Burden, duration and costs of hospital bed closures due to acute gastroenteritis in England per winter, 2010/11–2015/16

Abstract

SummaryBackgroundBed closures due to acute gastroenteritis put hospitals under pressure each winter. In England, the National Health Service (NHS) has monitored the winter situation for all acute trusts since 2010/11.AimTo estimate the burden, duration and costs of hospital bed closures due to acute gastroenteritis in winter.MethodsA retrospective analysis of routinely collected time-series data of bed closures due to diarrhoea and vomiting was conducted for the winters 2010/11 to 2015/16. Two key issues were addressed by imputing non-randomly missing values at provider level, and filtering observations to a range of dates recorded in all six winters. The lowest and highest values imputed were taken to represent the best- and worst-case scenarios. Bed-days were costed using NHS reference costs, and potential staff absence costs were based on previous studies.FindingsIn the best-to-worst case, a median of 88,000–113,000 beds were closed due to gastroenteritis each winter. Of these, 19.6–20.4% were unoccupied. On average, 80% of providers were affected, and had closed beds for a median of 15–21 days each winter. Hospital costs of closed beds were £5.7–£7.5 million, which increased to £6.9–£10.0 million when including staff absence costs due to illness.ConclusionsThe median number of hospital beds closed due to acute gastroenteritis per winter was equivalent to all general and acute hospital beds in England being unavailable for a median of 0.88–1.12 days. Costs for hospitals are high but vary with closures each winter.

Revisiting the winter burden of acute gastroenteritis on hospital beds in England: change in data collection supports analytical method for previously missing values

“…Furthermore, an infectious cause for the majority of these bed closures due to acute gastroenteritis -as we have speculated in the original paper [1] -has been confirmed in our follow-up study that regressed the SitReps data against enteric pathogen counts in laboratory-based surveillance, finding that norovirus explained more than 94% of the observed variation in bed-days closed due to acute gastroenteritis [3] . It seems thus justified for NHS England to label the records with "bed-days closed due to diarrhoea and vomiting/norovirus-like symptoms" [2] .…”

“…The same follow-up study estimated a median total of 290,000 bed-days being unavailable (occupied and unoccupied) due to norovirus-associated gastroenteritis in acute care and community hospitals in England on an annual basis [3] , with financial costs for the NHS of £107.6 million per year and economic costs of up to £297.7 million; an estimated 57,800 alternative patients could have been admitted were the beds made available to them. The findings described in this letter here suggest the actual burden of norovirus is likely higher than this estimated burden, as we based the unoccupied bed-days closed conservatively on the best-case scenario results with the lowest imputations [1] .…”

“…In a study published previously in this journal [1] , we estimated that all-cause acute gastroenteritis resulted in a median 88,000-113,000 bed closures in hospitals in England each winter, placing a financial burden on hospitals of at least £5.7-£7.5 million (or £6.9-£10.0 million when including staff absence costs). Our analysis used routinely-collected data of the number of occupied and unoccupied hospital beds closed (i.e., unavailable to other patients) due to diarrhoea and vomiting during the winters 2010/11 to 2015/16.…”

“…We previously used two imputation scenarios to estimate the number of bed-days lost to acute gastroenteritis on weekends and public holidays, therefore addressing the issue of the fact one third of days were missing [1] . The imputation scenarios were based on the number of bed-days recorded as closed immediately before and after the weekends and public holidays, using the lowest and highest numbers as best-to-worst case scenarios.…”

“…The imputation scenarios were based on the number of bed-days recorded as closed immediately before and after the weekends and public holidays, using the lowest and highest numbers as best-to-worst case scenarios. There are a number of limitations and assumptions underlying this approach, as discussed in the original paper [1] , which led us to conclude: "With the information available, it is not possible to determine where exactly the number of beds closed lies on the range of the lowest to highest imputations; this inaccuracy is unavoidable though due to the values missing." [1] .…”