Mark Starinsky, AHFI, CFE, CHC, SA, is a healthcare Fraud, Waste and Abuse (FWA) subject matter expert and senior product manager at Shift Technology.
This instalment of the “Four Questions with” series follows a similar conversation we had with Mandy Fogle and addresses the technology considerations health payer organizations must take when adopting prepay improper payment/FWA strategies.
Shift Technology: From a technology perspective, how is the industry currently approaching prepayment improper payment detection/FWA, and what is Shift doing differently?
Mark Starinsky: I think a lot of people in the industry are taking the approach of, “the best predictor of future behavior is past behavior.” So, if they’re even thinking about expanded prepay activities, they’re focusing those strategies on things like National Correct Coding Initiative (NCII) edits, historical claims data or even just the results of investigations which they're currently undertaking.
In the case of historical claims data or investigation results, health payers may complete a post payment investigation and from those findings they’ll say something along the lines of, “we should implement some sort of prepay for this guy. Maybe we should put this provider on prepay review.”
The major limitation of this strategy, from both a technical and a business perspective, is that it’s a binary approach when it comes to edit capabilities. Based on the edits provided, the computer says “yes” or the computer says “no.” That’s all there is to it. They’re essentially using a deterministic logic engine in a circumstance that requires much more nuance. Perhaps a provider does have a history of billing a high level of E/M codes, with little justification. Maybe it is prudent to flag all of a provider’s claims for review prior to payment and endure everything that entails - the time to request and receive medical records, the time and cost of a licensed medical records reviewer, the provider abrasion. I could go on and on. But what if the investigation simply uncovered an outlier in the provider’s claim history?
Shift’s approach is significantly different in that we believe the principles that drive successful postpay claim review and investigation can drive prepay as well. By analyzing post-payment claims data we can develop more effective and even new claims edits. The outputs of post-payment investigations can help health payers effectively identify those individual claims that warrant further review and separate them from legitimate claims, even when the claims are submitted by the same provider.
Shift Technology: How does Artificial Intelligence (AI) fit into Shift’s approach to improper payment detection?
Mark Starinsky: First, let’s dig a little deeper into what happens currently when a provider is put on prepay review. It's not a refined process. It’s mostly all or nothing — all of a provider’s claims will be reviewed before payment. And so what tends to happen is that the provider gets upset about the position they’ve been put in. This can cause a lot of abrasion.
When a provider has been put on prepay review nothing gets paid until the claim has been cleared, and that means a review of every single claim that’s submitted.
The real benefit of thinking about prepay differently, and in taking a more sophisticated approach that employs artificial intelligence to analyze claims is that you now have the power to refine the scope of review. Instead of assuming all claims need review prior to payment, AI is able to determine which individual claims, independent of which provider has submitted them, show signs of suspicious behavior. And when those models are informed by data uncovered as the result of postpay investigations, combined with other external data sources, you’ve created a closed loop system that learns and adapts and gets really good at spotting the outliers. When you can quickly, efficiently and fairly separate questionable claims from legitimate claims the business benefits are tremendous.
Shift Technology: Where do you see AI making the greatest impact related to enhancing prepayment decisions and detection for plans?
Mark Starinsky: In the near term, I believe AI’s greatest impact will be focused on providing greater nuance to how claims are identified for further review and investigation. AI-powered strategies allow health plans to avoid putting providers or in some cases entire provider networks on prepay review. I can’t express strongly enough how much easier life can be for health plans and providers if prepay review is focused on the claims that need it as opposed to every single claim associated with a provider or provider network.
Yet, there’s something even more exciting on the horizon. I’ve talked a bit about medical record review in the context of prepay. This is an arduous, highly manual and time consuming task. I think where we're going to go next with AI is tackling the problem of medical record review. We’ve seen the power of Generative AI and Large Language Models. These solutions can be trained on medical records, claims data, and relevant external data. Focused prompt engineering can be applied. All of which gets us to the point where AI can help us determine if the information contained in the medical record support the requested payment.
It’s important to remember here that we’re not talking about AI replacing qualified human record review. Applying AI to this problem is about augmenting the human experience rather than replacing it. AI provides the focus and resulting information to help a professional make the final decision.
Shift Technology: In your opinion, how should health plans begin adopting an evolved prepay strategy?
Mark Starinsky: Like most anything new, the first step is to acknowledge that there may be a more effective way to think about improper payments/FWA. The siloed approach isn’t effective. Pay and chase isn’t effective. And traditional prepay review isn’t effective. There needs to be a more holistic approach to the problem.
And the good news is that from a technology perspective, the underpinnings for a holistic approach are already there. We have AI technology that can effectively analyze claims to spot suspicious behavior and provider networks. These technologies can be applied both prepay and postpay to create a continuous learning loop that makes analyzing claims on both sides of the equation fast, accurate and fair.
The industry has a set of approved APIs and guidelines for use that can make the sharing of medical records and other related documents safe and secure, opening the door to AI-assisted medical record review, making the whole process even more efficient. It’s truly an exciting time as health payer organizations rethink how to best identify improper payments and fight FWA.
