What’s Wrong with Payment Integrity in Healthcare

payment integrity in healthcare

A typical healthcare payment scenario: Your payment integrity vendor has been working on a case for the last two and a half years. They’ve finally completed the medical record review, identified an extrapolated overpayment of over a million dollars, and mailed the audit report to the provider. Then you receive an angry, agitated call from your provider and find out your vendor has spent thousands of dollars in man hours identifying an issue covered by your payment policies. And what about your claim audits? Have you ever run analytics on your claims processing to see how many claims are being paid improperly?

Payment integrity in healthcare is one of the biggest challenges facing healthcare payers. It’s possible that 20% or more of improper payments are missed by traditional payment integrity solutions, leading to over $500B in unidentified losses for healthcare payers each year. It’s time for a change in the function of the payment integrity industry to identify and stop more improper payments.

Importance of Payment Integrity

Payment integrity is crucial in the healthcare industry. The FBI estimates insurance fraud costs the average family between $400-$700 in premiums annually.1 And as RevCycleIntelligence notes, “fraud in healthcare is not as clear cut as in other industries.”2 Based on analyzing market data, we estimate having a solid payment integrity program in place could help save 10-20% of healthcare costs by identifying and stopping improper payments on claims.

Recovering funds from claims already paid, aka post-pay recovery or “pay-and-chase,” is even more complex and costly. Determining what claims were overpaid often involves an investigation into individual providers (which could be hundreds or thousands), a statistically valid random sample, review of the medical records for each of the sampled claims, and extrapolation of any overpayment identified to the remaining claims in the target subset. This process includes the efforts of an investigator, a claims auditor, and a statistician, and typically takes months to complete.

From there, the overpayment is communicated to the provider, typically via an Audit Recovery Letter, which can be 20+ pages long and takes hours to produce. Providers often request education calls and negotiate the overpayment, reducing the amount the health plan can recover. Collection of the overpayment may be immediate or could take years to recoup. Offsetting claims is one option, but if the provider then stops accepting the insurance, how is the payer to recover the lost payments? These losses are passed on to consumers through increases in premiums and deductibles. This places the importance of capturing wasteful and abusive claims early in the claims process.

Cutting Through the Red Tape

Traditional fraud, waste, and abuse (FWA) vendors operate using a structured library of schemes or rules of commonly known FWA activities. To stay relevant, these libraries may be updated annually with new rules identified by investigators or data scientists throughout the year. Because the rule libraries used by traditional payment integrity vendors are static, they may be missing forms of FWA activities another vendor has already identified and included in their rules library. The process of identifying and building the rules in the traditional fashion is timely, which is why updates are only made once a year. 

Static rule libraries also don’t take into account payer specific payment policies. For example, when it comes to Medicare Advantage plans, the Office of Inspector General (OIG) noted that potentially 20% of denied claims actually met payer policies, causing avoidable delays, creating extra steps in the payment process, and causing provider abrasion.3 So, having static rules can reduce the effectiveness and efficiency of your payment integrity program. 

In an attempt to thwart as much FWA as possible, most healthcare payers stack vendors in pass positions to review claims before and after payment. The cost for each pass position typically increases, thus increasing the overall cost to the payer. The most cost-effective structure for healthcare payers would be to invest in a single solution that could cover both pre-pay and post-pay avenues. To date, the only payment integrity system that has this capability are FWA products that utilize machine learning models.  

Streamlining the Process

So, if pay-and-chase and structured rules libraries are the “old way,” what should healthcare payers be looking for from a vendor?  

A single vendor with machine learning capabilities could have the same effect as multiple vendors manually updating rules libraries, but at a fraction of the cost and with increased accuracy. Include an analytical dashboard component, and your Special Investigations Unit (SIU) will be able to easily target suspect providers, while your provider relations team will know which providers may need educational intervention. 

Unsupervised machine learning models are best for identifying new patterns of potential FWA. Once confirmed by an investigator or coding certified data scientist, labels can be attached, and a supervised model can be generated. The quality of the supervised model can be effectively monitored, and the accuracy could identify and prevent 20% or more of improper payments. With the right quality measures and effective policies, there can be one layer with maximum efficiency and effectiveness for payment integrity in healthcare. 

But healthcare payers should be cautious who they choose to build their machine learning models.  

Choosing the Right Vendor

There are a number of vendors out there right now who claim to have a running machine learning model, but the operation of a high performing machine learning program is a skilled production. Healthcare payers should review potential vendors for verified success with artificial intelligence and machine learning models. Vendors should be able to provide proven case studies with outstanding results. You should be looking to replace all of your vendors, not just add another. Question your vendors’ security protocols. Do they meet your expectations with data governance in information security? Are they familiar with the claims process and security protocols in place when handling protected health information (PHI) and personally identifiable information (PII)? 

A solid payment integrity program recognizes that fraud in healthcare is broken into three categories—fraud, waste, and abuse. Unlike retail and banking fraud, the ability to show intent for the billing of fraudulent healthcare claims is more complex than, say, a fraudulent check. A series of overbilled claims could come down to a glitch in the provider’s claims software, not actual intent to defraud the insurance company.   

The ability to identify which claims fall into the waste and abuse categories could help save health plans hundreds of thousands of dollars. On average, an FWA investigation in healthcare costs $7,500. The recovery on those investigations has the chance to go sideways in the snap of a finger, as mentioned at the beginning of this article. Effectively and accurately identifying suspect claims during the claims processing cycle is the most beneficial way for healthcare payers to stop improper payments before they happen and to reduce provider abrasion.  

Learn how our Payment Integrity Analytics solution can be deployed anywhere in the pre-payment or post-payment lifecycle to help reduce FWA and ensure payment integrity in healthcare.

1 Insurance Fraud, NAIC, June 13, 2023.  

2 What Payment Integrity Means for Providers, How to Avoid Claim Issues”, RevCycleIntelligence, Jacqueline LaPointe, July 11, 2023.  

3 Some Medicare Advantage Organization Denials of Prior Authorization Requests Raise Concerns About Beneficiary Access to Medically Necessary Care,” Office of Inspector General, April 27, 2022. 


Devon Snyder

Devon Snyder

Director, Payment Integrity

Srikant Bishoyi

Srikant Bishoyi

Associate Director, Payment Integrity

Contact Concentrix

Let’s Connect