Can a smartphone + AI predict my MS future?

My smartphone can do just about anything. Could it also predict my
multiple sclerosis (MS) progression?

Maybe.

Scientists at Google have teamed with others at the University of California
and the Duke University School of Medicine to develop a way to use
artificial intelligence (AI) learning methods to analyze data collected by
smartphones and wearable devices. They report they’ve been able to predict the appearance of five, high-severity MS symptoms – fatigue, sensory disturbance, walking instability, depression or anxiety and cramps/spasms – three months before they’re actually experienced. They say they can also predict the symptoms’ future severity.

Moving data collection outside the clinic

Currently, MS care relies, in large part, on data that’s obtained by reviewing magnetic resonance imaging, in-clinic laboratory exams and tests, and clinical histories. But months may pass between those tests and, during that time, subtle physiological changes that occur may be missed. This is where AI, smartphones, and wearable devices can team up to help, by keeping track of tasks such as walking, dexterity, and balance during those gaps.

Using an app called MS Mosaic, which was downloadable during the study, participants performed versions of several MS-related tests: The timed 25-foot walk, the two finger tapping test, the Paced Auditory Serial Addition Test (PASAT), the nine-hole peg test, and a spatial memory test. The app also tracked symptoms and some participants also used Apple’s Health app – the standard app that comes with the iPhone - to measure step counts, sleep data, and heart rate.

Knowledge is power

Researchers conducting the Mosaic study think that by predicting high-severity symptoms before patients actually notice them people with MS (PwMS) could better anticipate and manage them. That, in turn, would likely improve their life quality. Forecasting a change in walking ability or a sensory disturbance, for example, could allow a health care provider (HCP) to better prepare a patient to handle those anticipated symptoms, or the HCP could be more proactive at prescribing a therapy. The researchers think combining smartphone tracking with AI analysis might also give those HCPs a leg-up on which new symptoms are likely to appear, why a patient is experiencing current ones, and what treatments are best suited to prescribe. In their paper, the researchers say this approach has “the potential to empower subjects as experts of their own experience [my emphasis] in order to improve symptom management.”

I love that! I’ve always believed that PwMS need to be actively involved in their treatments. I’ve also used some symptom tracking apps that are already available, such as iCompanion and My MS Manager. Although they don’t analyze the data they collect, I’ve found them to be useful to chart my symptoms for later discussion with my neurologist. The Mosaic MS app could provide an even greater level of symptom management empowerment for those of us who live with this illness.

(A version of this post first appeared on the Rare Disease Advisor website.)