Using Social Groups to Locate Areas with High Emergency Department Attendance, Subsequent Inpatient Admission and Need for Critical Care

Aims: To study the role of social group and deprivation on the proportion of patients admitted via the emergency department who then progress to the critical care unit.

Study Design: Two large data bases were compared using the output area code of the patient’s home address.

Place and Duration of Study: Patients attending the emergency department (ED) and admitted into the general surgical and medical critical care unit (CCU) at the King’s College University Hospital, London, during a three-year period (2013/14 to 2015/16).

Methodology: The output area of the patient’s home address was used to link ED attendances with CCU admissions. Each output area has an associated social group and deprivation score. Various ratios were calculated such as the proportion admitted in each social group, and a method based on Poisson statistics was used to measure statistical significance. Each output area contains around 300 persons of roughly similar social and demographic characteristics. The Cartesian co-ordinates (distance east and north of the UK reference point as the X- and Y-axis respectively) for each output area have been used to map ED attendances without resort to GIS software.

Results: Output area social grouping, via the London Output Area Classification (LOAC) and deprivation (Index of Multiple Deprivation) are shown to affect the attendance rates at the ED, the proportion admitted from the ED, and conversion rates for patients admitted via the ED into an adult general surgical and medical CCU. The overall conversion rate from an admitted inpatient to the CCU was 6.6%, however, higher than average conversion rates of 14.3% were associated with the ‘Settled Asians’ Super Group, while lower conversion rates of 3.3% were associated the with the ‘Ageing City Fringe’ Super Group and other ‘Affluent Suburbs’ Super Groups. Social group rather than deprivation per se appeared to delineate high and low conversion rates.

Conclusion: Small areas characterised by particular social groups (and relative deprivation) were identified having either high/low rates of attendance at the ED, high/low conversion to an inpatient from an ED attendance, and high/low conversion into CCU admission. This will enable Primary Care Organisations to target admission avoidance and/or appropriate end-of-life care to those social groups yielding the greatest benefit.