Salmonellosis is the most significant foodborne bacterial infection in terms of its impact on morbidity and mortality in humans. Surveillance of salmonellosis cases by the National Reference Center for Salmonella (CNR) shows a 33% decline between 1997 and 2001. This reduction coincides with the implementation in 1998 of control measures in poultry farms targeting S. Enteritidis (SE) and S. Typhimurium (ST), the two serotypes most commonly associated with human infections. To assess the relationship between the observed decline in the number of human salmonellosis cases and the control measures implemented at the poultry farm level. This is an ecological study with an etiological focus based on the analysis of time series. Time series of the monthly number of SE and ST salmonellosis cases and a "control" series composed of non-avian serotypes were analyzed using data from the CNR from 1992 to 2003. Analysis of these three series using the Box & Jenkins method and the identification of intervention models made it possible to assess the impact of the control measures implemented in 1998. Unlike the "control" series, for which no effect of the intervention could be demonstrated, we found an annual decrease of 555 SE salmonellosis cases (95% CI=[148–964]) and 492 ST salmonellosis cases (95% CI=[0–1092]), representing respective decreases of 21% and 18% in the number of cases following the implementation of the intervention. For SE, the decrease in the number of cases followed immediately after the implementation of the measures, whereas for ST, it followed a decline in the number of cases that had begun in January 1998. Our study, which includes a "control" series, highlights a link between salmonella control measures in poultry farms and the decrease in the number of SE and ST salmonellosis cases. However, for ST, the link is less clear; the decrease observed prior to the intervention may also reflect measures implemented by the food industry in the dairy sector and the ground beef sector. (R.A.)