The postdoctoral researcher at Amsterdam University Medical Center talked about a tool that helps assess the likelihood of transitioning to secondary progressive multiple sclerosis, allowing clinicians to make informed treatment decisions in a timely manner.
Research indicates that secondary progressive multiple sclerosis (MS) is linked to a poorer prognosis, making early predictive tools valuable for clinicians in assessing disease risk. A recent study published in Multiple Sclerosis and Related Disorders demonstrated that the DAAE score, a newly developed tool, consistently predicted the risk of transitioning to secondary progressive MS across various international datasets over a 5-year period. Although the DAAE score has proven to be user-friendly, further validation in larger patient populations may be required to establish its reliability for clinical use.
In the study, lead author
In a new iteration of
Tom Fuchs, MD, PhD: To support personalized medicine for patients with MS, physicians and researchers like myself develop disease-monitoring tools to answer questions like, How well is my patient doing right now on the therapy they’re receiving? or What is the best-fit therapy for the patient in front of me at this moment?
What my colleagues and I, both here at the MS center in Amsterdam and in Buffalo, New York, developed is the DAAE score, which predicts the risk of transition to secondary progressive MS over five years. It helps answer the question, How likely is it that my patient’s disease will worsen in the years ahead? We focused on secondary progressive MS because these patients accumulate disability more rapidly, respond less to the rehabilitation therapies we have, and no longer benefit from most of the disease-modifying therapies available. Predicting this outcome is important.
The DAAE score is similar to how a weather report might tell you the probability of rain, helping you decide whether to take an umbrella. In this case, the DAAE score helps us ask, How likely is my patient’s disease to worsen? We can then make informed decisions, such as starting or stopping a particular therapy or targeting high-risk patients for timely rehabilitation. That’s the purpose and driving goal behind the DAAE score.
The main takeaway from our research is that we succeeded, after a lot of hard work, in creating a clinically usable tool. I’ll walk through the stages, but first, the end result: the DAAE score can be used in a clinical setting in under one minute. The score is based on the final predictors in our model—disease duration, age, age at disease onset, and the Expanded Disability Status Scale, which measures physical disability.
All of this information is readily available in clinical environments. By inputting it, you get a score from 0 to 12 points, which allows you to assess your patient’s risk of transitioning to secondary progressive MS over five years. For example, a patient in a low-risk group might have an 8.2% chance of transitioning in the next five years.
This predictive algorithm, which can be used in less than a minute, is available for free online. We’ve developed an app that can be accessed on a smartphone or desktop, and for those without access to technology, we also created a paper and pencil version with charts for calculating scores. We’ve already tested the DAAE score with physicians in over 20 countries, and they’ve rated it highly for usability and usefulness. We’re confident in how it works.
The development process, however, was much more complicated. We broke it into four stages. First, we identified predictors of transition to secondary progressive MS by reviewing over 60 publications. Then, we performed feature selection, choosing the predictors that worked together without overlap. Using machine learning techniques, we tuned the weight of each predictor for the final model.
The most important step was testing the algorithm across multiple datasets. We developed it using data from the U.S., then tested it on entirely separate datasets, including one from another country. This ensured the prediction tool worked consistently and was generalizable. Once we were satisfied with it, we converted it into a usable format—both as an app and as paper reference tables. It’s similar to existing predictive tools in neurology, like the CHADS2-VASc score.
I’d like to highlight our next steps. Even though we’ve made a significant achievement, we’re already looking at how to improve the DAAE score. I’m collaborating with colleagues in Amsterdam, Buffalo, and other countries to further develop the tool. Many physicians who’ve used it say it would be helpful if the score integrated information about current disease-modifying therapies or MRI findings, such as what's happening in the brain and spine. Knowing that this information would help physicians make more informed decisions, we’re working to incorporate it into the score with international collaboration.
Once that’s done, we believe it will be even more useful for personalized care in MS. We expect to release this updated version within the next year.
Lastly, I’d like to emphasize something for others interested in developing what I call “translation tools”—tools that can be immediately implemented in clinical environments. When building clinically usable tools, we need to consider the limitations of a clinical setting. Physicians often have no more than 10 minutes with a patient, and they may not have access to advanced software or Linux-based systems. These constraints can prevent scientific advancements from being used in practice.
When developing the DAAE score, my colleagues and I were mindful of these limitations, and I believe we succeeded in creating a tool that’s usable today—not just in the future. I encourage other scientists working on their own projects to keep these constraints in mind as they develop clinically usable tools.
Transcript edited for clarity.
