Population and patient factors affecting emergency department attendance in London: retrospective cohort analysis of linked primary and secondary care records

BACKGROUND: Population factors, including social deprivation and morbidity, predict the use of emergency departments (EDs). AIM: To link patient-level primary and secondary care data to determine whether the association between deprivation and ED attendance is explained by multimorbidity and other clinical factors in the GP record. DESIGN AND SETTING: Retrospective cohort study based in East London. METHOD: Primary care demographic, consultation, diagnostic, and clinical data were linked with ED attendance data. GP Patient Survey (GPPS) access questions were linked to practices. RESULTS: Adjusted multilevel analysis for adults showed a progressive rise in ED attendance with increasing numbers of long-term conditions (LTCs). Comparing two LTCs with no conditions, the odds ratio (OR) is 1.28 (95% confidence interval CI = 1.25 to 1.31); comparing four or more conditions with no conditions, the OR is 2.55 (95% CI = 2.44 to 2.66). Increasing annual GP consultations predicted ED attendance: comparing zero with more than two consultations, the OR is 2.44 (95% CI = 2.40 to 2.48). Smoking (OR 1.30, 95% CI = 1.28 to 1.32), being housebound (OR 2.01, 95% CI = 1.86 to 2.18), and age also predicted attendance. Patient-reported access scores from the GPPS were not a significant predictor. For children, younger age, male sex, white ethnicity, and higher GP consultation rates predicted attendance. CONCLUSION: Using patient-level data rather than practice-level data, the authors demonstrate that the burden of multimorbidity is the strongest clinical predictor of ED attendance, which is independently associated with social deprivation. Low use of the GP surgery is associated with low attendance at ED. Unlike other studies, the authors found that adult patient experience of GP access, reported at practice level, did not predict use.