Rob Rolf

Using analytics to predict, prevent and recover improper medical claims

Around the world, millions of times a day, clinicians, hospitals and pharmacies submit claims to healthcare plans. In the U.S. alone, nearly $3.2 trillion dollars in payments flow from public and private payers to providers annually. These payers are under constant pressure to process and pay claims quickly, leaving little time for complex analyses of large numbers of claims. While most reimbursements are straightforward, some should not be paid because they are erroneous, inflated or even fraudulent.

Specialized software using advanced analytical algorithms can help healthcare payers find improper claims, prevent a payment or recover a payment already made. Different claim situations call for different analytical approaches, including outlier analysis and predictive modeling. Machine learning is a particularly promising technology—that’s where self-learning models use known outcomes to teach themselves what is a true positive and what is a false positive, further automating the decision-making process. As big data becomes ever more integrated with cognitive analytics, machine learning and the dynamic computing power of the cloud, more payers will be able to predict fraud even before it happens while maintaining compliance with prompt pay regulations.

Already today, advanced algorithms can test claims against business logic to:

  • Identify hidden patterns and anomalies within a payer’s entire claims universe, such as counter-to-policy, unusual or logically impossible combinations of procedures and codes
  • Uncover trends and outliers to suggest potential areas for additional investigation
  • Reveal patterns such as doctor-shopping and excessive provider services in the business day
  • Highlight abnormal behavior between providers and laboratories
  • Indicate which claims to suspend, reject at pre-payment, or audit and recover
  • Provide insight for cases with the highest potential for recovery more than doubling the accuracy of traditional rule-based algorithms and correctly eliminating false positives 98% of the time

Additional data sources have the potential to further enrich claims data for these kinds of analytics. This includes electronic medical records that could provide an independent verification of the diagnosis and procedure when matched against a claim automatically. Sources like death records, residency records and driver licenses also could help verify that patients are who they say they are through list matching and link analysis.

As long as there are insurance reimbursements for claims, there will be erroneous, inflated or fraudulent claims. As an IT and business process services partner working on behalf of healthcare payers globally, CGI provides expert services and advanced analytics in our CGI ProperPay solution to help ensure integrity in the claims process. 

Blog moderation guidelines and term of use