Kong Zihe, Kong Dexing, Kong Jiangming, Xing Yuxin, Liang Ping
Department of Government, School of Public Affairs, Zhejiang University, Hangzhou, China.
School of Mathematical Sciences, Zhejiang University, Hangzhou, China.
BMC Health Serv Res. 2025 Sep 2;25(1):1179. doi: 10.1186/s12913-025-13344-x.
Artificial intelligence (AI) has been regarded as a major success in healthcare services. At present, few studies have successfully empirically analyzed this view through large-scale multi-center trails, especially in China as well as other less-developed regions. The research aim of this work is to empirically reveal how artificial intelligence benefits in less-developed regions and reallocates medical resources.
This work takes the "non-perception-perception" public service performance evaluation model as the theoretical framework for evaluation, and the "task-periphery" performance structure model as the basis for forming performance indicators. This work has also adopted literature research, expert consultation, questionnaire measurement, and statistical inference as methodology, as well as a representative, advanced large-scale multi-center medical policy pilot case for performance evaluation. The case is conducted in the entire region of Puyang Prefecture, Henan Province, China. As of June 2024, 108 public healthcare institutions have been equipped with 291 modules and screened 281,663 people. 88.34 million RMB has been invested. A total of 493 questionnaires were collected (429 valid questionnaires).
Based on the non-perceptual mode, the AI system has technical advantages, including more accurate diagnostic results (20.72% higher than the conventional rate), more detailed diagnostic data (precise to two decimal places), faster reporting (down to 0.2 seconds), standardized data collection procedures, unified healthcare collaboration platforms and lower healthcare insurance (reduced 85.7%-92.9%). Based on the perceptual mode, the overall performance value is relatively high (5.19/7 on average). The public value created by the system application is more distinct than the direct economic value.
This advanced, representative, large-scale multi-center pilot case reveals that AI has effectively promoted data standardization, regional medical cooperation, and reduced medical insurance expenditures mainly by improving the accuracy, precision, and speed of diagnosis, enabling less-developed regions to access more efficient and fair medical resources. The application of the AI system not only creates very significant economic value for primary-level medical and health institutions, but also generates huge public value (sustainable development, social satisfaction, etc.). This work points out a referential path for the healthcare development in the Global South from the perspective of AI.
The online version contains supplementary material available at 10.1186/s12913-025-13344-x.
人工智能（AI）已被视为医疗服务领域的一项重大成就。目前，很少有研究通过大规模多中心试验成功地对这一观点进行实证分析，尤其是在中国以及其他欠发达地区。这项工作的研究目的是实证揭示人工智能在欠发达地区的益处以及如何重新分配医疗资源。
这项工作以“非感知 - 感知”公共服务绩效评估模型作为评估的理论框架，以“任务 - 外围”绩效结构模型作为形成绩效指标的基础。这项工作还采用了文献研究、专家咨询、问卷调查和统计推断等方法，以及一个具有代表性、先进性的大规模多中心医疗政策试点案例进行绩效评估。该案例在中国河南省濮阳市全域开展。截至2024年6月，108家公立医疗机构配备了291个模块，筛查了281,663人，投入资金8834万元。共收集问卷493份（有效问卷429份）。
基于非感知模式，人工智能系统具有技术优势，包括诊断结果更准确（比传统准确率高20.72%）、诊断数据更详细（精确到小数点后两位）、报告速度更快（低至0.2秒）、数据收集程序标准化、统一的医疗协作平台以及更低的医疗保险费用（降低85.7% - 92.9%）。基于感知模式，整体绩效值相对较高（平均5.19/7）。系统应用所创造的公共价值比直接经济价值更为显著。
这个先进的、具有代表性的大规模多中心试点案例表明，人工智能主要通过提高诊断的准确性、精确性和速度，有效地促进了数据标准化、区域医疗合作，并降低了医疗保险支出，使欠发达地区能够获得更高效、公平的医疗资源。人工智能系统的应用不仅为基层医疗卫生机构创造了非常显著的经济价值，还产生了巨大的公共价值（可持续发展、社会满意度等）。这项工作从人工智能的角度为全球南方的医疗发展指出了一条可供参考的路径。
在线版本包含可在10.1186/s12913 - 025 - 13344 - x获取的补充材料。
