Retrospective observational study of emergency admission, readmission and the 'weekend effect'

Abstract

Objectives: Excess mortality following weekend hospital admission has been observed but not explained. As readmissions have greater age, comorbidity and social deprivation, outcomes following emergency index admission and readmission were examined for temporal and demographic associations to confirm whether weekend readmissions contribute towards excess mortality. Design: A retrospective observational study. Individual patient Hospital Episode Statistics were linked and 2 categories created: index admissions (not within 60 days of discharge from an emergency hospitalisation) and readmissions (within 60 days of discharge from an emergency hospitalisation). Logistic regression examined associations between admission category, weekend and weekday mortality, age, gender, season, comorbidity and social deprivation. Setting: A single acute National Health Service (NHS) trust serving a population of 550 000 via 3 emergency departments. Participants: Emergency admissions between 1 January 2010 and 31 March 2015. Outcome measure: All-cause 30-day mortality. Results: Over 5 years there were 128 966 index admissions (74.7% weekday/25.3% weekend) and 20 030 readmissions (74.9% weekday/25.1% weekend). Adjusted 30-day death rates for weekday/ weekend admissions were 6.93%/7.04% for index cases and 12.26%/13.27% for readmissions. Weekend readmissions had a higher mortality risk relative to weekday readmissions (OR 1.10 (95% CI 1.01 to 1.20)) without differences in comorbidity or deprivation. Weekend index admissions did not have a significantly increased mortality risk (OR 1.04 (95% CI 0.98 to 1.11)) but deaths which did occur were associated with lower deprivation (OR 1.24 (95% CI 1.11 to 1.38)) and an absence of comorbidities (OR 1.17 (1.02 to 1.34)). Conclusions: Associations with emergency hospitalisation were not identical for index admissions and readmissions. Further research is needed to confirm what factors are responsible for the 'weekend effect'.