Insights > Blog

Case Study: Using Big Data to Easily Identify Fraud with Zelis

By Redapt Marketing | Posted on November 29, 2019 | Posted in Microsoft Azure, Featured

A Big Data Solution to a Big Insurance Problem

When healthcare technology company Zelis needed a Big Data solution to easily identify fraudulent claims, Redapt’s data scientists provided them with a framework.

The problem

Zelis needed a Big Data framework for identifying potential fraud in millions of insurance claims.

The solution

Redapt built a platform that allowed Zelis to run models based on historical data to pick out suspicious claims.

The outcome

Zelis is able to employ a flexible rules engine to help them stay in front of new trends and areas of fraud so they can focus on outcome-based healthcare.

Moment of clarity

clarity-developer-code-thinking

By building a framework for our data scientists to run various models, we developed nearly 50 distinct ways for Zelis to process claims and identify potential irregularities based on mountains of data.

What the company needed

A core part of Zelis’s business is processing out-of-network claims between doctors and insurance providers. Flagging cases of fraud, waste, and abuse (or FWA) is a major challenge when millions of claims are being processed from hundreds of different providers.

In order to keep up with the never-ending stream of data, Zelis needed the ability to predict and isolate potential fraudulent claims—a classic needle in the haystack scenario.

Our recommendation

planning-whiteboard-solution

A framework for efficiently sifting through data to identify and fingerprint historic costs related to procedures.

A flexible rules engine that can be modified with new information derived from a constant stream of data.

Leveraging Microsoft Azure for data scientists to run data models efficiently and with far less resources than on-premises.

The end result

Ability to continually improve algorithms in order to stay ahead of new trends in fraud and abuse.

Single pane of glass visibility for data model outcomes based on hundreds of different providers.

Accelerated processing of non-fraudulent claims.