OBJECTIVES: To assess the comparative costs and outcomes of HIV treatment algorithms with and without the use of raltegravir for treatment naïve patients. METHODS: A cost-utility analysis was constructed using a Markov model to assess the incremental cost per QALY gained of a HIV treatment algorithm which included raltegravir as a first line treatment compared to standard practice in Australia. Comparator treatment algorithms were constructed for patients initiating treatment with a non-nucleoside reverse transcriptase inhibitor (NNRTI) containing regimen and patients initiating treatment with a protease-inhibitor (PI) containing regimen. The economic model was a Markov cohort model with the cohort transited through five health states which reflect four lines of the treatment in each algorithm and death. The modelled population cycled through the treatment algorithm upon loss of virologic response and at each cycle are at risk of adverse events, AIDS events and death based on the treatment regimen received, viral load and CD4 count. The efficacy and safety of each treatment at each line of the algorithm was based on randomised clinical trial data and a mixed treatment comparison. Costs were calculated from the perspective of the Australian healthcare system. The economic model was a lifetime model. RESULTS: The raltegravir algorithm was more costly than the NNRTI algorithm (A$4487 per patient) but less costly than the PI algorithm (A$18,383 per patient). The raltegravir algorithm had greater QALYs than both the comparator algorithms (0.17 and 0.22 per patient compared to the NNRTI and PI algorithms respectively). The incremental cost per QALY of the raltegravir algorithm compared to the NNRTI regimen was A$26,896 which is considered within an acceptable threshold in Australia. The raltegravir algorithm dominated the PI algorithm (lower costs and greater QALYs). CONCLUSIONS: Raltegravir for treatment naïve patients is an effective and cost-effective intervention relative to current practice in Australia.