Background Hospital admission at weekends has been found to have an adverse impact on patient outcomes for both emergency and elective conditions. However, due to data limitations, studies to date have only assessed day – not time – of admission. Consequently, weekends have been defined as 00.00 am Saturday to 11.59 pm Sunday, which does not accurately reflect working patterns. In addition, it has been impossible to examine whether or not there is a similar impact on patients outcomes for night-time admission, when service levels are also reduced. The objective of our study was to estimate the association between adverse patient outcomes and out-of-hours admissions by defining weekend and night appropriately using an enhanced dataset, including hour and minute of admission.

Methods Our sample contains 244,639 emergency spells extracted from a large teaching hospital in England between April 2004 and March 2014. Patient outcomes were measured by 30-day, 7-day and in-hospital mortalities. We conducted Probit regressions to estimate the impact of two key predictors on the risk of mortality, adjusting for a range of patient demographics and complexity. The first predictor was admission at weekends, defined as 7.00 pm Friday to 6.59 am Monday. The second predictor was night-time admission, defined as 7.00 pm to 6.59 am. We then categorised out-of-hours admissions in three groups: weekday-night, weekend-day and weekend-night, to compare with admissions within normal operational hours (weekday-day).

Results Age, gender, ethnicity, deprivation, and complexity were significant predictors for mortality. After adjustment, weekend admission had an independent and significant association with adverse outcomes: risk increased by 0.4 percentage points for 30-day mortality, and 0.3 percentage points for 7-day and in-hospital mortalities. Examining out-of-hours admission by day and night suggested that the greatest risk of mortality was associated with admission at weekends on night shifts: 30-day mortality increased by 0.6 percentage points compared with being admitted during normal operational hours, whereas 7-day mortality increased by 0.4 percentage points and in-hospital mortality increased by 0.5 percentage points. Weekend admissions during day time had a similar effect on mortality to weekday admissions during night time. The lowest risk of mortality was for admission during normal operational hours (weekday day-time).

Conclusion When out-of-hours periods are properly defined, there appears to be an increased risk of mortality for patients admitted at weekends and during the night-time. These effects are additive, so that the greatest risk of mortality occurs in patients admitted during the night at weekends.