Personalized information service has become an important interest in medical library due to the rapid development of Internet technology. The SOA framework-based personalized information service system architecture was proposed by analyzing the contents and characteristics of personalized information service in medical library, which is characterized by standards, loose couple, low cost modification and operating maintenance, and can thus ensure both personalized information service and daily professional activities in medical library.

Based on the new trend of transition from information service to knowledge service currently emerging in the field of information service,the paper presents the construction of a knowledge service system based on clinical data center.It introduces the architecture of the system,analyzes its features,describes its application effect and discusses the issues to be concerned for the further development of the knowledge service system in the future.

Chronic diseases have grown into a major public health threat affecting social development. The traditional management of chronic diseases tends to characterize " prioritizing treatment and neglecting prevention" . In this study, big data application, knowledge management, knowledge service and other theories and methods in the information management discipline were put in practice of chronic disease management. Thanks to the construction of a multi-source and heterogeneous chronic disease database, a chronic disease health knowledge analysis platform and a chronic disease service application platform, the following objectives were achieved. These included the data interconnections between patients′ self-testing data and medical institutions′ data in chronic disease data management, the integration and accurate delivery of chronic disease health knowledge, and personalized and accurate services covering the whole process of chronic disease management, namely prevention, diagnosis, treatment and rehabilitation.Since August 2019, this management mode has been used to build a diabetes big data platform. By January 2022, 380 million pieces of diagnosis, treatment and management data were collected from about 2.16 million residents, with the accuracy rate of 97.46% and the integrity rate of 96.07%. Corresponding knowledge base and service platform were built, and personalized service was provided to diabetes patients. These measures improvd the awareness rate, treatment rate, treatment control rate, science popularization and communication rate covering both diabetes patients and community residents.