TDWI | Training & Research | Business Intelligence, Analytics, Big Data, Data Warehousing

"Stream"-lining Healthcare Claims Processing

Healthcare must innovate to meet the quantity and quality demands of claims submission and processing. Moving to real-time big data is a critical step for companies that wish to streamline claims processing.

• By Jake Dolezal
• May 6, 2016

Healthcare is a tough business these days. Practitioners and insurers are under enormous pressure to streamline their operations. The battle for efficiency and cost savings is won or lost in the trenches of claims processing, which represents one of the most intense, time-consuming, and costly workloads a healthcare organization can face.

A decade ago, large healthcare organizations deployed electronic data interchange (EDI) claim submission solutions and automated systems to accelerate claims filing and processing. This advance allowed organizations to accommodate the massive volume of claims sent or received every day.

Today, organizations are still struggling to handle the quantity as well as the quality of automated processing.

In many cases, when the quality of claims processing declines, claims drop out of the automated queue for manual processing. Even worse, some claims are processed and paid or denied incorrectly -- causing claims to come back for reprocessing. This can result in delays in payment for healthcare, generate possible regulatory turnaround-time infractions, or even add interest or penalties on top of the claim payment. It is a lose-lose situation for providers and insurers.

With millions of claims per day, a healthcare company is not just in the healthcare business but in the business of data. With these massive amounts of data and such high stakes, companies wishing to control costs and maximize claims production must leverage the cutting edge of information management.

Many companies have turned to Hadoop to store their claims-related data, using it to meet regulatory compliance requirements, track the progress of claims, and improve the claims and billing process through analytics. The most innovative companies are taking this one step further.

The cutting edge in healthcare claims processing is using streaming big data to analyze claims processing and automation in real time. Streaming data technologies, such as Kafka, Kinesis, Storm, Flume, and Spark Streaming, offer opportunities to "stream"-line healthcare claims processing. By treating claims as a real-time data stream, companies can make claims data instantly available to all parties who can react to and course-correct a troubled claim while it is still in the processing flow.

By inserting predictive analytics to identify claims that are likely to be denied or rejected in the approval and examination process, claims can be handled by fewer people, processed once, and paid faster. With the processing power of a technology such as Spark Streaming, potentially millions of claims and claim details can be accommodated in a small time window.

In addition, implementing streaming data means that claims processing goes from sending or receiving claims in large batches a few times a day to tiny batches every few minutes.

Using the messaging capabilities of a technology such as Kafka, it is even possible to alert a provider or claim submitter of an issue that will lead to denial -- missing or invalid information, coordination of benefits, duplicate claims or procedures, etc. -- within minutes of submission. This could allow the submitter to correct the claim and resubmit it within the same day, even before the insurer or claims processor touches the claim for the first time.

The barrier for many companies' leadership against considering this type of solution is the need to shift both their architecture and their thinking. They must take a hard look at their architecture and ask themselves if it is designed for the past or the future.

Companies struggling to streamline their claims submission and processing workflows may be fighting an uphill battle by relying on EDI and relational database automation solutions. Thinking outside yesterday's IT architecture box and looking at what's available today will create new possibilities once inconceivable.