PRIMARY-AI: Outcomes-based standards to safeguard primary care in the AI era

Nature Medicine

Vol. 32, Issue 3

Date of Publication: February 11 2026

This article introduces the PRIMARY-AI framework, an outcomes-based approach for evaluating artificial intelligence in primary care. It argues that rapid deployment of AI tools risks undermining core primary care functions without proper standards. The framework proposes assessing AI based on its impact on continuity, coordination, comprehensiveness, and person-centered care to ensure technologies strengthen health systems rather than fragment them.

Citations

Zeng, Dian, Lorainne Tudor Car, Kamlesh Khunti, Yuanli Liu, Till Bärnighausen, Niels H. Chavannes, Pearse A. Keane, Holger Kunz, Lan Xue, Joseph J. Y. Sung, Yih Chung Tham, Lorenzo Righetto, Rupa Sarker, Samuel Yeung Shan Wong, Donald Boudreau, Qionghai Dai, Weiping Jia, Yang Liu, Dinggang Shen, Jia Liu, Weixing Shen, John S. Ji, Zhong Wang, Zhiyi Wang, Haibo Wang, Shenglan Tang, Chenyang Pei, Zehua Jiang, Zihao Zou, Yiming Qin, Huating Li, Yasha Wang, Dinesh Visva Gunasekeran, Sabrina Wong, Dong Xu, Ryan Urbanowicz, Liliana Laranjo, Ana Luisa Neves, Nan Liu, Yulan He, Phuoc Van Le, Neil Bressler, Rifat Atun, David C. Klonoff, Bin Sheng, Nigam Shah, Josip Car, and Tien Yin Wong. 2026. PRIMARY-AI: outcomes-based standards to safeguard primary care in the AI era. Nature Medicine 32 (March): 778–781. https://doi.org/10.1038/s41591-025-04178-5