The Great Lakes Water Quality Agreement specifies “fish tumors or other deformities” as one of the 14 beneficial use impairments that can be used to declare a geographic area within the Great Lakes an Area of Concern (AOC). The International Joint Commission has suggested that the fish tumor impairment can be delisted when fish tumor incidence in the AOC does not exceed rates at unimpacted control sites. This paper presents a statistical technique utilizing Bayesian hierarchical logistic models to estimate tumor incidence on brown bullheads (Ameiurus nebulosus) in an AOC and in candidate least impacted control sites (LICS). Liver and skin tumor incidence are estimated using age, length, weight, and gender as possible covariates using a hierarchical framework to account for a sampling design in which sites are sampled over multiple years and/or at multiple sublocations within the site. By utilizing a Bayesian approach, estimates of uncertainty for tumor incidence in sites with no observed tumors can be obtained. The posterior distributions of tumor incidence can then be used to identify LICS for the watershed and subsequently compare the tumor incidence in the AOC to the LICS using a Bayesian form of the two one-side tests for equivalence procedure. Presque Isle Bay (Erie, PA) in the Lake Erie watershed is used as a case study to demonstrate the technique.