Objective: identify risk factors of nosocomial infection and develop a risk of nosocomial infection scale in pediatric intensive care unit Risk factors were analysed with logistic regression, value of nsk of nosocomial infection. Study design: A prospective cohort study was carried out in Children's Hospital No.1 within 14 months. Main variables were nosocomial infection, demographic, nutrition status, underlying disease, PRISM, intenvention. Risk factors were analysed with the logistic regression, the risk of nosocomial infecbon scale was identified cut-off value with ROC. Results: There were 671 patients enrolled in the study. The Incidence of nosocomial infecdon was 23%. The risk factors of nosocomial infection were: age <
1 year (RR = 1,9 (1,1 -3,4))
malnutrition II, III (RR = 2,0 (1,1 -4,2))
PRISM >
10 (RR = 5,4 (2,4 - 11,9))
intubation (RR = 2,3 (1,3 -4,0)
parenteral nutntion (RR = 2,8 (1,3-5,9)
central venous catheter (RR = 2,8 (1,3 - 5,9)
bladder catheter (RR = 5,7 (3,0 - 10,7)). Risk of nosoajmial tnfecdon scale = 6 x II x) + (27 x PRISM >
10 x) + (10 x CVPx) + (8 X intubation x) + (17 x bladder catheter %) + (10x parenteral nutrition x) (z = 1 if nsk factor appears
X=0 if not), the risk of nosocomial infection is high with sensitivity of 86,4% and specificity of 81,2% if scale IS more than 32. Conclusion: The risk of nosocomial infection scale includes 7 factors: <
1 year old
malnutrition >
II
PRISM>
10
CVP
intubabon
bladder catheter
parenteral nutrition. The risk of nosocomial infection is high with value of scale more than 32. The risk of nosocomial infecbon scale needs to be enhanced and applied to help the control of nosocomial infection in the pediatric intensive care units.