Claims Data Can Provide New Insight Into the Condition of Patients With MS

Laura Joszt

Analyses of real-world data have broadened the understanding of multiple sclerosis (MS) and provided a snapshot into patient conditions and healthcare costs in the years leading up to an MS diagnosis, explained Bruce Pyenson, FSA, MAAA, Principal, Consulting Actuary, Milliman, Inc, during a session highlighting findings of a recent Milliman white paper at Asembia’s 15th annual Specialty Pharmacy Summit, held April 29 to May 2 in Las Vegas, Nevada.

Progression MS in patients has been described using the Expanded Disability Status Scale (EDSS), which includes information such as the patient’s ability to walk without assistance for 500 meters or if the patient is restricted to a bed or wheelchair. However, while these kinds of metrics are widely used in randomized controlled trials, they are hard to find in real-world data.

“Clinicians that take care of patients don’t use this scale, don’t record this even in the electronic medical records,” Pyenson said. “…If you try to find things like this in things like claims, it’s almost impossible.”

Knowing this, Milliman has taken the EDSS scale and converted it into markers of disability that can be found in the administrative data—for example, if someone has a prescription for a wheelchair or if they are taking medications for neurological impairments, such as fatigue, anxiety, or depression.

While those 3 disability markers are important for patients with MS, they barely show up in the EDSS, Pyenson noted. He added that one of the limitations of clinical trials is that they don’t necessarily correspond with what’s most important for patient progression or to the patients themselves. To patients with MS, the neurological impairments that are prevalent with the disease, such as difficulty concentrating, difficulty remembering, and pain, as well as the previous 3 mentioned, are important.

“These are particularly important for a commercial population: people who have family responsibilities or people who have jobs and are going to have trouble staying employed or taking care of their families, because of these [neurological impairments],” Pyenson said.

Milliman looked at claims data for adults (18-64 years) in 2 ways: The snapshot analysis found that MS affects 0.21% of the commercial population, with 10% of MS patients newly diagnosed in a year. That proportion of newly diagnosed patients is important because the most intense period of investigation and treatment for patients with MS is that first year, Pyenson said.

The variability for costs for treatment was quite big, he added. For instance, the per patient per month cost, including both insurer-paid and patient cost-sharing amounts, for DMT ranged from $0 for patients in the 10th percentile to $7901 for patients in the 95th percentile, with a median of $5317. For non-DMT costs, it ranged from $89 for patients in the 10th percentile to $6871 for patients in the 95th percentile, with a median of $654.

The longitudinal analysis helped to identify how MS is a progressive condition. It showed that the number of impairments grows for a year or 2 before MS is diagnosed, and by the time it is diagnosed almost half (47%) of the MS population has 1 or more indicators of EDSS-derived disability or related neurological impairment.

Most of the cost is in that first year of diagnosis, but there is an increase in costs during those 1 to 2 years before diagnosis when the number of impairments grows.

“Something is going on with these patients before MS is diagnosed, and that shows up in the administrative data,” Pyenson said. “Interestingly, in the year after diagnosis, costs tend to fall and remain flat after that.”

The implication of the longitudinal findings for payers is that administrative data can be used to follow and identify the condition of patients with MS, as well as the progression and pattern of impairments. The findings show that there are things that are not in the clinical trials that can be found in the claims data, which provides a different view of patient quality of life, outcomes, and more.

“There’s a lot of discussion…about predictive modeling and the value of big data,” Pyenson. “I think you can get a hint of some of the potential for following patients through administrative data to apply some of those techniques and to figure out how to profile patients so they can get better care.”
Print | AJMC Printing...