An AI-powered study shows a surge in global rheumatoid arthritis since 1980, revealing local hotspots, socioeconomic disparities, and worsening inequalities in disease burden. This image depicts the geographic distribution of age-standardized incidence rates in 2021. Credit: Annals of the Rheumatic Diseases / Jin et al.

The most comprehensive analysis of rheumatoid arthritis data to date reveals that demographic changes and uneven health infrastructure have exacerbated the rheumatoid arthritis burden since 1980 and shows global disparities on a granular level.

The AI-powered study in the Annals of the Rheumatic Diseases, utilized deep learning techniques and policy simulations to uncover actionable insights for localized interventions that national-level studies have previously missed. Its design yielded highly precise, dynamic projections of further disease burden to 2040.

Principal investigator Queran Lin, MPH, WHO Collaborating Center for Public Health Education and Training, Faculty of Medicine, Imperial College London; and Clinical Research Design Division, Clinical Research Center, Sun Yat-Sen Memorial Hospital, Guangzhou, explains, "While previous Global Burden of Disease (GBD) studies have provided important insights, they have largely focused on high-level descriptions and visualizations at global and national scales, failing to capture local disparities or the dynamic interactions between socioeconomic development and disease trends.

"With access to sufficient computational resources and advanced analytical capabilities, our Global-to-Local Burden of Disease Collaboration aims to unlock the full potential of the GBD dataset (pioneered by the Institute for Health Metrics and Evaluation, University of Washington).

"By employing cutting-edge approaches such as transformer-based deep learning models, we were able to generate the most granular disease burden estimates to date, offering a new framework for guiding precision public health across diverse populations."

Using GBD data, the study integrates the largest spatiotemporal rheumatoid arthritis dataset spanning 953 global to local locations from 1980 to 2021 with a novel deep learning framework to reveal how demographic aging, population growth, and uneven health care infrastructure exacerbate rheumatoid arthritis burdens differently across regions.

It also enabled investigators to analyze the prevalence, incidence, mortality, disability-adjusted life years (DALYs), years of life lost (YLLs), and years lived with disability (YLDs) of rheumatoid arthritis, as well as their socioeconomic inequalities and achievable disease control based on socioeconomic development level (frontiers) and forecast long-term burdens until 2040 with scenario simulations.

The study observed that globally there were significant absolute and relative sociodemographic index (SDI)-related inequalities, with a disproportionately higher burden shouldered by countries with high and high-middle SDI. Among the key findings of the study are:

Co-lead author Baozhen Huang, Ph.D., Department of Biomedical Sciences, City University of Hong Kong, says, "Japan's sustained decline in DALYs despite a high SDI proves that socioeconomic status alone doesn't dictate outcomes; proactive health care policies such as early diagnosis programs can reverse trends."

Many regions around the world still lack the necessary evidence base to inform precision health policy and targeted interventions. These data are intended to support more informed clinical decisions and health policy planning, especially in settings where reliable subnational evidence has historically been scarce.

Co-lead author Wenyi Jin, MD, Ph.D., Department of Orthopedics, Renmin Hospital of Wuhan University; and Department of Biomedical Sciences, City University of Hong Kong, concludes, "The adoption of this advanced framework quantifies the expected impact of feasible intervention scenarios in public health, supplying policymakers at global, national, and local levels with more reliable, dynamic evidence, redefining the very paradigm of health surveillance."

More information: Spatiotemporal distributions and regional disparities of rheumatoid arthritis in 953 global to local locations, 1980-2040, with deep learning-empowered forecasts and evaluation of interventional policies' benefits, Annals of the Rheumatic Diseases (2025). DOI: 10.1016/j.ard.2025.04.009  Journal information: Annals of the Rheumatic Diseases