OBJECTIVES: To explore the contribution made by disease transmission modelling to cost effectiveness analyses. METHODS: Traditional cost-effectiveness analysis quantifies the costs and effects accumulated by an average individual exposed to a particular intervention, relative to one or more suitable comparators. When reacting to an infectious disease, however, many interventions alter the natural history of an infection or individuals' behaviour in ways that affect the onward transmission of the pathogen. This, in turn, may influence the number of secondary infections generated by each infectious case. For a cost-effectiveness analysis to account for the averted/additional cases, a population level perspective including disease transmission modelling is required, but this comes at a cost. Transmission models take time to construct and parameterise and are often data hungry. Do the insights these models provide justify the investment in time and expertise they require? RESULTS: This exposition outlines the basic concepts underlying disease transmission modelling, presents a simple model for a directly transmitted disease (such as influenza) and demonstrates the enormous impact population level effects can have on the outcome of a cost-effectiveness analysis. This impact is most acute when the members of a large population can mix freely (there is little compartmentalisation), for example across different age groups, there is significant transmission (the mean number of secondary infections generated by each primary infection is greater than one) and where transmission is rapid (short incubation period). This final condition is important if the cost-effectiveness analysis has a short time horizon, particularly where costs are accumulated sooner than benefits (as in a vaccination program). This has been demonstrated in the analysis of both varicella zoster and hepatitis B virus vaccination programs. The circumstances in which transmission modelling is of greatest value will be further discussed, as too will those in which an individual based analysis suffices.