The receptor-binding assay (RBA) method for the detection of paralytic shellfish poisoning (PSP) toxins was evaluated for its overall performance in comparison with the mouse bioassay (MBA). An initial study to evaluate the effects of filtering shellfish extracts prior to running the RBA indicated no significant difference between filtered and unfiltered extracts on the determined saxitoxin (STX) concentrations. Next, we tested the RBA assay on 295 naturally contaminated mussel tissue samples, ranging in concentrations from 320 µg STX equiv. kg-1 to 13,000 µg STX equiv. kg-1 by MBA. An overall trend was observed with the RBA giving higher results (256 µg STX equiv. kg-1 on average) than the MBA; however, at low concentrations (< 500 µg STX equiv. kg-1) the RBA results were marginally lower. A third study was conducted using spiked mussel tissue analysed by three independent laboratories, two of which performed the RBA and one the MBA. This multi-laboratory study again showed the RBA to give higher results than the MBA; however, it also revealed that STX determination was accurate by the RBA, unlike the MBA. To optimise the assay for efficient usage under regulatory practice, three suggestions have been made: the use of an initial screening plate to separate those samples that exceed the alert level; use of rapid PSP test kits in the field and in the laboratory for screening negative samples and for early detection of toxicity; and use of an alternate commercially available porcine membrane in place of the laboratory-prepared rat membrane homogenate. The large number of samples analysed and the diversity of the tests conducted in this study further support the RBA as an affordable rapid method for STX detection that is also free of the routine sacrifice of live animals.