A generalized SIRVS model incorporating non-Markovian infection processes and waning immunity

The Markovian approach, which assumes constant transmission rates and thus leads to exponentially distributed inter-infection times, is dominant in epidemic modeling. However, this assumption is unrealistic as an individual's infectiousness depends on its viral load and varies over time. In this paper, we present a SIRVS epidemic model incorporating non-Markovian infection processes. The model can be easily adapted to accurately capture the generation time distributions of emerging infectious diseases, which is essential for accurate epidemic prediction. We observe noticeable variations in the transient behavior under different infectiousness profiles and the same basic reproduction number R0. The theoretical analyses show that only R0 and the mean immunity period of the vaccinated individuals have an impact on the critical vaccination rate needed to achieve herd immunity. A vaccination level at the critical vaccination rate can ensure a relatively low incidence among the population in case of future epidemics, regardless of the infectiousness profiles.