Healthcare Delivery, Costs and Quality: Evidence from Emergency Departments

It is well understood by medical practitioners that healthcare delivery systems are far from optimal. Numerous examples abound of individuals who receive inappropriate care. Furthermore, many health economists believe that growth in healthcare costs may be driven in part by the provision of costly care with little marginal clinical benefit - leading many observers to suggest that certain areas of medicine may have reached the flat part of the health-production curve.

Behind all care decisions is a medical decision-maker, typically a physician, who is helping make underlying care management decisions for patients that influence both resource utilization and medical outcomes. Given this context, it is useful to understand the extent to which there is variation in the practice management styles across physicians, the causes for these variations, and the relation between physician practice management style, quality of care and healthcare costs. This study seeks to evaluate the impact of practice style by different emergency department (ED) physicians on health outcomes and healthcare costs and to understand the mechanisms behind any variations in outcomes. Emergency departments, along with intensive care units, are the most common sites for misdiagnosis, given the acute and time-sensitive nature of the diagnostic and therapeutic process. EDs are a prominent place of treatment for a number of widely-prevalent diseases such as ischemic heart disease and cerebrovascular disease, both of which are of significant importance to an aging population, with high rates of ED revisit or hospitalization for these conditions.

We seek to identify the impact of ED physician management style by making use of a unique dataset and quasi-experiment. The starting point of our analysis is an initial ED visit. We have data on the universe of patients who visit an ED on the Island of Montreal, Quebec, Canada, during the period April 1, 2006 to December 31, 2006. For each patient, we observe the diagnoses and procedures performed at the initial visit as well as the ED physician of Initial treatment. Importantly, our data links the visit with all future encounters of the patient with the health system during the 90 days following the initial visit. Thus, we observe the extent of future hospitalizations, as well as outpatient, office and ED visits. At EDs on the Island of Montreal, ED physicians typically see patients with a full range of diagnoses and levels of complication. In most EDs, doctors rotate across shifts between simple cases and difficult cases, following an initial triage of patients. In one ED, patients are assigned to doctors based purely on who is first available. This procedure leads to a quasi-random assignment of patients conditional on treatment at a particular ED. We exploit the quasi-random assignment of physicians: our study asks the question of how the initial assignment of ED physician affects the future pattern of health outcomes and healthcare costs. This then allows us to understand the extent and causes of the variation in attributes across ED physicians.