A March 23 news release said the study included 2,000 participants in the U.S. to gather symptom and environmental data, like pollen counts. When a participant uploads symptoms into their e-diary, the smartphone sensor captures physical activity data and geolocation.
With these inputs, the algorithm was able to predict the severity of allergies with greater than 80 percent accuracy, the release said.
“To date, machine learning models in healthcare have focused on clinicians with decision support or administrators with forecasting, rather than emphasizing patient-facing direct care models,” Nirav R. Shah, MD, CMO of Sharecare, said in the release. “When predictive models like those in our study are capable of real-time personalized predictors and offer a strategy for tailored clinical care and prevention, they will be integrated into care delivery.”
More articles on artificial intelligence:
FDA authorizes 1st machine learning COVID-19 screening tool
Machine learning tool can predict severe illness, death from COVID-19 in patients
New AI tool being tested in UF Health study to improve Parkison’s diagnosis
