Peak incident management level affects rates of injury on large federal wildland fires

Background/Purpose: Wildland firefighters are a high risk occupational group about which little is known. Our objective was to evaluate the effect of peak incident management level (PIML) on rate of injury for federal wildfires. Methods: Data for injuries, person-days worked and fire characteristics for federal fires reported to the National Interagency Fire Center from 2003-2007 were used. Logistic regression was used to assess fire-level risk factors for the odds of any injury. Negative binomial regression was used to examine incidence rate ratios associated with fire-level risk-factors. Results/Outcomes: The adjusted odds ratios for the odds of any injury reported versus no injury were 5.36 (95% CI 2.98 – 9.64) for PIML I fires as compared to PIML III fires and 6.49 (95% CI 4.36 – 9.65) for PIML II fires as compared to fires with a PIML of III. In the adjusted logistic model, both PIML and the fire's resistance to control were significantly predictive of any injury occurrence. The adjusted incident rate ratio (IRR) for fires with a PIML I as compared to fires with a PIML III was 0.36 (95% CI 0.25 – 0.53). The adjusted incident rate ratio for fires with a PIML II as compared to PIML III was 0.74 (95% CI 0.56 – 0.99). In the adjusted model, the only risk factor significantly predictive of rate of injury was the PIML. Conclusions: Peak incident management level of a wildfire is predictive of both odds of any injury occurrence and IRR on wildfires, but with contrasting effect.