Background Significant variation in length of inpatient stay (LOS) exists across hospitals in the UK.¹ It is not clear whether demographic, clinical and/or process factors, in particular socio-economic position (SEP) can explain this. Studies examining the association between LOS and SEP have been inconclusive, possibly due to differences in the definition of prolonged LOS.^2–7

Study design A retrospective audit of 2889 children aged less than 16 years admitted to two District General Hospitals in London from 1st April 2009 to 30^th March 2010, each with different models of paediatric accident and emergency (A&E) service delivery. Hospital A A&E was paediatric and Hospital B A&E led.

Exposure measures Demographic, clinical characteristics and process factors were examined (Tables 1 & 2). SEP was measured by quartile of Income deprivation affecting children index (IDACI).

• Download figure
• Open in new tab
• Download powerpoint

Abstract G478(P) Table 1

LOS: Length of stay; IDACI: Income deprivation affecting children index; A&E: Accident and emergency

• Download figure
• Open in new tab
• Download powerpoint

Abstract G478(P) Table 2

LOS: Length of stay; IDACI: Income deprivation affecting children index; A&E: Accident and Emergency

Methods Prolonged LOS was defined as greater than or equal to the mean length of stay (1.8 days). Multivariable logistic and linear regression analyses were performed. Sensitivity analysis was carried out using two other definitions of length of stay.

Results A fifth of children stayed less than 24 hours and almost two fifths more than the mean of 1.8 days. Adjusted for age, sex and admission hospital, being in the 3^rd most deprived IDACI quartile was associated with increased odds of prolonged LOS (Model 1: Odds Ratio (OR) 1.26, 95% Confidence interval (CI) 1.01, 1.56) (Table 3). This association was attenuated by additional adjustment for clinical and process factors (Table 3, Models 3 & 4). In contrast, adjusting for demographic, clinical and process factors strengthened the association between admission to Hospital B and prolonged LOS (Model 4: OR 1.41, 95% CI 1.16, 1.71). These associations were not consistent using different definitions of prolonged LOS.

• Download figure
• Open in new tab
• Download powerpoint

Abstract G478(P) Table 3

OR: Odds ratio; CI: confidence interval; A&E: Accident and emergency; ref: reference category; IDACI: Income deprivation affecting children index; LOS: Length of stay; β co-e: β co-efficient

Conclusion The association between SEP and prolonged LOS is weak and inconsistent and depends on the definition used and the consideration of other potential confounding factors. These findings suggest that service configuration may account for some of the variation in LOS between hospitals in the UK; alternatively they could be due to residual confounding.

Model 1: Adjusted for age (groups), gender admission hospital and quartile of IDACI; as defined in Tables 1 and 2

Model 2: additionally adjusted for ethnicity (group) – as defined in Tables 1 and 2.

Model 3: Additionally adjusted for final diagnosis, medical or surgical specialty and number of hospital admissions during the study period

Model 4: Additionally adjusted for source of admission, weekend or week day admission, season of admission and time of admission – as defined in Tables 1 and 2