政府新闻
本地医生开发的医疗人工智能模型，与ChatGPT 4.0相当 2024-11-12
Shanghai doctors have developed a new generation of multimodal AI models, targeting Google's Med-Gemini family, for clinical diagnostics that can help with clinical practice, medical research, and education.
According to Shanghai East Hospital doctors, which developed Med-Go with the Institute of Software under the Chinese Academy of Sciences, the AI system is comparable to ChatGPT 4.0 in clinical testing and practice.
Med-Go has passed China's medical licensing exam and won several medical information competitions.
"We offer medical textbooks, materials, papers, and latest developments daily. Over 20 billion medical data are available. Its medical expertise is fresher, better, and more complete than a professor," according to emergency department physician Zhang Haitao, the system's developer.
Med-Go is in use in Shanghai East Hospital and 15 neighborhood health centers in the Pudong New Area. Some hospitals in Jiangsu and Shanxi provinces and health institutions have deployed the technology.
Dr Zhang Haitao (second right) talks with his colleagues while using Med-Go.
In clinical practice, doctors simply have to input the patient's response to previous treatments, lab results, medical, surgical, and social history, as well as the symptoms.
Within seconds, the system can provide a prompt, accurate, and precise diagnosis along with recommendations for follow-up treatment and processing. It prevents wrong diagnoses and significantly enhances the effectiveness and quality of healthcare practice.
Med-Go can give diagnosis and follow-up suggestions on treatment within seconds.
Zhang presented a case involving a 9-year-old boy who had been admitted to a prestigious Beijing hospital twice and who, after nearly a year of numerous group meetings with physicians from various departments and numerous examinations and tests, was given the right diagnosis.
"A doctor needs more than an hour to read all the medical records and papers. Only a few seconds after inputting the medical records, Med-Go provided an accurate diagnosis. Top doctors in Beijing agreed with the assessment of a highly uncommon autoimmune disease," Zhang said.
"The approach has improved clinical practice, particularly for young and entry-level physicians. It covers all departments because medical practitioners often concentrate on one. The system may consider 202 factors, including the patient's age, living circumstances, and symptoms while choosing a medication. A single physician cannot consider drug use from 202 perspectives."
Additionally, it provides quality control on medical record writing in line with diagnostic requirements to improve medical quality, as well as assistance with medical research and education, he said.
To further enhance medical efficiency and quality control, hospital officials announced that Med-Go will be incorporated into their smart hospital administration system.
"In the future, we will further perfect the function of Med-Go to make it a better tool for doctors to achieve a deep integration of AI and medicine," Xu Zhaohui, the vice president of Shanghai East Hospital, said.
Dr Liu Fangyuan of Lujiazui Neighborhood Health Center said that she has benefited from using Med-Go.
Source: Shanghai Daily
