A three-tier AI solution for equitable glaucoma diagnosis across China's hierarchical healthcare system.

Journal: NPJ digital medicine
Published Date: Jul 3, 2025

Abstract

Artificial intelligence (AI) offers a solution to glaucoma care inequities driven by uneven resource distribution, but its real-world implementation remains limited. Here, we introduce Multi-Glau, an three-tier AI system tailored to China's hierarchical healthcare system to promote health equity in glaucoma care, even in settings with limited equipment. The system comprises three modules: (1) a screening module for primary hospitals that eliminates reliance on imaging; (2) a pre-diagnosis module for handling incomplete data in secondary hospitals, and (3) a definitive diagnosis module for the precise diagnosis of glaucoma severity in tertiary hospitals. Multi-Glau achieved high performance (AUC: 0.9254 for screening, 0.8650 for pre-diagnosis, and 0.9516 for definitive diagnosis), with its generalizability confirmed through multicenter validation. Multi-Glau outperformed state-of-the-art models, particularly in handling missing data and providing precise glaucoma severity diagnosis, while improving ophthalmologists' performance. These results demonstrate Multi-Glau's potential to bridge diagnostic gaps across hospital tiers and enhance equitable healthcare access.

Authors

  • Yi Zhou
    Eye Center of Xiangya Hospital, Central South University, Changsha, Hunan, China.
  • Haitao Nie
    School of Instrumentation and Opto-Electronic Engineering, Beihang University, Beijing, China.
  • Xinyu Gong
    School of Chemical, Materials, and Biomedical Engineering, College of Engineering, The University of Georgia, Athens, GA 30602, USA.
  • Minhui Dai
    Eye Center of Xiangya Hospital, Central South University, Changsha, Hunan, China.
  • Zhaohong Guo
    Department of Ophthalmology, Yiyang Central Hospital, Yiyang, Hunan, China.
  • Xiaoling Deng
    School College of Electronic Engineering and School College of Artificial Intelligence, South China Agricultural University, Guangzhou 510642, China.
  • Mengyang Li
    The Faculty of Hepatopancreatobiliary Surgery, The First Medical Center, Chinese People's Liberation Army General Hospital, Beijing, China.
  • Yong Liu
    Department of Critical care medicine, Shenzhen Hospital, Southern Medical University, Guangdong, Shenzhen, China.
  • Lingyu Sun
    Eye Center of Xiangya Hospital, Central South University, Changsha, Hunan, China.
  • Xiangyi Tang
    Eye Center of Xiangya Hospital, Central South University, Changsha, Hunan, China.
  • Ling Zhou
    Eye Center of Xiangya Hospital, Central South University, Changsha, Hunan, China.
  • Zhiyao Tang
    Xiangya School of Nursing, Central South University, Changsha, Hunan, China.
  • Ziqing Xia
    School of Automation, Central South University, Changsha, Hunan, China.
  • Lemeng Feng
    Eye Center of Xiangya Hospital, Central South University, Changsha, Hunan, China.
  • Wulong Zhang
    Eye Center of Xiangya Hospital, Central South University, Changsha, Hunan, China.
  • Qingqing Yi
    School of Business Administration, Faculty of Business Administration, Southwestern University of Finance and Economics, Chengdu 611130, China; Institute of Big Data, Southwestern University of Finance and Economics, Chengdu 611130, China. Electronic address: yiqingqing@smail.swufe.edu.cn.
  • Xiaobo Xia
    Eye Center of Xiangya Hospital, Central South University, Changsha, Hunan, China.
  • Bin Xie
    School of Automation, Central South University, Changsha, China. xiebin@csu.edu.cn.
  • Weitao Song
    School of Electrical and Electronic Engineering, Nanyang Technological University, 50 Nanyang Avenue, Singapore, 639798, Singapore.

Keywords

No keywords available for this article.

External Resources

Popular Topics
Recent Journals