Machine Learning-Driven Phenomics
Are there clinical characteristics that are important predictors of response to migraine prophylactic treatment?
Thus far, no one has looked carefully. New machine-learning techniques will allow us to investigate this possibility. In the Migraine Research Program's headache practice at the Mayo Clinic, the team is carefully and quantitatively documenting treatment response patterns of people with migraines. The group has a large database of cross-sectional information gathered as the patients begin treatment.
The program's researchers propose to use rapidly emerging artificial intelligence technologies to interrogate not only the structured data in his database, but also to study the vast amount of unstructured data in anonymized patient health records to identify factors that correlate with therapeutic response to preventive migraine treatments. This information will be used in conjunction with genomic data to develop methods that no longer rely on chance, but rather focus on using an individual's characteristics to improve the likelihood and rapidity of treatment success.