Machine Learning Times
Machine Learning Times
EXCLUSIVE HIGHLIGHTS
Three Best Practices for Unilever’s Global Analytics Initiatives
    This article from Morgan Vawter, Global Vice...
Getting Machine Learning Projects from Idea to Execution
 Originally published in Harvard Business Review Machine learning might...
Eric Siegel on Bloomberg Businessweek
  Listen to Eric Siegel, former Columbia University Professor,...
Effective Machine Learning Needs Leadership — Not AI Hype
 Originally published in BigThink, Feb 12, 2024.  Excerpted from The...
SHARE THIS:

This excerpt is from Propertycasualty360. To view the whole article click here.  

9 years ago
How Predictive Analytics can Improve Claims Outcomes

 

If insurers could somehow forecast the future, what a safer, more productive, and more profitable future it might become. Loss prevention departments could greatly reduce both claim frequency and severity, while growth could be focused to mitigate catastrophe risk.

While we await a crystal ball, all kinds and sizes of losses are taking place. The good news is that tools are now available to help shine a light on important issues within the insurance industry. Big data and predictive analytics have held a firm foothold in underwriting and actuarial departments for years, and many are finding they also have a place in handling Workers’ Compensation claims. As models are refined, more insurers are learning how to improve outcomes and their bottom line.

The key to successfully incorporating a predictive model is to target opportunities where data can give a claims organization the biggest lift, with the lowest-hanging fruit clearly being claim severity. At the National Council on Compensation Insurance 2015 Annual Issues Symposium, 32 states reported an increase in average annual Workers’ Compensation claim severity between 2009 and 2013, and this is a trend that deserves attention. From claim assignment through settlement, identifying claim severity early and accurately can help put claims departments in the driver’s seat, allowing a proactive and informed approach to claims handling.

When claims spiral out of control

Even the most seasoned adjusters occasionally find that claims can spiral out of control. This can happen for a number of reasons: heavy caseloads, frequent file transfers, lack of information and inexperience in spotting the severity indicators of a particular claim. Assigning the right claim to the right adjuster is only half the battle. The adjuster must have the experience and skill necessary to identify severity indicators early on in the life of the claim.

The adjuster must also devise an appropriate plan to bring the claim to resolution and get the injured claimant to maximum medical recovery and then back to work, if possible. No matter how deep their experience, adjusters will likely encounter claims with unfamiliar characteristics. For a newer adjuster, this can be daunting. Controlling costs becomes even more difficult when claims are reassigned and each successive adjuster must make a seamless transition in identifying the proper steps to take in handling the claim. With each reassignment, the likelihood of potentially severe claims falling through the cracks — resulting in skyrocketing costs — increases exponentially.

Using data and analytics to improve outcomes

Spotting high-risk claims early in their lifecycle can be key to mitigating costs and improving outcomes for all parties involved. With the increasing availability of big data, it has never been easier to incorporate predictive modeling into the claims process. This changes the process from relying on the experience of a single adjuster to drawing on the claims experience at the company level, which may mean hundreds of thousands of claims — or even to the industry level, which may involve millions of claims.

The challenge of correctly identifying many potentially severe claims early-on can confound even the most seasoned adjusters. Data supplied by a solid predictive model can accelerate experiential learning and provide a safety net for adjusters by providing data-driven indicators to flag claims that may spiral out of control without proper attention. Pairing severity identification with business process can yield even greater results.

Given a predictive model providing a range of severity scores, for example, it makes sense to assign the least severe claims to newer adjusters, providing them with low-risk claims so they can learn the fundamentals of claims handling. As experience levels rise, so should the complexity of claims assigned. This allows adjusters to deepen their experience level while optimizing claim outcomes for claimants, employers and insurers.

Beyond triaging claim assignments, severity thresholds can be built into the claims handling process. For example, auto adjudication may be incorporated for the lowest-severity claims, while actions such as mandating nurse case management and increased managerial oversight may be considered for the highest-potential claims. This allows a claims department to focus its time, energy and expenses where they can have the greatest impact on high-severity claims.

This excerpt is from Propertycasualty360. To view the whole article click here.

By: Adam Wesson, director of claims solutions, ISO Claims Partners
Originally published at www.propertycasualty360.com

Leave a Reply