Background: Lifelong prophylactic replacement therapy with clotting factor concentrates is recommended for severe haemophilia. The prophylactic dose determines both clinical outcome and treatment cost. In the absence of clinical studies, computer simulation was used to explore lifelong effects and clotting factor consumption for various prophylactic dose levels, and optimize strategies for switching between prophylactic and on-demand treatment. Design and Methods: Individual patients' lifetime joint bleeds, radiological arthropathy (Pettersson score, 0-78) and consumption were simulated for each treatment strategy. Treatment effectiveness (expressed as % of patients maintaining a lifetime Pettersson score ≤14) and clotting factor consumption were modelled for lifelong prophylaxis at dose levels 1000-4500 IU kg À1 year
À1, for on-demand treatment and for switching strategies. Treatment efficiency (consumption per unit of effectiveness) was used to compare strategies. Results: Compared to lifelong on-demand treatment, lifelong prophylaxis at 1000 IU kg À1 year À1 increased effectiveness from 21 to 36%, at an additional consumption of 0.9 9 10 6 IU kg À1 . For lifelong prophylaxis, each additional 1000 IU kg À1 year À1 resulted in a proportional increase in consumption by AE5 9 10 6 IU kg À1 but a less than proportional reduction in arthropathy by AE50%; consequently, increasing consumption progressively diminished treatment efficiency. Switching strategies slightly reduce effectiveness and consumption. Optimum switching criteria were similar across prophylactic dose levels. Conclusion: According to the simulation model, low-dose prophylaxis (1000 IU kg À1 year
À1) improved outcome at a limited increase in consumption compared to ondemand treatment. Increasing prophylactic dose further improved health outcomes, but at decreasing efficiency. Optimal prophylactic dose should therefore be selected balancing acceptable health impact and available budget.