Estimating Rotavirus Burden and Mortality

We estimated approximately 10 000 symptomatic rotavirus gastroenteritis cases per 100 000 children aged younger than 5 years per year on the basis of a global systematic review and meta-analysis by Bilcke and colleagues.²⁶ We used WHO region estimates of the proportion of all-cause gastroenteritis cases that are severe (defined as children with moderate or severe dehydration), as a proxy for the proportion of rotavirus gastroenteritis cases that were severe (and non-severe).²⁷ (link) To calculate numbers of rotavirus deaths in each country (without vaccination), we estimated means (and 95% CIs) using country-specific estimates from three difference sources (Institute for Health Metrics and Evaluation, Maternal Child Epidemiology Estimation, and WHO US Centers for Diseases Control and Prevention) for the year 2015.1 (link), 3 (link), 4 (link) We elected to use the mean because the range reported by the three different sources was from 158 000 to 202 000 deaths a year in children younger than 5 years for the group of 73 countries. Comparison and discussion of methods and results from the three sources have been published elsewhere.⁵ If a country had already introduced the vaccine in 2015, then the mortality for the most recent prevaccination year was used, using WHO–UNICEF joint estimates of national immunisation coverage to determine the most recent prevaccine year.²⁸ In absence of vaccination, we assumed that rotavirus mortality would decrease at the same rate as all-cause mortality for children younger than 5 years of age. Rotavirus age distributions were based on a systematic review and statistical analysis of over 90 hospital datasets.²⁹ We assumed that 20% of severe rotavirus gastroenteritis cases would require a hospital admission and further reduced this proportion to account for those without access to hospital, using coverage of the first dose of diphtheria–tetanus–pertussis vaccine (DTP1) as a proxy for access to care. This method generated rates of rotavirus gastroenteritis hospitalisations that were consistent with prevaccination rates previously reported (around 350 per 100 000 per year, among children younger than 5 years).30 , 31 , 32 (link), 33 (link), 34 , 35 (link), 36 (link) We assumed that 100% of severe rotavirus gastroenteritis cases and 10% of non-severe cases would require a clinic visit, and again used DTP1 coverage to adjust for access to care. DALY weights were taken from the 2013 Global Burden of Disease study,³⁷ (link) using values reported for moderate diarrhoea as a proxy of non-severe rotavirus gastroenteritis and for severe diarrhoea as a proxy of severe rotavirus gastroenteritis. We assumed a duration of illness of 4 days for non-severe rotavirus gastroenteritis and 6 days for severe rotavirus gastroenteritis cases and explored longer and shorter durations in probabilistic analysis.³⁸ Input values and ranges for DALY weights and duration of illness are available in the appendix.