Even with good surveillance programmes, hospital-acquired infections (HAIs) are not always recognized and this may lead to an outbreak. In order to reduce this risk, we propose a model for prompt detection of HAIs, based on the use of a real-time epidemiological information system called VIGI@ct (bioMèrieux, Las Balmas, France) and on the rapid confirmation or exclusion of the genetic relationship among pathogens using fluorescent amplified length fragment polymorphism (f-AFLP) microbial fingerprinting. We present the results of one year's experience with the system, which identified a total of 306 suspicious HAIs. Of these, 281 (92%) were 'confirmed' by clinical evidence, 16 (5%) were considered to be simple colonization and the latter nine (3%) were archived as 'not answered' because of the absence of the physician's cooperation. There were seven suspected outbreaks; of these, f-AFLP analysis confirmed the clonal relationship among the isolates in four cases: outbreak 1 (four isolates of Pseudomonas aeruginosa), outbreak 2 (three Escherichia coli isolates), outbreak 6 (two Candida parapsilosis isolates) and outbreak 7 (30 ESbetaL-producing Klebsiella pneumoniae subsp. pneumoniae). Based on our results, we conclude that the combination of VIGI@ct and f-AFLP is useful in the rapid assessment of an outbreak due to Gram-positive or Gram-negative bacteria and yeasts.