Modeling Physician Distribution Uncertainty in Three Common Health Workforce Data

Context: There is widespread concern about the adequacy and distribution of our nation’s health workforce. For planners and policymakers, the correct identification of physicians’ practice locations is critical, yet tremendous uncertainty endures in their use of existing national workforce datasets.

Purpose: To explore spatial uncertainty in commonly used physician workforce data and understand the probable patterns of physician distribution.

Design: We triangulated three physician workforce data, 2014 American Medical Association Physician Masterfile (AMA), 2014 National Provider Identifier (NPI) Database, and 2014 Georgia State Physician Licensure (LIC) data in two counties in Atlanta metro areas. These data are also linked to land use classification with parcel data.

Population Studied: Patient care physicians

Principal Findings: More than 25% Georgia licensed physicians listed addresses outside the state. 92% of LIC data are matched to AMA data, 89% of AMA data are matched to NPI data, and 90% of LIC data are matched to both AMA and NPI data. 13.3% of NPI physician addresses, 15% of AMA masterfile physician addresses, and 24.9% of state licensure physician addresses are found in residential neighborhood.

Conclusions:  There are significant practice location uncertainty in physician workforce data. NPI data appears to have the best practice location and state licensure data has the worst practice location information. This study calls for a unified national data system for physician practice locations which is needed for better health workforce planning and deployment that is critical for ensuring public health. More research are needed to improve practice location accuracy.