Twenty-two finisher farms were randomly assigned to an intervention or a comparison group. The intervention group implemented a package of hygiene and biosecurity measures to reduce Salmonella infection, measured by culture of pooled pen faecal samples and use of the meat juice enzyme-linked immunosorbent assay (MJ-ELISA). Data on hygiene and biosecurity practices were used to calculate compliance scores, which were significantly greater on intervention farms. Salmonella was isolated from 38% (95% confidence interval [ci] 22% - 53%) of pens on intervention farms and 42% (95% ci 27% - 58%) of pens on comparison farms. The prevalence of MJ-ELISA positive pigs on intervention farms was 40% (95% ci 26% - 58%) and 58% (95% ci 41% - 75%) on comparison farms. These differences were not statistically significant. The power of this study was reduced by a strong farm effect. The prevalence of infection amongst introduced pigs at the start of the finisher cycle had a significant impact upon overall pen prevalence.