To estimate the effects of key uncertainties on the effectiveness and cost-effectiveness of breast cancer prevention with tamoxifen. The incremental cost-effectiveness ratio of tamoxifen therapy relative to placebo was estimated using decision analysis with Markov modelling of health states, outcomes and costs for a simulated cohort of women at high risk for breast cancer. Relative effects of tamoxifen's benefits and harms were estimated from meta-analyses of randomised controlled trials. Cost estimates were based on Australian treatment patterns and costs. The main outcome measure was cost per quality-adjusted life year (QALY) gained with costs and effects discounted at a 5% annual rate. Tamoxifen therapy over five years reduces the incidence of breast cancer by approximately 1.4%, which is offset by an increase in endometrial cancer of 0.7% and pulmonary embolism of 0.2%. If the reduction is permanent (preventing new breast cancers emerging over five years and no further treatment effect thereafter), the model estimates an increase in life expectancy of 0.057 QALYs and an extra cost of $2,193; or $38,271/QALY gained. A model assuming further treatment effects of tamoxifen preventing new breast cancers emerging for up to 10 years results in an incremental cost of $19,354/QALY. However, if five years of tamoxifen therapy merely delays when these breast cancers appear (such that by 10 years there is no longer a reduced incidence), the incremental cost per QALY saved is estimated to be $199,149. Tamoxifen is potentially cost-effective in preventing breast cancer in women at high risk. However, its cost-effectiveness as a preventive therapy is highly sensitive to whether these cancers are permanently prevented or their clinical presentation is only delayed. Long-term follow-up in randomised controlled trials is therefore crucial in forming health policy.