Reducing Non-urgent Visits and Emergency Department Congestion : Pricing or Improving Perceptions?

A major reason behind crowding in emergency departments (EDs) is non-urgent patients' visits to ED. In this paper we study how patients' imperfect perception of their urgency and self-interested choice affect non-urgent ED visits and social cost. We then investigate how perception-improvement measures and economic incentives influence patients' choice, social cost, and profit of the healthcare network. We model patients' problem of choosing between an expensive/congested ED and general practitioners (GPs), who refer urgent patients to the ED, as a queueing game. We analytically characterize equilibrium and socially optimum patient flows, and find that improving patients' perceptions may increase non-urgent ED visits and social cost under certain equilibria. We identify sufficient conditions on ED/GP fees, service costs and waiting time externalities under which this happens. We further show that perception-improvement guarantees to reduce the social cost if socially optimum flows are induced at equilibrium. Motivated by a real case, we design a novel differential pricing "GP-r" mechanism (incorporating GP-referral discount) to induce optimum patient flows at equilibrium, and analyze its properties/benefits. We derive all results for general ED waiting times that are increasing and convex in arrival rates, which makes our results applicable to a wide variety of ED queueing systems. Our results imply that perception-improvement measures, implemented alone, may not always prove beneficial, unless they have high accuracy. This limitation can be overcome by deploying them together with economic incentives that can induce optimum patient flows. In terms of economic incentives, the proposed GP-r mechanism provides a benefit of lowering the ED fee required to induce optimum flows, compared to a traditional pricing mechanism that only raises ED fee. This benefit can be achieved without extra budget investment in the healthcare network, and is most prominent when status quo ED fee is not too large compared to GP fee, but the resources/congestion at ED are much costlier compared to GP