Pharma companies annually invest more than $12 billion to provide co-pay cards that ease out-of-pocket costs for prescription drugs and the people who need them. For healthcare consumers who are underinsured or require financial assistance, these programs help make essential medications more affordable and accessible. They also make tempting targets for fraudsters. To mitigate losses, pharma is deploying advanced analytics and AI-led fraud detection systems and modernizing validation processes.

At a leading U.S.-based biopharmaceutical company, managers suspected its co-pay card program for a specialty therapy had fallen victim to fraudulent claims after hypothesized business rules identified anomalous pharmacy co-pay transactions. Spotting and stopping co-pay card fraud remains a persistent challenge for drug companies because of the sheer number of transactions, making it hard to check each one closely. Fraudsters often team up with pharmacies and healthcare providers to create fake prescriptions and claims. They mix fake prescription fills with real ones, making it hard to tell them apart.
 

Fraud can come from both outside and inside a pharmaceutical company. On the outside, healthcare providers and pharmacies might take advantage of the system. On the inside, bad data management and outdated processes can make things worse. Traditional methods, such as basic rule-based systems or manual checks, often cannot catch complex schemes. The lack of advanced analytics makes spotting sophisticated fraudsters and uncovering hidden patterns even harder.
 

Poor data management exacerbates the problem because messy data and weak monitoring allows fraudulent claims to go unnoticed. The constantly changing nature of fraud, especially with expensive specialty products, adds another layer of difficulty. Effective fraud detection hinges on teamwork across different drug company departments that can be tough to achieve and maintain.
 

To tackle these issues, drug companies increasingly turn to AI, machine learning (ML) and big data analytics. These tools boost their ability to detect fraud and reduce revenue loss, offering a stronger defense against the ever-changing threats.

ZS developed a proprietary ML-based approach to flag anomalous transactions. Primary objectives were to identify and characterize pharmacies associated with anomalous co-pay-card transactions and identify and profile healthcare providers and prescribers associated with anomalous transactions.
 

The system analyzes large data sets to detect unusual patterns and anomalies that indicate fraudulent activities, such as high rejection rates for payer claims and frequent fills per co-pay card. The three-phase approach includes:

• Assessment. Aggregating data and reviewing product spend to estimate revenue leakage risk and identify vulnerable areas.
• Quick concept (proof of concept). Demonstrating AI’s value in detecting discrepancies that routine reviews might miss, often revealing gaps in third-party vendor processes.
• Scale and operationalize. Building and scaling an anomaly detection engine that continuously monitors new data, generates alerts and prevents misuse. This system integrates human feedback, creating a semi-supervised approach that adapts to evolving fraud patterns.

The investigation into pharmacy misuse exposed significant fraudulent activities, prompting several important actions and policy changes. It identified 70+ pharmacies linked to misuse. Among these, 20+ pharmacies were involved in approximately $25 million worth of fraud and were subsequently blocked to prevent further financial losses. Additionally, approximately 50+ pharmacies were placed on a watchlist for closer monitoring. Legal proceedings were initiated against a co-pay vendor to recover lost funds.
 

To prevent future misuse, eligibility criteria were revised. New validations for benefits exceeding a certain limit help ensure only eligible participants receive them and mitigate the risk of fraud. Enhanced eligibility measures support a more secure and reliable co-pay program that doesn’t burden recipients or stand between them and their critical therapies.