Danish Veterinary Institute, Department for Poultry, Fish, and Fur-Bearing Animals, Hangøvej 2, 8200, Aarhus N, Denmark

The same person visited all 78 broiler houses, and farm-related factors were listed both from personal observations and by interrogating the person in charge of attending to the broilers. Additional factors (noted at the visits) were recorded in a later telephone interview. Altogether, 187 variables were analyzed.

Broiler houses were divided into single-infection houses (which had one or both of the salmonella types in only one crop), and multiple-infection houses (where infection occurred in multiple crops). This was done both generally and separately for each of the salmonella types.

Factors associated with house status—but that were beyond the control of the farmer—were the salmonella type and the percentage of salmonella positive samples in the first crop. Factors which could be controlled by the farmer and which were associated with reduced prevalence of multiple-infection houses included antiseptic soap and water for washing hands in the anteroom, hygiene barriers when removing dead broilers, gravel alongside the broiler house, systematic checks of indoor rodent-bait depots, and combined surface and pulse-fogging disinfection.