OBJECTIVES: We demonstrate an approach to assess the impact of uncertainty in risk predictions on health-economic outcomes in risk-stratified prevention strategies, illustrated for preventive statin treatment based on 10-year coronary heart disease (CHD) risk predicted by the Framingham risk score (FRS). METHODS: We refitted the FRS to three random samples of increasing size (N=2,500, N=1,000, N=500) from a population-based cohort. A Markov decision-analytic model was used to simulate cohorts with preventive statin treatment in high-risk (FRS≥20%) individuals (ATPIII guideline). This treatment threshold was incrementally lowered to T=0.0% with 0.5% decrements. Using the cohort the distribution of individuals over the low (<0.5T%), intermediate (0.5T%-T%), and high (≥T%) risk category and corresponding observed CHD-risks were calculated. The Net Health Benefit (NHB) (willingness-to-pay of $50,000/QALY) was calculated and uncertainty in outcomes was assessed with probabilistic sensitivity analysis. The NHB for each 0.5% risk category was calculated to assess the impact of risk prediction uncertainty on associated uncertainty in NHB. RESULTS: Prediction model performance was fair in all samples (c-statistic: 0.70-0.75). Prediction uncertainty resulted in probabilities of incorrect treatment decisions of up to 0.4 (N=2,500) and 0.5 (N=500) for risk around T=20.0%. The NHBs per risk category ranged from -0.031 for a predicted risk of [3.5%;4%] to 0.020 for [10.0%;10.5%] in men and from -0.067 at [0.0%;0.5%] to 0.045 for [2.0%;2.5%] in women. The NHB was positive for predicted risks >12.5% in men and >11.5% in women. For individuals with predicted risks <7.5% or >20%, in 95% of PSA simulations the NHB was negative or positive, respectively. CONCLUSIONS: Risk-stratified prevention is increasingly recommended. While uncertainty in risk predictions may lead to incorrect treatment decisions, associated impact on long-term health-economic outcomes is unknown. Assessing this impact can guide studies aiming to improve prediction models by focusing on individuals for which improvement may actually improve health-economic outcomes.