Figure 1From: Improved hospital-level risk adjustment for surveillance of healthcare-associated bloodstream infections: a retrospective cohort study Longitudinal plots of overall bloodstream infection rates in a sample of public hospitals in Queensland, Australia. A dashed vertical line splits training and validation subsets of the data, collected from 2001 to 2007. Hospitals 5, 7, 8 and 10 were level I hospitals and the remaining hospitals were level II hospitals.Back to article page