Therapeutic implications
Dr. Ezzati believes that machine learning and data analysis could point the way to a future of more tailored migraine therapies. “I think we have to in general go down the path of using more evidence and more data to inform us about individualized planning for patients. For that we need larger clinical studies and larger epidemiological studies to help us identify more homogeneous subtypes of patients that we can eventually target in clinical trials,” he said.
Catherine Chong, MD, who chaired the session where the research was presented, praised the study in an interview. “Episodic migraine and chronic migraine have been developed [as categories] by headache frequency per month, and it was basically based on consensus in committee. They made basically a determination that 15 and under migraine days would be episodic migraine and over would be chronic migraine. So they dichotomized migraine, in a way, based on what people thought in the field. Looking at the data freely, and letting the algorithm determine the different subtypes, and putting everybody with migraine in it, and having these groups naturally appear from the data, I think is fascinating,” Dr. Chong said.
She echoed Dr. Ezzati’s call for further research that could create even more subgroups. “Is it really truly the case that somebody with less than 15 migraine days [per month], that 14 migraines days would be so different than somebody with 15 or over, or 8? I think we need to look at it further to see whether there are additional subgroups within that data. I think there are probably more [groups identifiable] from different data that we have out there,” said Dr. Chong.
Dr. Ezzati has consulted for or been a reviewer or advisory board member for Corium, Eisai, GlaxoSmithKline, Mint Research, and Health Care Horizon Scanning System. He has received research funding from Amgen. Dr. Chong has no relevant financial disclosures.