Only the Beginning: Crackdown on Health Care Fraud to Accelerate in Wake of COVID-19

The COVID-19 pandemic has ushered in a new era of health care fraud, making it critical for health plans to ensure the integrity of payments to providers.

The US Department of Justice (DOJ) recently charged 138 medical professionals with committing $1.4 billion in health care fraud using telehealth services, COVID-19 relief funds, substance abuse treatment facilities, and illegal opioid distribution schemes across the country.

This comes after the DOJ recently established the COVID-19 Fraud Enforcement Task Force, which marshals the resources of several agencies across the government to enhance enforcement efforts against COVID-19-related fraud.

Earlier, the US Centers for Medicare and Medicaid Services (CMS) created the Medicare Advantage Risk Adjustment Data Validation program to beef up audits of Medicare Advantage insurers. For 2022, CMS also doubled its budget for fraud, waste, and abuse (FWA) investigations and requested an additional increase for 2023 with a focus on innovative efforts supporting prepay claim review, predictive analytics, and states’ abilities to detect and deter fraud and abuse.

CMS’ investment in combating FWA in Medicare Advantage is consistent with much of the rest of the industry. For example, the global health care fraud analytics market is projected to reach $6.65 billion by 2027, with significant growth between now and then, driven by increased spending on health insurance and government spending to stop fraud, according to Emergen Research. The predictive segment is expected to lead the market, expanding at a compound annual growth rate of 28% during the same period.

Analyzing Patient Records for FWA

Much of the recent activity around FWA can be attributed to fraudsters attempting to capture COVID-19 relief funds from the federal government and collect illegal, inappropriate payments from public and private health plans. In health care fraud, the most common schemes revolve around billing for unnecessary services, applying incorrect coding and billing, and identity theft.

For payers, this surge in fraud has created a situation in which they must wade through massive amounts of patient electronic health records (EHRs), claims, and bills to resolve expenses and identify instances of FWA. As payers are performing these tasks, they are likely to spot an unprecedented amount of FWA as a result of COVID-19, as recent trends are likely to continue into the foreseeable future—whether the pandemic is “over” or not.

How successfully health plans manage these challenges will be determined by their ability to replace time-consuming and expensive manual processes with artificial intelligence-based tools that comb patient records to identify potential fraud, assess patient risk, and confirm payment accuracy.

Three Ways Natural Language Processing Improves FWA Detection

In the traditional approach to payment integrity, payers employed time-consuming and expensive chart reviews to identify and extract important patient data from EHRs, such as whether a patient’s diagnoses and symptoms support the need for certain COVID-19 tests. Today, however, an increasing number of health plans are employing natural language processing (NLP) technology, as opposed to manual chart reviews.

NLP is a type of technology based on artificial intelligence that reproduces the human ability to interpret language, minus the errors sometimes associated with fatigue and bias. NLP enables payers to analyze large, unwieldy chunks of health care information to identify specific clinical data that may raise red flags and alert the need for further investigation. By extrapolating meaningful data points from thousands of pages of medical records, NLP can present key insights for reviewers to quickly make determinations on FWA outcomes.

With the greater amount of scrutiny on the payment integrity of Medicare Advantage plans and FWA, NLP will play a growing role in enabling payers to pinpoint fraud before CMS audits do it instead. Here are 3 ways payers can rely on NLP to boost FWA detection:

#1 Identify Patterns

In FWA cases, patterned anomalies are often observed, such as a high proportion of patient records that share recurring templates, duplicate charges, or rare diagnostic clusters, for example. NLP enables health plans to detect these patterns.

#2 Spot Outliers

Additionally, NLP enables health plans to pinpoint anomalous data that could indicate fraud, like costly tests that are not medically justified. Because it helps payers comb through large amounts of unstructured data to identify contradictions in patient records, NLP can show whether critical data is contained in records.

#3 Achieve Scale

Even the most hard-working humans experience limitations in performing large amounts of chart reviews in a narrow timeframe, but NLP automates reviews, driving substantial improvements in scalability. Because complex medical documents often run as long as thousands of pages, NLP helps payers perform reviews more quickly and cost-effectively.

For payers, concerns about fraud, waste and abuse will continue to proliferate, and the days of relying on tedious and expensive chart reviews will no longer be an effective method to combat the problem. Artificial intelligence-based technologies such as NLP help payers more quickly and accurately detect instances of FWA at scale—without CMS audits.