Using available data on the occurrence of Salmonella enteritidis (SE) in US layer flocks and eggs, and a probabilistic scenario tree method, an estimate of the fraction of SE-contaminated eggs produced annually is derived with attendant uncertainty. In lieu of a definitive prevalence survey, the approach presented here provides insight to the relative contribution of various pathways leading to contaminated eggs. A Monte Carlo model with four branches is developed. The first branch predicts the proportion of all US flocks that are SE-affected. The second branch apportions SE-affected flocks into three categories (high, moderate, and low level affected flocks) based on population-adjusted epidemiologic data. The third branch predicts the proportion of affected flocks that are molted and producing eggs during a high risk period subsequent to molt. The fourth branch predicts the fraction of contaminated eggs produced by flocks of the type described by the pathway (e.g. high level affected flocks that are not molted) based on egg sampling evidence from naturally infected flocks. The model is simulated to account for uncertainty in the data used to estimate the branch probabilities. Correlation analysis is used to estimate the sensitivity of model output to various model inputs. The output of this model is an uncertainty distribution for the fraction of all eggs that are SE-contaminated during 1 year of production in the US. The expected value of this distribution is approximately one SE-affected egg in every 20,000 eggs annually produced, and the 90% certainty interval is between one SE-contaminated egg in 30,000 eggs, and one SE-contaminated egg in 12,000 eggs. The model estimates that an average of 14% of all eggs (i.e. contaminated and not contaminated) from affected flocks are produced by high level, non-molted affected flocks, but these flocks are estimated to account for more than two-thirds of the total fraction of contaminated eggs produced annually. Sensitivity analysis also suggests that the proportion of affected flocks that are high level flocks - and the egg contamination frequency for these types of flocks - are the most sensitive model inputs. The model's pathways provide a framework for evaluating interventions to reduce the number of contaminated eggs produced in the US. Furthermore, sensitivity analysis of the model identifies those inputs whose uncertainty is most influential on the model's output. Future farm-level research priorities can be established on the basis of this analysis, but public policy decisions require a fuller exposure assessment and dose-response analysis to account for microbial growth dynamics, meal preparation, and consumption demographics among US egg consumers.