Fig. 1.

Forest plots of individual study results and meta-rate ratios for select cancers among male career firefighters compared to a general, uniformed service, or working population referent.^†

† Random-effects models were used with the restricted maximum-likelihood estimator. Hartung-Knapp-Sidik-Jonkman (HKSJ) adjustments and an ad hoc variance correction were used to calculate confidence intervals for summary estimates. Calculated study intervals may differ from reported values due to differences in variance estimation methods. Presented cancer sites were chosen for having positive associations and relevant exposures related to the firefighting occupation.