Conditional vaccine performance benchmarks for reducing MRSA burden in an idealised national transmission model for Italy

Vaccine development against antimicrobial-resistant pathogens benefits from quantitative benchmarks, but national burden estimates cannot identify all conditions required for impact. We developed a deterministic compartmental model of methicillin-susceptible and methicillin-resistant Staphylococcus aureus colonization, infection, recovery, attributable death, vaccination, and waning immunity, representing an idealised national risk pool for Italy. Annual ECDC MRSA infection estimates for 2016–2020 were the primary calibration targets. Weakly identifiable parameters were sampled from literature-informed distributions and retained using an explicit acceptance algorithm with biological plausibility constraints. The ensemble contained 450 accepted parameter sets from 12,000 attempts. In the primary scenario, efficacy was 70% against MRSA acquisition and progression, uptake was 20%, and mean protection duration was 3 years. Median 10-year reductions were 83.8% for cumulative MRSA infections (2.5th–97.5th percentiles, 68.5%–94.3%) and 83.3% for attributable deaths (67.9%–94.0%). The fraction attaining at least 75% infection reduction was 0.90, and the median protected proportion in year 10 was 36.0%. Assuming that 25% or 50% of acquisition pressure arose from an untreated external reservoir reduced median infection impact to 53.7% and 40.2%. These are conditional scenario benchmarks, not policy thresholds or forecasts for a vaccine. Protection against acquisition was the main driver of impact, whereas external reservoirs and strain ecology reduced or widened projected effects. The framework supports target-product discussions and prioritisation of data collection rather than deployment decisions.