Sea lice, Caligus rogercresseyi, are ectoparasitic copepods, which severely affect the salmon farming industry in southern Chile, reducing the health status of fish and producing both direct and indirect economic losses. Local farmers have reported increasing infestation levels since 2004, reaching a peak in 2007. In response to this situation, the Chilean Fisheries Service (Sernapesca) developed a surveillance programme; the first step of which consisted of a general survey of salmon farms. This survey included documenting counts of parasite burdens on fish and measurements of several husbandry and environmental factors providing an evaluation of risk factors for the observed infestation levels. The information collected was analysed using a linear mixed model technique, which takes into account the clustered structure of data, decomposing the unexplained variation and assigning it to different aggregation levels of the productive system. Geographical zones, fish species, treatment against sea lice performed 1 month before sampling, stocking density, fish weight and water salinity were the variables significantly associated with sea lice burdens. In contrast, treatments performed 2-3 months before sampling, use of photoperiod in sea cages and water temperature, were not significant. There was significant unexplained variation at all aggregation levels, i.e. sub-zone, fish farm and cage level, with the fish farm level showing the greatest variation.