s up to plans to flag providers who try to game this system to their advantage by committing fraud. But where should they start?
Common Schemes
When looking for signs of fraud, waste, and abuse in behavioral health, it’s crucial to look for outliers. The schemes that we’re increasingly seeing fall into three general categories:
As mentioned earlier, fraudsters are having a field day with the rate of change in behavioral healthcare coverage, policy and codes; by targeting areas that have recently undergone updates — slipping through the cracks of evolving coverage. What’s more, many ABA patients are not privy to the claims that are being submitted and whether those services were in fact provided to them.
Another complex example is an individual receiving excessive or improper ABA treatment over a period of time. Plan members or beneficiaries could be using ABA treatment as alternative childcare or, potentially, for financially beneficial connections between providers, facilities, and members that could spark a complex fraudulent ring of bad actors.
Also, situations like diagnoses in an age range that is not usually associated with a behavioral health matter or a provider that billing for more time than should be possible with their staff during any given week are red flags. These instances must be validated with SIU investigative work; if documentation indicates that only a few minutes were spent on treatment and a full hour is billed, there could be more going on. It is up to investigators to find the necessary documentation, claims data, and various external data sources before conclusions can be made. However, gathering that necessary investigative detail is easier said than done.
You’ll notice that these schemes, while sometimes complex, are detectable with investigative resources, attention, and a lot of legwork. The problem is that there aren’t enough legs. Even when there’s an obvious discrepancy between a plan and what’s documented, a potentially high value case can easily get missed simply because there are too many cases to investigate, which is what most of these fraudsters are counting on and what costs health plans millions of lost dollars.
Rules-Based Detections vs. AI-Powered Detection
Traditional healthcare fraud detection strategies can be manual, rules-based, or automated. These methods are designed to help health plans deal with problems at the pace of fraud, but differ in how they detect it.
What's the difference?
Essentially, rules-based detection requires significant manual input, research, and analysis — building, monitoring, and maintaining “if-then” rules-based scenarios. Data scientists and teams have to do the work of programming rules for detection, like payments being made in areas that a member or beneficiary doesn't frequent for services. If a person makes a claim in Ohio but their home address is in Massachusetts, for example, the solution would flag it for investigation. Similarly, multiple claims could be flagged as fraudulent if they are filed by the same provider on the same day at the same time — but the claim could still be legitimate if there are multiple providers at the same office.
Detection with advanced AI, on the other hand, uses machine learning to gather data, make connections, flag outlying or anomalous services, and deliver actionable insights all while continuously learning and refining its logic for increasingly accurate detection. These solutions are designed to streamline many of the manual information gathering and detection tactics, and capable of finding new fraud patterns on their own.
Which method is better suited to fighting healthcare fraud?
Traditional, rule-based fraud detection solutions are at a disadvantage in this situation. The need for manually updating each rule when policy or code changes take place can quickly leave an health plan behind on the latest regulatory updates. And this is what fraudsters are hoping for — that team don't have the bandwidth to keep track of every policy change and their schemes go unnoticed.
Detection methods that leverage AI have the upper hand in other areas as well. Since models can be trained to gather data and generate insights for services, providers, and compliance, they continuously “learn” and adapt to new scenarios. Unlike static rules, AI-powered detection models can evolve with the ever-changing landscape of healthcare fraud, waste, and abuse.
Automated solutions can also pull in external data, such as provider reviews, exclusion lists, business/corporation data and even social media data to find connections and help validate possible fraud, waste, and abuse. For example, if a person claims they were performing or undergoing treatment during a time when they communicated via social media that they were in another country, it might aid the investigation process significantly.
Automated solutions are better in these cases because they are designed to keep up with the pace of policy changes and evolving fraud, leaving no stone unturned and catching activity that would fall outside of the parameters of rules-based detection methods.
Don't Miss the Misbehavior in Your Claims Data
The growing need for behavioral health services will continue to increase, but that doesn’t mean that health plans have to experience an increase in undetected fraud. Advanced AI and machine learning will not only detect fraud faster, but also greatly reduce average investigation time and bring much higher value cases to the forefront.
We expect to see more variations of these schemes in the behavioral health market as the year goes on. Watch our webinar with special guest Carlyn Hoffman, AHFI at Integrity Advantage, on the topic to learn more about how Advanced AI can help you get ahead of behavioral health schemes.