Smartphone app collects data on migraine triggers, medications from 600,000 users
Click Here to Manage Email Alerts
Self-reported migraine treatment records collected via a smartphone application provided insights into migraine triggers and medication effectiveness.
Researchers presented results of a big data analysis at the American Headache Society Virtual Annual Scientific Meeting.
“[The Migraine Buddy] is a popular headache and migraine tracking application that can be downloaded and installed to your mobile phone,” Mindy Fang, PhD, junior associate professor of biostatistics at the Health Innovation School in Japan, said during a presentation. “In this study, we have included more than 10 million records from English speakers living in over 50 countries and regions. Each migraine record contains information about the start and the end time of the migraine, pain intensity, triggers and symptoms, medications the patient has taken and the outcomes of the medication.”
Between 2014 and 2020, Fang and colleagues constructed more than 10 million migraine type records. They excluded records missing “essential information” or “ambiguous medication outcomes” and used the remaining approximately 8.7 million records from approximately 600,000 unique users in the data analysis. Men accounted for 44,000 of the total users and women for 393,000, with 151,000 choosing not to specify their sex. The average age of patients was in the mid-30s, Fang said.
Migraine Buddy incorporated a pain scale from zero to 10. Users had an average migraine pain intensity of between five and six, with the average pain lasting approximately 30 hours and the average user experiencing more than 10 migraine attacks.
Fang and colleagues applied a two-level nested logistic regression model to compare the effectiveness of users’ migraine medications. They evaluated 22 of the most commonly used drug molecules and another group of medications they termed “others,” and considered outcomes positive if users reported that the medication was “helpful” or “somewhat helpful.” In all, the analysis included 23 medication groups, Fang said.
Triptans had the best ORs compared with the other groups. Results appeared “relatively consistent” across the United States, the U.K. and Canada. Fang noted that all of the receiver operator characteristic curves for the logistic regression models had areas under the curve of approximately 0.85, which suggested the models had “relatively good fit.”
Results of network analysis of the triggers and symptoms of the migraines showed common migraine triggers for women included neck pain, stress, anxiety, lack of sleep and dehydration. The most common triggers for men included anxiety, stress, lack of sleep and neck pain. Male patients were more likely than female patients to report having no idea of what triggered their migraines.
“More female patients reported nausea, neck pain, sensitivity to light and noise, fatigue, throbbing pain and nasal congestion,” Fang said. “Male symptoms are also similar.”