Objective To evaluate the targeted surveillance of health care-associated infections in ICU. Methods PubMed, EMBASE ( Excerpt Medica Database ) , China Biology Medicine disc, China National Knowledge Infrastructure, Vip Journal Integration Platform and Wanfang Data were searched from 2009 to 2013 by two researchers. Full-text studies on targeted surveillance of healthcare-associated infections in ICUs were included. Studies were evacuated using STROBE and GRADE. Random effects meta-analyses were used to generate summary estimates of infection rates. Results Totals of 21 studies met eligibility criteria, which were all Chinese research papers. Taking the infection rate as an example, 74 600 cases were observed within which 7 318 cases were infected. Pooled meta-analyses of eligible studies revealed that adjusted infection rate:infection rate, case infection rate, daily infection rate, case daily infection rate were 3. 27%, 4. 92%, 4. 58‰, 6. 49‰, respectively. The infection rates were 23. 12‰ in ventilator associated pneumonia, 5. 26‰ in catheter-related bloodstream infections, and 2. 38‰ in catheter-associated urinary tract infections. The lower respiratory tract, the blood, and the urinary tract were the main infection sites, accounting for 58. 07%, 18. 16%, and 15. 35%, respectively. A. baumanii, P. Aeruginosa, fungi, S. aureus, K. pneumoniae, E. coli were the main pathogens, accounting for 20. 54%, 14. 28%, 14. 35%, 11. 62%, 10. 92% and 6. 86%, respectively. Conclusions The burden of health-care-associated infection in ICU is high. Our findings indicate a need to improve surveillance and infection-control practices.

Objective To discuss the effects of community diabetes specialist nurse training practice from the perspective of knowledge management, in order to provide some references for personnel training mode of the community nurses. Methods We established community diabetes specialist nurse training practices, screened 80 community nurses to participate in training, and implemented the training practices. The training objects were surveyed with the diabetes knowledge and performed duties questionnaires before and after the training. Diabetes knowledge level and the degree of willingness to perform duties were as evaluation indexes, SPSS 18. 0 software was used for analyzing and processing the date, and studying training effects from the perspective of knowledge management. Results Diabetes knowledge test accuracy of community nurses were 37. 0%, 68. 4% before and after the training (Z=7. 36, P<0. 01);willingness to perform their duties showed that options of very willing were 52. 5%(42/80), and options of willing were 46. 2%(37/80), and only one person said reluctantly willing. Conclusions The training practice can guarantee quality of nursing knowledge resources which were imported community health service centers, promote effectively spread, flow and use of knowledge resources in the community health service centers, then to promote knowledge utilization value.