2012 Innovation Awards: Next Generation Predictive ModelReprints
Liberty Mutual Insurance Co.
Next Generation Predictive Model
When Boston-based Liberty Mutual Insurance Co. launched its first workers compensation predictive model in 2004, the technology was limited to simply correlating reported injuries with high claim costs.
That's useful information, said George Neale, executive vp and general manager of commercial claims for Liberty Mutual, but it falls short of addressing a critical problem facing workers comp benefits managers: claims that begin as relatively minor injuries or illnesses but become high-cost problems over time, largely because of other undiagnosed or untreated medical conditions.
“You have issues with claims that come through the door that are not immediately evident at the time of the first report,” Mr. Neale said. As a result, benefits managers, insurers and third-party administrators aren't able to effectively plan for the potential impact of seemingly unrelated conditions or health risks on a claim's long-term cost.
“It's very easy to have issues hidden within that information,” Mr. Neale said.
Liberty Mutual's Next Generation Predictive Model, which earned a 2012 Business Insurance Innovation Award, launched in August 2011, seeks to illuminate those issues by analyzing lost-time claims and individual medical bills, patient history and internal research. Designers said the new predictive model can provide not only greater clarity as to the medical and nonmedical conditions that can prolong workers comp claims, but an opportunity to mitigate those conditions early on.
“It puts us in a much stronger position to focus on that small subset of cases that have the greatest opportunity to become high-cost claims,” said Maureen McCarthy, senior vp of workers compensation and managed care at Liberty Mutual.
Ms. McCarthy said the model—which is offered free of charge to all Liberty Mutual workers comp insurance customers—compares an incoming workers comp claim and the medical history of the injured worker to the more than 825,000 lost-time claims Liberty Mutual has on file. Depending on the type of injury and the number of identifiable ancillary risk factors, the program can produce a probable estimate of a claim's total cost and duration less than 30 days from the initial filing date.
Ms. McCarthy said she often cites the case of a mechanic who suffered a joint laceration to the pinky finger on his dominant hand. Liberty Mutual ran the injured employee's workers comp claim and medical history through the Next Generation predictive model and discovered that he had an above-average risk of developing diabetes.
Subsequent meetings with the worker, his employer, his treating physician and Liberty Mutual's case manager revealed that the worker already had developed diabetes and hypertension, and was obese and a heavy smoker, Ms. McCarthy said.
“Based on all of that information, we all decided together to postpone the surgery until those risk factors were under control,” Ms. McCarthy said. “It ended up being a very successful claim, but you never know what would have happened if he had gone ahead and done the surgery without managing those risk factors.”