Aquaculture is currently the primary source of fish for human consumption. However, the sustainability of this industry has been under scrutiny. In Chile, the main concern is the use of antibiotics to control piscirickettsiosis, and farmers have identified timely detection as a critical issue. Using data provided by the Fisheries and Aquaculture authority (Sernapesca), we assessed whether the sampling strategy or the time of the first detection of Piscirickettsia salmonis would affect the time to onset or the magnitude of an outbreak. We modeled time to onset using an Aalen additive-hazards survival model, and found that on average, shorter times between diagnosis and treatment were associated with earlier onset of disease. Piscirickettsiosis also occurred earlier when fish received a sea lice bath treatment at any point before the outbreak, and when farms submitted the samples, but these effects waned past 30 to 35 weeks into the production cycle. To assess the impact of these predictors on the proportion of fish that die during an outbreak, the magnitude of an outbreak, we used a mixed-effects linear model, controlling for the level of mortality when the treatment started. We found that mortality due to piscirickettsiosis was higher when the diagnosis was made closer to treatment; however, this was only observed when samples were obtained during an active surveillance activity. Our results showed that disease occurs later, and fewer fish die when the time between the detection of the agent and the occurrence of an outbreak is longer, especially when samples are collected during active surveillance activities. This suggests that active surveillance may be more likely to lead to early detection, and give the farmers time to implement adequate control measures. Despite the limitations of this study, the results suggest that the current surveillance program is effective in delaying and reducing the impact of piscirickettsiosis outbreaks, and that farmers may benefit from increasing their own surveillance efforts.