Targets and measures: Challenges associated with reporting low sea lice levels on Atlantic salmon farms

A popular framing of Goodhart's Law states, When a measure become a target, it ceases to be a good measure. The extent to which this may be the case in the reporting of sea louse infestation on salmon farms is explored here. Due to the importance of controlling sea louse infestation on salmon farms, monitoring programmes are active in most salmon producing regions and, in many, a maximum allowable sea louse level is specified. Using publicly accessible data from Norway and BC, Canada, this study investigated the extent to which the framing of these programmes, in particular the specification of low threshold levels, may be affecting the veracity of the reported sea louse infestation data. In BC, where the threshold level is set to 3 mobile Lepeophtheirus salmonis little evidence of anomalous patterns in the data and the overall proportion of females within the adult sea lice population is around 0.43. By contrast, in Norway where lower sea louse limits are in place (at either 0.5 or 0.2 adult female L. salmonis), there is evidence of unexpected and sharp reductions in the abundance of adult females reported around these threshold values. In addition, the average proportion of females is estimated to be only around 0.20 of the total adult L. salmonis population. The unexpected observations in the data were much more evident for farms in the southern areas of Norway and over the most recent years. These findings appear to support the case that the measurement of sea lice on salmon farms can be significantly influenced by targets (particularly those which are highly demanding), and that as such, researchers and fish health professionals should be aware of potential biases within these data. In addition, regulators should carefully consider the unintended consequences of setting certain sea louse thresholds and the ways in which the potential to effectively review data quality and accuracy may be impacted by the choice of sea louse stage(s) that are reported.

Automated summary

This study explores the impact of target setting on the reporting of sea louse infestation on salmon farms. The findings suggest that the accuracy of reported data can be influenced by the specification of threshold levels. Researchers and fish health professionals should be aware of potential biases in these data, and regulators should carefully consider the unintended consequences of setting certain threshold levels.