Availability of primary care and avoidable attendance at English emergency departments: A regression analysis.

Attendances at emergency departments (EDs) by patients who could have been treated in primary care increase waiting times and costs in EDs and may reduce quality of care. This study examines whether the probability that a patient's ED attendance is avoidable is associated with their characteristics and the quality, staffing, and availability of their general practice, particularly its extended hours provision. We estimate ED attendance level linear probability and logistic regressions using data on 10.16 M attendances at 144 major EDs by patients aged 16 or over from 6668 English practices. We use two definitions of avoidable ED attendance: the NHS definition (non-urgent) and a new wider definition (clinically inappropriate). 9.3 % of attendances were avoidable according to the NHS definition and 21.8 % with our definition. The probability of avoidable attendance was lower for older, female patients, those living in more socioeconomically deprived or sparsely populated areas, or those closer to their practice than to the ED attended. Attendances from practices where a higher proportion of patients get same-day GP appointments, or were aware of early morning extended hours, were less likely to be avoidable. The probability that an ED attendance was clinically inappropriate was about 0.5 % smaller during weekends or evenings when the practice had extended hours but was not associated with the overall provision of extended hours by the practice.