OBJECTIVES:  Microsimulation is the preferred modeling approach to assess comparative effectiveness and cost-effectiveness of cancer screening strategies. These models often involve multiple comparators from strategies formed by a combination of initiation/cessation age and screening intervals. Using probabilistic analysis to present findings in models with multiple comparators is challenging because the probability of cost-effectiveness not only depends on the relative costs and screening effectiveness among alternative strategies but also is sensitive to the number of comparators in the model. METHODS:  We developed a microsimulation model to compare the cost-effectiveness mammography screening strategies for average-risk women. Scenario A includes 10 strategies covering guidelines from three professional societies with three cessation ages (75/80/none), plus the no-screening option. Scenario B expands to 28 strategies by varying initiation (40/45/50), cessation (75/80/none) age, screening intervals (annual/biennial), and hybrid strategies that transition from annual to biennial. We obtained clinical parameters from the literature or statistical modeling (e.g., age-dependent sojourn time) and cost inputs from Medicare fee schedule and analyzing SEER-Medicare data. To address uncertainties, we ran 100 repetitions of the model, each simulating individual woman’s lifetime natural history for a birth cohort of 100,000. Simulation results were converted to relative risk ratios (RRR) using multinomial logistic regressions. RESULTS:  At $100,000/QALY willingness-to-pay, Scenario A showed the most cost-effective strategy (annual 45-54, biennial 55-75) had 31% chance to yield the highest net benefit, whereas the probability reduced to 19% for the most cost-effective strategy (biennial 40-75) in Scenario B. Using the next-best strategy as the base category, multinomial model showed that the most cost-effective strategy was 1.15 and 1.36 times more likely to yield the highest net benefit in Scenario A and B, respectively. CONCLUSIONS: Presenting uncertainty in RRR can mitigate trends toward a lower probability for the most cost-effective strategy in models with a larger number of comparators.