We adjusted our model to reflect the new and higher incidence of rotavirus gastroenteritis and included the indirect protection for individuals aged 5 years and older using data from Friesema et al.
] and Lopman et al.
]. In the previously published study, we also included herd immunity but only in one specific scenario analysis and only for individuals aged less than 5 years. In the current analysis, herd effects for children under five years of age were calculated exactly as it was done in the original study. It meant we assumed herd protection benefits for those not yet (fully) protected by the vaccine (either too young to be vaccinated or those who had not yet received the complete set of doses) and non-vaccinated children (5% of a birth cohort for the Dutch situation), assuming protection would be as effective as the vaccination would be after the completion of all doses
]. To account for herd effects for individuals of 5 years and older, we used data from Lopman et al.
], where the rotavirus discharge rates for different age groups, e.g., 5–14; 15–24; 25–65 and
65 years old were compared between the post and prior rotavirus vaccination periods. However, we decided to include herd protection only for age groups of 5–14 and 15–24, where there were statistically significant changes in reduced RVGE hospitalized cases
]. In order to capture this indirect protection resulting from rotavirus vaccination, we have reduced the incidence of hospitalized cases due to rotavirus infection by the same rate as indicated by the US rotavirus discharges. In details, the authors reported the change of RGVE hospitalized cases comparing the pre- (2000–2006) and post-vaccination eras (2008) using the relative risk (RR). The RRs for rotavirus-coded hospitalization discharges in 2008 compared with the prevaccine period of 2000–2006 were 0.29, 0.35 for age groups of 5–14 and 15–24, respectively. The choice to base the indirect effects on this specific study was made as it reported the actual decrease in RVGE cases from a large database covering ~20% of all US hospital admissions after 2-year follow up since the introduction of rotavirus vaccination in the US in 2006. We felt that using these actual epidemiological data was a conservative approach and would provide a better picture of the impact of rotavirus vaccination as it also reflected the change in vaccine uptake. With the increasing evidence of herd protection in unvaccinated or otherwise unprotected children below 5 years of age, we also focused on the CoRoVa-model scenario that included that type of herd protection
In the original study, we included conservative incidence numbers based on the data provided by Mangen et al.
], where the RVGE hospitalized cases were estimated at 3,600
]. As mentioned, recent studies indicated relevantly higher annual hospitalizations of at least 5,000 cases below 5 years of age. Conservatively, this lower bound was used in the analyses as ballpark figure for total rotavirus hospital cases (including nosocomial infections) in the Netherlands
]. For the sake of comparability, we recalculated results applied to the birth cohort of 180,000 newborns as used in the original publication. In order to calculate the age-specific disease distribution, we applied the age-specific hospitalization distribution which divided the total estimated number of cases by the different age groups exactly as indicated in the original study
]. The relative increase for the hospitalizations <5 years was also applied to those above 5 years of age (which may again be considered conservative
Except for updating the new hospitalized RVGE cases and including the herd effects, we applied the same vaccine efficacy, waning vaccine protection, Dutch discount rates for health effects and costs of 1.5% and 4%, respectively; 2010 consumer price index acquired from the Netherlands’ Central Bureau of Statistics, no care-giver quality of life loss and treatment cost for rotavirus cases of all categories, as we did in the original study
]. In the absence of the vaccine price, we applied a total cost of €75 per fully vaccinated child. The threshold willingness-to-pay was conservatively assumed to be €20,000 per quality-adjusted-life-year (QALY), with additional calculations at €50,000/QALY
]. All the analyses were done from the societal perspective, inclusive parental work absenteeism, however excluding indirect costs from any deaths in infants due to rotavirus according to the Dutch friction costing method for indirect productivity costs
]. As there was no data on QALYs for children and adults five years and older and there were no deaths reported due to rotavirus infection for older children and adults, we only estimated the direct and indirect costs saved due to hospitalized cases and conservatively excluded any QALY effects.