Chinese researchers inject AI power to evidence-based medicine

LANZHOU -- Chinese researchers are integrating the growing power of AI into the Digital Intelligent Evidence-Based Medicine (i-EBM), creating a more dynamic, precise and efficient approach to medical evidence.

"As a new modern medical paradigm, i-EBM features three dimensions -- multi-source data integration, intelligent evidence analysis, and individualized decision support," said Ge Long, a professor at the School of Public Health of Lanzhou University and leader of the study.

"The burgeoning AI technologies are more active in processing massive data and transforming them into reference treatment plans, enabling the traditional evidence-based medicine (EBM) to advance in the AI era," Ge added.

Published in the journal Chinese Science Bulletin, the study is led by Ge's colleagues in collaboration with evidence-based medicine scholars.

EBM, a cornerstone of modern medical decision-making, combines the best research evidence, clinical judgment and patients' concerns. For decades, it has continued to significantly enhance the scientific understanding behind medical decisions.

"However, the limitations of EBM, such as the time lag, isolated evidence, and insufficient individualized consideration to each patient, have caused troubles in today's complex medical environment," Ge said.

"In the AI era, i-EBM is not a disruptive force to the traditional EBM model, but rather an inevitable evolution, which is targeted at an in-depth collaboration between machine intelligence and human expertise," Ge added.

I-EBM builds a unified digital intelligence foundation by integrating high-quality scientific research evidence, biomedical knowledge systems, and multimodal real-world data including electronic medical records and medical images, among others. On this basis, it combines with AI technologies, individualized diagnosis and treatment suggestions to assist doctors in making precise decisions.

"Of course, in the initial stage, what i-EBM can offer is the standardized plans, which can serve as a reference for doctors," Ge said.

"Theoretically speaking, EBM considers information of various aspects. However, in practice, it is difficult for clinicians to search, sequence, and analyze such a large amount of information," Ge said.

I-EBM is pushing the limits of traditional EBM. By means of AI technologies such as knowledge graphs, it can build deep cross-domain connections across scientific research data, clinical medical records, medical images, and environmental meteorology information -- tasks that were previously carried out manually with a great degree of difficulty.

Moreover, AI can greatly reduce repetitive work. In just a few hours, or in some cases even minutes, they can complete the literature screening and evidence update that used to take months to finish.

According to Ge, the new system can assist with data integration, pattern recognition and causal analysis in complex research of the traditional Chinese medicine (TCM) to help reveal the internal laws of TCM evidence and optimize its evaluation system.

By now, the study team has applied their study achievements by developing multiple digital and intelligent products in fields such as the medication guide of Chinese patent medicine, facilitating medical research and practice.

The team has also cooperated with medical institutions to carry out research and application of i-EBM in childhood pneumonia treatment. I-EBM integrates imaging, laboratory and clinical data to build a multimodal database and facilitate more scientific diagnosis and treatment plans.

"The era of AI is coming. We will further our i-EBM research to support the improvement of medical standards, promote equal access to medical care, so as to play a role in improving people's health and well-being," Ge said.