Printed from BusinessInsurance.com

Predictive models flag problem claims

Posted On: Jun. 5, 2017 12:00 AM CST

Some workers compensation insurers are doing what the property insurance industry does when it examines a building’s potential to survive catastrophe: examining the person and the injury and predicting whether there is a likelihood of problems later on.

One model is Travelers Cos. Inc.’s Early Severity Predictor program, which helps payers identify individuals whose injuries and prescriptions might prove problematic later on, said Rich Ives, Hartford, Connecticut-based vice president of claims for workers compensation with the insurer.

It’s just one approach to tackling opioid prescribing that’s gaining steam.

“The flagging model is not our approach,” he said of a common practice among insurers of flagging opioid claims at the onset. “We’ve gone further to say there are certain variables that have a predictive quality (for opioid abuse or dependency).”

For Travelers, the process is personal and involves a battery of questions for the patient. For example: Do you have a chronic condition? What other medications do you take? Have you dealt with pain in the past? What is the type of injury? Where are we three weeks out from the injury? Are you complying with the medical guidance?

“From there, we reassess them,” said Mr. Ives. “We gain a little bit of that human element, how they’re coping. If we truly believe they are not making progress, (managing the prescription) would be a collaborative effort.” Patients are constantly reassessed, he said. “The model always runs in the background,” he said of the oversight that has been applied to 36,000 claims in the past 18 months, with 50% of them seeing improved outcomes.

ESIS Inc. uses a similar approach, with its risk-scoring program examining the facets of a claim that might predict problems later on, according to Deb Gleason, the company’s Philadelphia-based clinical resources manager.

“There are algorithms that manage and monitor an injured worker’s score,” she said. “(For example), if you had an injured worker who at 120 days was on both short- and long-acting narcotics, they would have a high score (and) our team would come back and make a clinical recommendation.”

Recommendations can include weaning or other therapies, she said.

What goes into the score includes everything from medical history and type of injury to the state the worker lives in and other demographics — as some locales have a higher probability for opioid abuse, she said.

“With this we can identify early on whether a claim has the potential to go south,” he said. “That’s really a key driver in early intervention.”