Artificial Intelligence in Medicine in the United States, China and India

Objective: To compare the development status of artificial intelligence (AI) in medicine among the United States (US), China and India with bibliometric analysis. Methods: Articles involving AI in medicine published from 2015 to 2019 were retrieved on March 30, 2020 from Web of Science Core Collection. The country-level and the institution-level performance of the US, China and India in the field of AI in medicine were compared with indicators including the amount of papers, 5-year Compound Annual Growth Rate (CAGR) of the amount of papers, the amount of highly-cited papers, the proportion of highly-cited papers and the average citations per paper. In addition, the research hotspots and international cooperation of the three countries in recent 5 years were compared by conducting keywords co-occurrence analysis and co-authorship analysis in VOSviewer. Results: From 2015 to 2019, The US has published 7838 papers and 154 highly-cited papers in the field of AI in medicine, with an average citations per paper to be 9.3, and the proportion of highly-cited papers to be 2.0 %. China has output 6635 papers and 73 highly-cited papers in this field, with an average citations per paper to be 5.3, and the proportion of highly-cited papers to be 1.1%. India has output 3895 papers and 22 highly-cited papers in this field, with an average citations per paper to be 3.6, and the proportion of highly-cited papers to be 0.6%. The 5-year CAGR of the US, China and India in the period of 2015~2019 were 16.0%, 25.4% and 2.4%, respectively. At the institutional level, most of these indicators were significantly better for the US institutions than for Chinese and Indian ones. There were four research hotspots in this field, namely medical imaging technology, health big data mining, disease prediction with biomarkers and genetic information, and early diagnosis of neurological disease. The three countries focused on different hotspots, with China focusing relatively less on health big data mining, while the US and India being complementary to each other. As to international cooperation, the average links per paper to other countries were 0.60, 0.40 and 0.20, respectively, for the US, China and India. Conclusions: In the field of AI in medicine, the US, with a number of competitive institutions in AI and medical researches, is taking a definitely leading role, having conducted many innovative researches and cooperated extensively with other countries. China is taking the second leading role at the country level, with top institutions somewhat less productive than those in the US. India is the third productive country, with top institutions obvious less productive than those in the US, and with research hotspots exactly complementary to the US.