Congestion Spillovers in Multi-Stage Public Services: Evidence from England's Emergency Care System
Abstract
In multi-stage public service systems, congestion at one stage can propagate to others, but credible evidence on the magnitude of such spillovers is scarce. We study the emergency healthcare system in England, where hospital discharge delays create upstream congestion that traps ambulances at hospital doors and slows emergency response in the community. We exploit a £700 million discharge fund introduced during the winter of 2022–2023 that targeted hospital patient flow without directly intervening in ambulance services. Using regression discontinuity in time and difference-in-discontinuities design around the structural break in discharge rates, we find that the policy increased return of spontaneous circulation for cardiac arrest patients by 3.8 percentage points and 30-day survival by 2.4 percentage points. The mechanism operated through accelerated patient throughput which reduced ambulance handover delays and freed vehicles for emergency dispatch. With a back-of-the-envelope calculation, we show that the survival gains for cardiac arrest patients alone offset 11% of the programme's cost, given that the cardiac arrest patients represent just 0.6% of all hospital admissions. Our findings shed light on the importance of accounting for cross-stage spillovers when evaluating public service interventions, and suggest that targeted investments to relieve bottlenecks can yield large system-wide benefits in emergency care.