This study aimed to correlate nosocomial bloodstream infections (NBIs) across time against the ecological effect of infection control activities. All patients hospitalised >or=48h in the haematology and intensive care departments of a university hospital and discharged between 1 January 2004 and 30 June 2006 were prospectively included. The case definition of NBI infection was: (1) at least one positive blood culture justified by clinical signs, or (2) at least two positive blood cultures when the micro-organism was one of the following: coagulase-negative staphylococci, Bacillus spp. (except Bacillus anthracis), Corynebacterium spp., Propionibacterium spp., Micrococcus spp., or other commensal with similar pathogenicity, if occurring >or=48h after patient admission. NBI incidences were correlated in quarterly intervals using Spearman's test and linear regression. In total, 3829 patients accounting for 46 474 patient-days at risk were included. We identified 101 NBIs in intensive care and 286 NBIs in haematology. There was a correlation between NBI incidence in haematology with the NBI incidence in intensive care (r=0.68, P=0.042). The linear model for NBI incidences between departments was R(2)=0.52, with a positive trend (P=0.029). A common determinant such as improved hygiene measures is the most likely reason for this association.