Algorithms fueled by artificial intelligence (AI) to analyze the data collected on a smartphone application could predict if a person affected plates sclerosis (Ms) will experience a certain high severity symptoms Over the next three months, according to a study.
Scientists believe that this will help patients better understand their disease and work more easily with their health team to make disease management decisions. The study, “Performance of automatic learning models to predict high severity symptoms in multiple sclerosis“, Was published in Scientific relationships.
In the MS, the immune system wrongly attacks healthy parts of the brain and spinal cord. Symptoms and severity of the disease may vary considerably depending on the most affected regions and the extent of damage. As a general rule, people with MS have clinical assessments sometimes a year, as well as a MRIBut these infrequent snapshots may not be enough to fully capture the impact of their lives between visits.
Digital surveillance tools such as smartphone applications could fill this gap and allow the collection of a wide range of daily data. In turn, AI algorithms can analyze data and predicts the progression of the disease to facilitate clinical decision -making.
“Mobile technology allows continuous data collection and can pave the way for forecasting complex aspects of MS such as symptoms and sickness courses,” the researchers wrote. “We could learn what symptoms of the SEP subjects are likely to develop, why they experience those they have and what treatments are more suitable for their current illness.”
Help patients manage their disease
An observational study called Ms. Mosaic (NCT02845635) was initiated to collect data from adults with MP in the United States via a three-year mobile application. The application was developed by scientists from Duke in North Carolina University and was made available for download during the study.
Participants accomplished tasks that included demographic surveys, daily reports on the symptoms of MS and other active functional tests. The smartphone has also passively collected data related to the number of steps, sleep and heart rate. The researchers then teamed up with Google Data Scientists to develop AI algorithms which permanently predict in the event of symptoms. The analyzes included data from 713 people who used the application for at least three months, most of which had a SEP.
Scientists have tested several different approaches to predict on a weekly basis if one person had one of the five SEP symptoms – fatiguesensory disturbances, walking instability, depression or anxietyAnd cramps or muscle spasms – with at least moderate gravity.
The team perfected a type of algorithm that had the best predictive capacities, with diagnostic accuracy between 80% and 90% for each symptom. This model was particularly precise given all the features evaluated collected on the smartphone, but which seemed most important to predict future symptoms was a past story of this symptom.
When this feature has been removed from the algorithm, its predictive capacities have decreased, although it can still achieve a fairly high performance by examining all other data.
“This highlights the importance of considering all the data available collectively, which proves the need for methods that can analyze a wide range of data simultaneously,” wrote the researchers, noting a method based on applications to predict the symptoms of MP which could give patients more certainty in their daily lives. “Symptomatic uncertainties associated with MS can have a substantial impact on quality of life, because individuals find it difficult to anticipate how they will feel every day, if their symptoms will hinder daily tasks or if the particular symptoms come from their MS.”
“By focusing on the exploitable prediction of high severity symptoms, the algorithm described here could improve anticipation advice and symptom management,” they said.
For example, if a person knows that their ability to walk could soon decrease, they can consider physiotherapy or other interventions before it happens. The application has been designed to provide patients with a summary ratio of symptoms that can facilitate such conversations.
Overall, “this approach has the potential to allow subjects as the experts in their own experience in order to improve the management of symptoms and to optimize interactions often limited with doctors and clinical experts,” the researchers wrote.
