It’s finally official: Around 900,000 Americans and quite possibly more than that have multiple sclerosis (MS) — easily double the long-accepted figure of 400,000.
Since MS News Today first reported on this finding in November 2017, the National Multiple Sclerosis Society (NMSS) study, which reached that conclusion, has undergone a peer review process and has been accepted for publication in Neurology, the journal of the American Academy of Neurology (AAN).

It appeared online Feb. 15 in the form of three articles: “Validation of an algorithm for identifying MS cases in administrative health claims datasets,” “The prevalence of MS in the United States: A population-based estimate using health claims data,” and “A new way to estimate neurologic disease prevalence in the United States: Illustrated with MS”. The print edition will be published March 5.
The $1 million NMSS study — nearly five years in the making — followed complaints that scientists had deeply underestimated the number of people living with the disease.
“After a lot of time hearing about how the original estimates were based on the best science at the time, our supporters felt we weren’t getting everyone,” Tim Coetzee, who heads advocacy, services, and research at the NMSS, said in a phone interview.
Join the MS forums: an online community especially for patients with MS.
“This study and the innovation behind it give us confidence that we now have a stronger representation of the number of people living with MS in the United States,” Coetzee said. “This gives us the ability to engage with policymakers, legislators, and companies about the impact MS has on people, and talk about the burden of the disease.”
Women now make up 74% of MS patients
Coetzee also said the unique approach pioneered by the 15 or so researchers involved in the study can be applied to other neurological diseases that don’t have national registries, such as Parkinson’s or Alzheimer’s.

“This taught us we could use existing data from government and private insurance companies to be able to get to an estimate fairly modestly,” he said. “The algorithms can be adapted. You’d need to tweak it for that particular disease, but it can be done.”