Who uses emergency departments inappropriately and when - a national cross-sectional study using a monitoring data system

Abstract

Background: Increasing pressures on emergency departments (ED) are straining services and creating inefficiencies in service delivery worldwide. A potentially avoidable pressure is inappropriate attendances (IA); typically low urgency, self-referred patients better managed by other services. This study examines demographics and temporal trends associated with IA to help inform measures to address them.

Methods: Using a national ED dataset, a cross-sectional examination of ED attendances in England from April 2011 to March 2012 (n = 15,056,095) was conducted. IA were defined as patients who were self-referred; were not attending a follow-up; received no investigation and either no treatment or 'guidance/advice only'; and were discharged with either no follow-up or follow-up with primary care. Small, nationally representative areas were used to assign each attendance to a residential measure of deprivation. Multivariate analysis was used to predict relationships between IA, demographics (age, gender, deprivation) and temporal factors (day, month, hour, bank holiday, Christmas period).

Results: Overall, 11.7% of attendances were categorized as inappropriate. IA peaked in early childhood (adjusted odds ratio (AOR) = 1.53 for both one and two year olds), and was elevated throughout late-teens and young adulthood, with odds reducing steadily from age 27 (reference category, age 40). Both IA and appropriate attendances (AA) were most frequent in the most deprived populations. However, relative to AA, those living in the least deprived areas had the highest odds of IA (AOR = 0.89 in most deprived quintile). Odds of IA were also higher for males (AOR = 0.95 in females). Both AA and IA were highest on Mondays, whilst weekends, bank holidays and the period between 8 am and 4 pm saw more IA relative to AA.

Conclusions: Prevention of IA would be best targeted at parents of young children and at older youths/young adults, and during weekends and bank holidays. Service provision focusing on access to primary care and EDs serving the most deprived communities would have the most benefit. Improvements in coverage and data quality of the national ED dataset, and the addition of an appropriateness field, would make this dataset an effective monitoring tool to evaluate interventions addressing this issue.

Figure 1

Adjusted odds ratios (reference age 40) and total attendances by year of age. Ages above 85 not included due to small totals. Confidence intervals were too small to be displayed. Logistic regression model controlled for age, gender and deprivation. Primary Y axis is a logarithmic scale (base 2). *Adjusted odds ratio.

Figure 2

Adjusted odds ratios for hour of arrival (reference hour is midnight). Confidence intervals were too small to be displayed. Logistic regression model controlled for hour of arrival, month, weekday, bank holiday and Christmas period.