Objective: We sought to systematically review economic analyses (EAs) of HER2 testing and trastuzumab therapy in all stages of breast cancer (BC) with specific attention to the methodological quality, quantification of uncertainty and incorporation of diagnostic test characteristics. Methods: EAs of trastuzumab in BC or HER2 diagnosis with either immunohistochemistry or fluorescence in situ hybridisation techniques were considered. Biosis, Cochrane, CRD, EconLit, Embase, HEED, Medline and PubMed databases were searched. The reference lists of each retrieved article, relevant reviews, and abstracts of the San Antonio Breast Cancer Symposium were hand-searched. Citations were reviewed in duplicate and relevant articles were qualitatively rated per Drummond. Results: Twenty studies, conference abstracts and health technology assessments were selected for full review from among 641 citations as of December 2007 (reviewer agreement kappa=0.85). Studies examined trastuzumab in metastatic (7/20) or adjuvant (10/20) settings or had a testing focus (4/20). HER2 diagnosis strategy and trastuzumab treatment were evaluated jointly in only one study. Few decision models were calibrated against epidemiological data (3/20). Probabilistic sensitivity analysis was infrequently used to characterise uncertainty (3/20) and decision uncertainty in the form of cost-effectiveness acceptability curves was presented in a single study. The overall reported quality of EAs was comparatively poor. Conclusion: Testing and treatment were rarely examined in tandem, despite a 2004 EA addressing this very issue in metastatic disease. Given the controversy around trastuzumab funding in many jurisdictions, the need for adequate attention to testing and uncertainty analysis is not met in the literature.