The recent focus of strategies to slow the rate of development of anthelmintic resistance in gastro-intestinal parasites has centred on the need to better manage the parasite population that is unexposed to the drug, i.e. in refugia. One strategy involves the use of targeted selective treatments (TST) where anthelmintics are only administered to those individuals that are likely to benefit from it, rather than to the whole flock. However, there are few TST indicators currently available for farmers in temperate environments. Here we describe the development and field evaluation of a performance-based decision support model to identify individual animals whose performance is suffering from nematodoses, and thus, likely to benefit from an anthelmintic treatment. One hundred and ninety two twin-suckled lambs were separated into eight groups and each grazed on a one-hectare paddock and exposed to one of four anthelmintic treatment regimes. Groups of lambs received either: a whole flock anthelmintic treatment either every four weeks (neo-suppressive treatment, NST); when clinical signs of disease were evident (metaphylactic/therapeutic treatment, MT); at pre-determined times throughout the grazing season (strategic treatment, SPT) or treatment on an individual animal basis as determined by a performance-based decision support model that relied on a calculation of the efficiency of nutrient utilization (targeted selective treatment, TST). Animal performance and parasitological data were recorded every two weeks. Mean faecal egg counts were low throughout the season for the NST group and moderate for the MT, SPT and TST groups. During the entire grazing season the NST, MT, SPT and TST animals each received on average 5.0, 2.0, 3.0 and 2.6 anthelmintic treatments, respectively. Compared to the NST animals, cumulative weight gains were 11% lower at the end of the grazing season in MT animals, whereas no reduction was observed in either the SPT or TST groups. The proposed decision support system appeared to be successful in discriminating between animals which were likely to respond favourably to anthelmintic treatment and those that were not and provided a quantitative estimate of the likely response. Furthermore, 87% of animals that had a pre-treatment efficiency of nutrient utilization value lower than the suggested threshold of the model responded positively to treatment. The potential benefits of using the decision support model as a TST indicator for managing the parasite population in refugia in temperate environments are discussed.