One of the ever-present challenges facing the Workers' Compensation industry is managing and making sense of the massive and growing amounts of information generated throughout the bill review process. This challenge increases daily with the introduction of new publicly available data resources, cheaper ways to store all this information and new tools that enable new different ways to manipulate information. The vast potential created by new tools and more resources creates a great need to make sense of and efficiently leverage big data.
The advent of big data initially brought the promise of more data intelligence, with the breadth, depth and sheer quantity of available information. However, enterprises everywhere are struggling to incorporate the insights offered by data analysis into their business operations. The "bottom up" approach makes it difficult for organizations to know what to do with the information they cultivate. Instead, they should focus on the questions they need answers for to drive improvement. Starting with the business need in mind, businesses can collect and review information to make better business decisions.
A three-pronged approach to leveraging big data is needed, and to demonstrate this approach, we look at a common need within the insurance industry as an example: How insurers can enable adjusters to quickly identify fraudulent providers and positively affect customer care.
Detailed below, the three-pronged approach harnesses the potential of big data to identify discrepancies within the business workflow enabling adjusters to appropriately focus their attention and efforts. Additionally, adjusters are also able to process and pay the right claims quickly, which will positively affect the overall quality of all care provided as a result.
1. Identify the core need
To identify the places where big data can be used, payors need to look for the areas that require the most effort, such as manual data reviews, where knowledge is not readily or consistently available to everyone involved, or where the business process involves looking at various resources to make a simple decision.
When it comes to identifying fraud, adjusters currently reference several files — some provided by their organizations' special investigations units, others available through paid subscriptions and some published publicly. All are separate reports, available in print or online, and all necessitate additional effort since none are integrated with existing technology and tools.
(Photo: Thinkstock)
2. Bring in data science
When using operational data sources, it's critical to ensure that data collected is "clean" data. Unclean data can be best described as data that has some type of error, and there are multiple ways that this can happen — some of these ways include missing fields of information, having information incorrectly spelled, data typos, different types of punctuation, or even data duplicates. These are just some of the ways that data can become unclean.
Data cleansing entails dealing with these multiple mistakes with the goal of standardizing, resolving the types of issues described here, and resulting in clean data that enables analytical efforts. A clean provider database lets payors give attention to identifying fraudulent behaviors. They can examine operational data, published fraud lists and use adjuster expertise to seek and identify potentially fraudulent patterns.
To begin, reference established common data resources and apply data tools to uncover data patterns and inconsistencies. Analyze these patterns and set benchmarks for "regular treatment" or treatment that is rudimentary and commonly seen.
When it comes to fraud, these "regular treatment" benchmarks need to be established both by state and by country, as treatment patterns vary by region. Such benchmarks make it much easier to identify claims that show inconsistencies. Payors can take the findings revealed by data resources and turn them into codified fraud lists. Rather than relying on a manual look-up, payors should automate whenever possible and expedite the ability to identify known issues, both public and internal.
3. Apply analytics
The clear set of anomalies and data patterns revealed through data science efforts can now be incorporated into an adjuster workflow. The key is to make the analytics actionable. To do this, adjusters should start by mapping out the existing workflow and design the places where the application of the appropriate data would enable better decisions. Taking stock of the current systems and tools in place will show the level of effort needed to incorporate the data analysis at the ideal time.
For example, when it comes to fraudulent claims, adjusters can create a set of services that can be integrated easily into current toolsets with minimal IT support. A customized online portal can let users identify, process or prevent fraud drastically streamlining the bill review process.
In sum, an approach that starts with the end in mind will help organizations track the right data sources to harness valuable information. Applying data science methods will reveal trends and patterns that yield actionable insights organizations can incorporate into an adjuster's workflow.
Vidya Dinamani joined Mitchell in 2011 and currently serves as the vice president of the company's product management and marketing auto casualty solutions division. In this capacity, She guides the product roadmap for Decision Point, Mitchell's medical claims billing software and holds seven U.S. patents for software technology.
© Arc, All Rights Reserved. Request academic re-use from www.copyright.com. All other uses, submit a request to TMSalesOperations@arc-network.com. For more information visit Asset & Logo Licensing.