Their model may therefore underestimate the number of symptomatic infections observed. Secondly, the models differ in assumptions regarding immunity and re-infection. The model Ruxolitinib presented here assumes that a fraction of individuals gain long-term immunity after each episode of disease. Pitzer et al. assumed a period of temporary but complete immunity after each infection waning at a constant rate with a mean duration of 9–12 months. We chose not to assume a period of complete protection, as studies looking at protection

conferred by natural infection in children have shown that up to four re-infections are possible within a two-year study period [15] and [18]. Thirdly, supported by household studies [19], [20], [21] and [22], we assumed that only symptomatic individuals are infectious and important in transmission, whereas Pitzer et al. assumed that all infections, to varying degrees, play a role in transmission (symptomatic infections > asymptomatic infections). In addition, we modelled all symptomatic infections in the population as opposed to modelling only severe symptomatic infections and, unlike Pitzer et al., we had an independent estimate of the reporting efficiency (under-ascertainment of rotavirus disease cases within the surveillance data), and so we did not have to estimate this and the transmission parameters (which could pose identifiability problems). In addition, we used a detailed dataset

www.selleckchem.com/products/ch5424802.html on contact patterns for Great Britain to improve parameterisation of the model and to help inform assumptions about mixing patterns between age groups. Despite these differences in model assumptions, the results of our model regarding the effect of vaccination are very similar to those of Pitzer et al., suggesting that the results are robust to slight differences in model structure.

Pitzer et al. also demonstrated that spatiotemporal variations in the size and timing of the peak in rotavirus disease could be explained by variations in birth rate. We incorporated into our model year-specific birth rates for England and Wales between 1998 and 2007. It did not improve the fit of the model or predict the slight fluctuations in the size or timing of the epidemics seen from year to year. Variability in birth rates over time observed in England and Wales are less marked than those in the United GPX6 States. This could explain why, unlike in the model developed by Pitzer et al., varying annual birth rates in our model was not important. Our model predicts that there will be an increasing decline in numbers of reported cases and delay in the start of the season in the first two years post-vaccination. Interestingly, a slight increase in numbers is predicted in the third post-vaccination year compared to the second. These predicted early dynamics capture the observed effects of vaccination seen in the United States [36] and [37] and are similar to those predicted by Pitzer et al. [29].