OBJECTIVES: The non-peptidic protease inhibitor (PI) tipranavir boosted with ritonavir (TPV/r) has shown superior efficacy compared to investigator selected ritonavir-boosted comparator PI (CPI/r) in treatment experienced HIV 1 infected patients in the RESIST 1 and 2 clinical trials. TPV/r or CPI/r were administered with an optimized background regimen (OBR). In addition to the clinical efficacy of TPV/r, healthcare decision-makers will be interested in the cost-effectiveness of TPV/r. METHODS: A previously published 3-stage Markov model was modified to reflect US practice patterns for treatment-experienced HIV patients using 2005 costs and combined 48-week RESIST 1 and 2 trial data. The 12 model health states (HS) were defined by patients' CD4+ count and viral load, with cost and risk of AIDS defining events (ADEs) being linked to each HS. Cycle length was three months and transition through the model continued until 90% of patients had died. Disease progression beyond the 48-week trial data was based on HAART treatment experienced patients from the Medical University of South Carolina database. Costs were estimated from the payer perspective, including AWP drug prices and costs for patient monitoring and ADEs from SC Medicaid patients. RESULTS: TPV/r patients remained longer in HS defined by higher CD4+ count and lower viral load compared to CPI/r patients. This reduced the rate of ADEs (12.35% over 5 years) and resulted in 9.6 quality-adjusted life-months gained over the model time horizon. The incremental cost-effectiveness ratio (ICER) of TPV/r vs. CPI/r was $56,668 per QALY (discounted at 3%). Subgroup analysis of patients not treated with enfuvirtide as part of their OBR reduced the ICER to $49,467 per QALY. CONCLUSIONS: The ICER for TPV/r falls well within the $/QALY range of $50,000-$100,000 considered acceptable. Treating patients with TPV/r in earlier treatment regimes, not requiring augmentation with enfuvirtide, yielded an ICER below $50,000 per QALY.