The Association of American Medical Colleges' Local Area Physician Workforce Modeling Project

Abstract

Physician workforce planning must address multiple concerns such as having sufficient numbers and adequate geographic distribution of physicians and pressures for physicians to adapt to new models of care and payment. Though there are national workforce planning tools, planning tools for local areas have been scarce. This article describes a dynamic simulation model developed as a pilot project to support physician workforce planning in 2 metropolitan areas, Cleveland and Albuquerque (February 2014-June 2016). This model serves as a prototype for planning tools that could be used by medical educators and local health systems to project the effect of different policies on physician supply and demand. System dynamics and group model building approaches were used to develop the model with the participation of local stakeholders to create the model's causal structure. The model included determinants of the demand for primary and specialty care for the local population and projected the effects of births and deaths, aging, level of chronic illness present, and migration on demand. Physician supply was disaggregated by primary versus specialty care, age, sex, and work setting and projected based on completions of local residency programs, physician migration in and out of the area, and retirements. Feedback relationships between supply and demand (e.g., adequacy of care affecting the distribution of chronic illnesses, demand for care influencing in- and out-migration of physicians) were also included and had important effects on the results produced by the model. Scenarios were simulated that projected increased demand for care (e.g., through expanded insurance coverage) and increased supply (e.g., through practice incentives to encourage in-migration) and a combination of these. An expanded advanced practice registered nurse and physician assistant capacity scenario was also simulated. In Albuquerque, the combination scenario yielded the greatest increases in local physician supply.