Answers are grounded in this paper and cite their sources.
Background: Previous forecasts of migraine burden have largely relied on traditional time-series models that assume stable trends and inadequately capture nonlinear dynamics. Using Global Burden of Disease (GBD) 2023 data, we applied a machine-learning approach incorporating demographic, economic, and urbanisation indicators to generate refined projections through 2050. Methods: We developed a hybrid forecasting framework integrating autoregressive integrated moving average with extreme gradient boosting, with calendar year, projected population, gross domestic product per capita, and urbanisation rate as covariates. We obtained migraine incidence, prevalence, and disability-adjusted life years (DALYs) from GBD 2023 and stratified them by World Bank income group, country, age, and sex. We evaluated predictive performance using a time-series split and rolling-origin cross-validation, quantifying forecast uncertainty with conformal prediction intervals. Results: By 2050, global migraine age-standardised rates (ASRs) are projected to remain largely stable. The ASRs are projected to decline by 0.92% for incidence (ASIR) and 1.23% for prevalence (ASPR), while the ASR for DALYs (ASDR) is projected to remain essentially unchanged (-0.25%). Substantial heterogeneity is expected across income groups, countries, sex, and age. High-income countries are projected to remain the highest-burden settings (ASPR = 14,963.99; ASDR = 523.33 per 100,000), although ASIR (-6.04%), ASPR (-1.61%), and ASDR (-1.70%) are all expected to decline. Females are projected to bear nearly twice the burden of males, with higher 2050 ASIRs (1,419.14 vs. 894.55), ASPRs (17,385.71 vs. 9678.20), and ASDRs (673.40 vs. 309.06). Prevalence is projected to peak in the 30s, with disability remaining elevated into the 40s and 50s. Conclusions: Global migraine ASRs are projected to remain stable, yet marked differences persist across income groups, countries, and sex-age subgroups. These are model-based projections conditional on the assumed demographic and economic trajectories rather than predictions of observed future burden.
No takes yet. Share an insight, caveat, or question.
Wang et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: