Abstract Johnsen et al. (2021) present a model that predicts lice-induced mortality of virtual post-smolt (VPS) Atlantic salmon migrating from 401 Norwegian salmon rivers. An important feature of this model is that the source of lice is farmed salmon. The model predicts lice-induced mortalities exceeding 30% for VPS from 82 rivers, 10–30% for 140 rivers, and less than 10% for 179 rivers. These are high numbers and have contributed to mandatory reductions in the production of farmed salmon in two of the 13 management areas in Norway. In this comment, the model used by Johnsen et al. (2021) is evaluated. Mortality estimates from observed lice numbers on post-smolts caught by trawling in outer fjord areas are compared with the model predicted mortalities for VPS migrating from rivers draining into the corresponding fjords. It is found that Johnsen et al. (2021) systematically overestimates lice-induced mortality for VPS compared to estimates from trawled smolt from corresponding fjords. The present analysis thus contradicts the concluding remark in Johnsen et al. (2021), that: “The estimated lice level on modelled post-smolts is shown to coincide with the level observed on captured wild post-smolt”.