Winter burden of hospital bed closures due to acute gastroenteritis in England, 2010–11 to 2015–16: a population-based observational study

Abstract

BackgroundWinter pressure on hospital beds is a recurring public health concern. We aimed to estimate and compare the extent of hospital bed closures due to gastroenteritis during the past six winters in England. MethodsWe obtained all available daily records of occupied and unoccupied hospital beds closed because of diarrhoea and vomiting for all acute Trusts in England for 2010–11 to 2015–16 from National Health Service (NHS) England's winter situation reports. Because recording lengths differed in all six winters, we focused on an overlapping range of dates (Nov 30 to Feb 20—ie, 83 days or 11·9 weeks). About one-third of values were missing non-randomly (ie, all weekends and bank holidays), for which we imputed numbers at provider-level through last-observation carried forward and next-observation carried backward. We considered the lowest and highest value imputed as best and worst case estimates. We report numbers of all general and acute beds available in England and of all beds closed; the total number of beds closed per winter and the proportion unoccupied; and Pearson's correlation coefficient for occupied and unoccupied beds closed. FindingsTotal hospital capacity was a median 8·30 million beds (IQR 8·28–8·35 million) to 8·36 million beds (8·35–8·39 million) with the lowest and highest value imputed, respectively, corresponding to 100 000 (99 500–101 000) and 101 000 (99 900–102 000) beds per day. The median number of beds closed because of diarrhoea and vomiting was 88 300 (70 900–123 000) to 113 000 (88 300–151 000) with the lowest and highest value imputed, with totals for each of the six winters of 98 500 to 123 000 (2010–11), 135 000 to 168 000 (2011–12), 131 000 to 161 000 (2012–13), 68 600 to 83 800 (2013–14), 78 100 to 102 000 (2014–15), and 37 800 to 50 100 (2015–16). Respective proportions of beds closed unoccupied were 0·21 and 0·22, 0·19 and 0·20, 0·18 and 0·20, 0·19 and 0·21, 0·18 and 0·18, and 0·19 and 0·19. The number of occupied and unoccupied beds closed per day was strongly positively correlated (r=0·91 and r=0·89). InterpretationBed closures were equivalent to all general and acute NHS beds in England being unavailable for a median of 0·88 to 1·12 days with the lowest and highest value imputed. 20% were lost when unoccupied. The peak bed closures in 2011–12 coincided with a novel norovirus strain and the decline after 2012–13 with rotavirus vaccine introduction in July, 2013. FundingThis study originates from FGS's doctoral research that is jointly supported by a PhD studentship from the London School of Hygiene & Tropical Medicine and Public Health England.