Modeling and predicting the long-term effects of various strategies and objectives of varicella-zoster vaccination campaigns.

Background: Susceptible, exposed, infected, and recovered (SEIR) models are increasingly developed and used, but their simplicity contrasts with the wide variety of scenarios before launching vaccination campaigns. Methods: We investigated the effects of some model-building choices (targets, pace, coverage rate) on the results of SEIR models in the case of vaccination against varicella and herpes zoster. Results: The analysis demonstrated the need for a progressive unvaccinated to vaccinated transition and a dynamic system-equilibrium before vaccination onset. When several doses are considered, new compartments are needed to account for vaccination histories. For varicella, the delay to reach the expected coverage rate and the pace until reaching this rate have significant impacts, especially on the short-term incidence. The impact of vaccination through herd immunity should be systematically investigated. Conclusions: Graphs help understanding the progress of instantaneous incidence; however, tables of cumulative average incidence over decades should be preferred because of higher stability.