Abstract: Objective  To explore the significance and value of refined management of SCI scientific papers in hospitals in improving the quantity and quality of scientific papers. Methods  Based on the issue time and characteristics of the management measures for hospital scientific papers, the period 2011-2020 is divided into four stages: 2011-2014 “rough”, 2015-2016 “simple”, 2017-2018 “performance” and 2019-2020 “refined”. Using the Web of Science (WOS) core database  as the data source, we retrieved SCI scientific papers published from 2011 to 2020 with Beijing You’an Hospital，Capital Medical University and Beijing Institute of Hepatology  as the first published address. The number, impact, and thematic evolution of scientific papers were visualized by bibliometrics software. Results   A total of 618 articles and review scientific papers were included in WOS from 2011 to 2020. The average number of scientific papers published in 2011-2014 was 32, in 2015-2016 was 53, in 2017-2018 was 71, and 2019-2020 was 112. From 2011 to 2020, there has 13 scientific papers with Normalized Total Citations (NTC) >5, including 4 in 2011-2014, 5 in 2017-2018, and 4 in 2019-2020. The thematic evolution analysis shows hepatitis B virus, liver failure, cirrhosis, hepatocellular carcinoma, and AIDS are the major subjects through 2011 to 2020. The research was mainly focused on clinical in 2011-2014 “rough” phase and 2015-2016 “simple” phase. The clinical and mechanism research were integrated in 2017-2018 “performance” phase and 2019-2020 “refined” phase, and the research area was expanded to multiple disciplines. Conclusion  The refined management of scientific papers has promoted the improvement of the number and influence of scientific papers in hospitals, enhanced the scientific research ability and innovation level of hospitals, and promoted the development of dominant subjects and cross integration of multiple subjects in hospitals.

Key words: refined management, scientific research management, SCI papers, bibliometrics, visualization analysis