OBJECTIVES To estimate the long-term health economic outcomes for patients treated with either bevacizumab added to chemotherapy or chemotherapy alone, as adjuvant treatment for patients with triple-negative breast cancer (TNBC). METHODS We used a three-state Markov Model (Disease-Free Survival (DFS), Progressed, Death) to estimate the impact of adding one year of adjuvant treatment with bevacizumab to an established chemotherapy regimen. The probability of disease recurrence was derived from a prospective clinical database analysis of 255 patients with early TNBC. The treatment effect of bevacizumab was incorporated by applying a relative risk reduction to the underlying risk of disease recurrence, using the expected hazard ratio (0.75), from the statistical power calculations of the on-going phase III BEATRICE study; a trial recruiting solely patients with early TNBC. Total direct medical costs and quality-adjusted life expectancy (QALYs) were projected over patient lifetimes from the perspective of the UK NHS. Clinical and economic outcomes were discounted at 3.5% per annum. Probabilistic and univariate sensitivity analyses were performed. RESULTS Mean discounted quality-adjusted life expectancy was projected to increase by 0.83 QALYs following the addition of bevacizumab to chemotherapy versus chemotherapy alone (11.35 vs 10.53 QALYs). Over patient lifetimes, discounted total direct costs were estimated to be higher with bevacizumab, primarily attributed to higher pharmacy costs and the increased time spent in DFS. These results corresponded to an incremental cost-effectiveness ratio of £43,804/QALY gained. Sensitivity analysis indicated results were most sensitive to the duration of the treatment effect of bevacizumab and patients' baseline characteristics (height and weight). CONCLUSIONS Short-term reductions in the probability of disease recurrence associated with adding bevacizumab to an adjuvant chemotherapy regimen for patients with early TNBC were projected to improve quality-adjusted survival and potentially be cost-effective given recent discussions concerning willingness-to-pay thresholds in the UK.