Travelers’ New Predictive Model Targets Chronic Pain, Opioid Use

In response to rising chronic pain claims, Travelers announced the development of a new patent-pending predictive model designed to identify claims and workers most at risk.

The company, which developed the Early Severity Predictor (ESP) program, manages more than 250,000 workplace injury claims and 3.5 billion medical treatments each year.

“One of the issues that we’re dealing with over the years has been an increase in chronic pain cases that we’re seeing,” said Dr. Adam Seidner, national medical director at Travelers. “We have a major case unit, which is basically our catastrophic claims. Most of the claims that go into there are brain injuries, spinal cord injuries, amputations, major burns, things like that.”

An analysis of claims Travelers conducted a couple of years ago revealed that more than half of the claims going into the major case unit were related to chronic pain.

“The only way that they could get into the unit is if they had a million dollars or more reserve on them,” Dr. Seidner said. “We said this is obviously a trend that’s been problematic.”

As a result, the insurer looked for ways to prevent injured workers from developing chronic pain. The company started with a flagging model but moved to a regression model so that the insurer could gauge the number of false positives and false negatives that were obtained with the previous model.

The model incorporates data like worker injury attributes, demographic information and claims information obtained through medical records. In addition, pharmacy and billing data are also included in the modeling. This information, in turn, may “trigger what is driving the potential for that individual to have a chronic pain case,” Dr. Seidner said.

While chronic pain impacts industries across the board, there are certain industries where workers are more likely to suffer from it.

“I would say that the manufacturers and other labor-type markets, we see more claims in and greater potential severity. The office workers, the folks that are more service industry that are not physical are less frequent. We still see chronic pain in those areas, if in fact, they are not managed appropriately,” Dr. Seidner explained.

He said the model helped identify the four main drivers of chronic pain.

Co-morbidity is one factor driving potential trends toward a chronic pain claim. Another factor the insurer found to be relevant was the nature of a worker’s musculoskeletal health – athletic versus sedentary. Pharmacy and a worker’s mental health status could also determine whether a claim would lead to chronic pain.

Once flagged the company adjuster and/or nurse assigned to the claim is notified.

“From there, we have contacts with employers, with the injured party, the employee as well as the employer. It’s all voluntary,” said Dr. Seidner.

He cited an example of a conversation with a worker who had a co-morbidity of diabetes flagged. During a conversation with the employee it was determined that despite being diabetic, the disease is treated with diet alone and his or her sugar levels revealed control.

“We would say, ‘OK, that comorbidity really isn’t of concern here because it’s a severity issue.’ If it was someone else that had multiple, high hemoglobin A1Cs, meaning poor control of their diabetes or they were on a lot of medicines…that might indicate greater severity and also may translate into a greater propensity for difficulty in healing soft tissue, and therefore produce pain,” said Dr. Seidner.

Not all states allow conversations with clinical providers without special permission.

“As long as we had no ex parte rules or we have that special permission we would be able to have a conversation with the doc, just to give him a head’s up on some of these issues and get their feedback in how they might want to best manage the patient to make sure that they don’t wind up in a chronic pain situation,” explained Dr. Seidner.

In addition, the predictive model reviews a variety of pharmaceuticals that control pain.

“It doesn’t take long for someone to get physically dependent on…opioids. It’s important to understand pain, the types of pain and really how quickly, within a couple of weeks, someone may be physically dependent on those medications if they’re taking them regularly,” he said. “Very frequently, they’ll experience withdrawal if they’re not tapered off of them.”

Dr. Seidner emphasized that medicine isn’t the only way to treat pain.

“It’s not reasonable to treat all of these folks with an opioid which is very addicting, and is not necessarily the best way. We’ve got a lot of folks with pain. We need to better understand how to manage it,” he explained. “We need to tailor it to the individual, not just their condition but maybe for their preferences as well, so that we don’t have them physically dependent and actually getting a negative return on using some of the medications such as the opioids.”

The model reviews codes for physical therapy, imaging and durable medical equipment.

“If someone was getting neuro modulation from a TENS unit, a transcutaneous electrical neurostimulator, that can also help people with pain in certain situations,” said Dr. Seidner.

Pharmacy utilization, underutilization, overutilization and missed opportunities are all part of that pharmacy analysis, he said.

Since early 2015, the model has been applied to more than 20,000 cases. Of those, the insurer reported that more than 9,000 injured employees were identified as being at risk of developing chronic pain. As a result, the employees underwent a sports medicine-like regimen of treatment to aid and accelerate their recovery.

Much like an athlete who gets injured and is quickly treated on the sidelines in order to get back in the game, Dr. Seidner said the same can be said for an injured worker.

“We want to get the person identified as early as possible so that the proper treatment can be started as soon as possible. Whether an image, physical therapy or other treatment needs to be done, getting that diagnosis early on, getting that person with the proper specialist is critical,” he explained.

The model offers a number of benefits, he said. Injured employees who participated in the program in the past year have, on average, recovered and returned to work more quickly. They were also far less likely to receive a prescription for opioids, and when they did, it was typically a lower dosage or only for short-term use. Medical expenses, which cost American employers an average of nearly $40,000 per injury according to the insurer, were reduced by nearly 50 percent.

“We’re really an agent for the employer, if you will,” said Dr. Seidner. “The employer is looking to make sure that they’ve got a healthier, productive employee…and not have issues around presenteeism, where someone is there, but not productive or even just being absent. We’re just helping to use the right resource at the right time and avoid any of the waste that can be part of the system. Using the wrong services, either at the wrong time or when they’re not necessary, that’s still waste.”

For now, the company plans on running the model for a few years before considering whether it will be licensed out to other insurers.