Anthropogenic influences on habitats often affect predation on species by introducing novel predators, supporting additional predators, or reducing animals’ ability to detect or avoid predators. Other changes may reduce the ability of animals to feed, or alter their energy use. An increase in predation risk is assumed to reduce prey populations by increasing mortality, reducing foraging and growth. Often animals don’t appear to have been adversely affected, or may even increase growth rate. However, theoretical predictions that may have been overlooked suggest that optimal foraging rate, mortality rate and growth rate may change in counter-intuitive ways, depending on exactly how predation risk or costs have been increased. Increasing predator density may increase mortality rate when foraging, reduce the safety of refuges, or alter the relationship between vigilance and attack likelihood. Increasing temperature may increase metabolic costs in ectotherms and reduce thermogenesis costs in endotherms, which affects the costs of active foraging and inactivity differently. Here, I review the theory on how predation risk and metabolic costs should affect foraging behaviour, mortality and growth in order to explain the great variation in behavioural responses. I show that in some situations animals may not respond behaviourally even though a change severely affects survival, and the mortality may be a poor metric of the impact of a change on population viability. In other situations a fitness proxy may change dramatically whilst fitness is unaffected due to compensatory changes in behaviour or life history. Other measures may change in a positive way whilst fitness declines. I describe how to identify the situations in the field and thereby make reliable measure of fitness in particular study systems. Overall, this work shows how behavioural theory can help understand the impacts of environmental change and highlights promising directions to better understand and mitigate their effects on ecosystems.