A multivariable predictive model including self-prediction and common migraine trigger factors was only slightly better than random for predicting episodic migraine attacks, suggesting that more work is needed to identify additional predictive values that can improve management of episodic migraine attacks, according to study results published in Headache.
The goal of this study was to develop a multivariable prediction model for patients with new onset migraine headaches. This model was largely based on self-prediction by the patients and frequent triggers.
A total of 178 adults (mean age, 37 years; 166 women) with episodic migraine were recruited through online forums for this 90-day prospective daily-diary cohort study. Participants completed a survey every evening pertaining to migraine occurrence and possible predictors. They answered questions on an iPhone-specific application.
These predictors, including stress, sleep, caffeine and alcohol consumption, weather, menstruation, and self-prediction, were imputed into multivariable multilevel logistic regression models for the risk of a new-onset migraine day compared with a healthy day. Study researchers then internally validated these models through repeated cross-validation.
In the primary analysis, participants reported 1870 individual migraine events on 10,696 at-risk days. Positive associations were found between next-day migraine risk and a decrease in caffeine consumption, increased self-predicted probability of headache, a greater level of stress, and being within 2 days of menstruation onset. The multivariable model that included these factors anticipated migraine risk just slightly better than chance (within-person C-statistic, 0.56; 95% CI, 0.54-0.58).
Limitations of this study were the inclusion of only participants with iPhones and the potential measurement error in the mobile-app-based surveys due to the brief nature of the questions.
The study researchers concluded that the incorporation of additional predictors in the model as well as the studying of a more “homogeneous migraine population may enable future predictive models to perform better, enable targeted medication use, and reduce the unpredictability of episodic migraine attacks.”
Holsteen KK, Hittle M, Barad M, Nelson LM. Development and internal validation of a multivariable prediction model for individual episodic migraine attacks based on daily trigger exposures. Headache. Published online October 6, 2020. doi:10.1111/head.13960