Interactive protocol for acquisition of migraine diaries with a mobile app and machine learning data analysis

• Tatiane Vieira Alves USP
• Kamila Rios da Hora Rodrigues USP
• Moacir Antonelli Ponti USP

Resumo

Migraine is a phenomenon that affects thousands of people on a daily basis around the world. It is not surprising that many studies have been carried out to better understand migraine using statistical and machine learning techniques. Since such analysis needs both good-quality and sufficient data, it becomes important to investigate tools that allow to improve the data collection protocol. A data acquisition tailored for such a problem would allow us to better understand the most determinant triggers that precede a migratory crisis, especially because those vary significantly for each individual. In this work we designed and evaluated a protocol for data collection comprising an initial questionnaire and a diary of habits and events. Using a decision tree model, we found the main domains of migraine triggers in the studied group are stress, caffeine intake, sleep and effects of the environment. More importantly, we report the most important aspects and challenges when designing such a protocol so that to better engage users, as well as users' testimonials when using the tool.

Palavras-chave: Migraine, Collection Protocol, Online Diaries, Machine Learning, Reports, Feedback