A data-based understanding of what drives increases in insurers’ claims costs has, for some time, been just out of reach. Data-rich clues like comorbidities or planned medical procedures that could be used to identify high-risk claims or emerging claims trends lie hidden deep within insurers’ claims files, making it hard to access this information.

But much has changed in recent years with the introduction of predictive models that use data mining and machine learning to access nontraditional data. These models have given insurers clear insights into claims cost drivers and the tools to streamline processes. Here are four of their success stories.

Nodal case study #1: A preemptive strike on high-risk claims

Mindful of the potential of new technologies and with a close eye on cost savings, an 80-member public risk pool felt sure it could make a huge difference in its claims costs if it could only identify its high-risk claims and curtail needless costs before they ballooned.

Milliman’s predictive model and decision-support system, known as Nodal, provided the solution. Designed as an early claims intervention tool, the model’s text finding algorithm reads through adjusters’ notes and other unstructured data, teasing out formerly hard-to-access data like comorbidities or upcoming medical procedures that might signal a sharp increase in costs. These claims, known as jumper claims, were assigned to more experienced adjusters.

After the first year of implementation, 19 of the pool’s 20 high-risk claims were identified and the pool saw:

• A 15% decrease in claims severity
• A 40% drop in loss cost
• A 25% decrease in claims frequency

Nodal case study #2: An analytic ecosystem comes to life

Shorthanded in its claims department, a self-insured Fortune 500 company was also battling large increases in its third-party auto liability claims. The staff shortage was likely contributing to the claims cost increases, but the company knew staffing wasn’t the primary reason for the increases.

While a full walkthrough of the company’s processes uncovered the need for some changes in staffing and processes, the primary solution was to build an analytic ecosystem that could tap into its unstructured claims data.

Milliman’s Nodal was able to pull out key information from the company’s claims files that was used to further customize the model. A robust analytic tool was married to the predictive model that allowed the company to perform detailed analyses including identifying key claims trends, comparing severity trends, and cataloguing references to narcotics or different types of comorbidities. In effect, the company had created a self-service analytic tool from which it could uncover specific drivers of claims costs.

Datalytics-Defense case study #1: Unsustainable increases in defense costs halted

Seeing its defense costs increase at an unsustainable rate, a large regional insurer realized its first step had to focus on gaining a much better understanding of the reason it was seeing such large increases in cost. It suspected that its claims were becoming more complicated and depositions seemed to be taking more time, but it didn’t have the data to back up these hunches.

Milliman’s Datalytics-Defense was used to electronically collect the insurer’s defense costs from attorney invoices and other vendors. After ensuring that these invoices conformed to the insurer’s billing guidelines, the Datalytics proprietary text mining algorithm was deployed to extract intelligence that had been hidden in the insurer’s invoices. Once revealed, the new data provided visibility into the insurer’s defense costs and a rationale for decisions that produced savings of nearly 20% in defense costs. Complementing these savings, the insurer was able to repurpose three full-time administrative staff as a result of a more efficient invoice review process.

Datalytics case study #2: E-billing brings efficiency to captive’s growing workload

With approximately $20 million in annual premium, a relatively small captive insurer’s claims staff was struggling to keep up with the growth of its parent company. Claims were increasing but staffing remained unchanged. Constrained by level funding, the captive wanted to streamline the review of its defense cost invoices and eliminate any inefficiencies in its claims handling process.

A review revealed that manually processing its defense costs consumed tremendous time and valuable resources. Invoice processing could be much better managed with Datalytics-Defense’s e-billing system, a fully integrated system that allows attorneys and other vendors to directly upload their invoices into Milliman’s web-based portal. After the invoices were presented to and approved by the captive, a simple click initiated a transfer of all payment details back to the captive and into its accounts payable system. The implementation eliminated all the manual keystroke entries the captive’s staff had performed, freed adjusters to devote more time to managing claims, and allowed the captive to maintain its current staffing levels while handling an increased workload.