BackgroundLiver fluke (Fasciola hepatica) is a widespread parasite of ruminants which can have significant economic impact on cattle production. Fluke infection status at the animal-level is captured during meat inspection of all animals processed for human consumption within Northern Ireland. These national datasets have not been analysed to assess their utility in uncovering patterns in fluke infection at animal- and herd-levels in Northern Ireland.MethodsWe utilised a dataset of 1.2 million animal records from ~18,000 herds across 3 years (2011–2013) to assess animal- and herd-level apparent prevalence and risk-factors associated with fluke infection. Animal-level apparent prevalence was measured as the proportion of animals exhibiting evidence of fluke infection at slaughter; between herd-level infection prevalence was measured by binary categorisation of herds (infected or not). “Within herd” infection prevalence was measured using the proportion of animals within a herd that showed evidence of fluke infection per year (ranging from 0–100 %). “Within herd” infection prevalence at the herd level was investigated using multivariable modelling.ResultsAt the animal level, the proportion of animals slaughtered that exhibited evidence of infection was 21–25 % amongst years. Across herds, the proportion of herds with at least one infected animal, varied between 61 and 65 %. However, there was a significant sampling effect at the herd-level; all herds where at least 105 animals slaughtered over the study period exhibited evidence of fluke infection (100 %). There was significant variation in terms of within-herd infection prevalence. Risk factors included herd type, long-term weather variation, geographic location (region) and the abattoir.ConclusionsLiver fluke apparent prevalence was high at the herd-level across years. However, there was lower prevalence at the animal level, which may indicate significant variation in the exposure to fluke infection within herds. The proportion infected within-herds varied significantly in time and space, and by abattoir, herd-type and some weather variables. These data are a useful source of information on a widespread endemic disease, despite known limitations in terms of test performance (low sensitivity). As well as informing on the distribution and severity of liver fluke infection, these analyses will be used to investigate the effect of co-infection on risk for bovine tuberculosis.Electronic supplementary materialThe online version of this article (doi:10.1186/s13071-016-1489-2) contains supplementary material, which is available to authorized users.