A Simulation-Based Approach for Strategic Evaluation of Resource Allocation for Improving Spatial Access to Radiation Oncology

Abstract

Access to radiation therapy is essential to oncologic care. Prior studies have examined geographic distribution of radiation oncologists and centers but with methodologies that fail to account for patient movement across artificially constructed boundaries. This analysis utilizes an approach agnostic to geographic boundaries as well as a mathematical metric to assess the equality of access to care and quantify the impact additional resource allocation would have on a geographic area. Government databases were utilized to collect information on supply of radiation oncologists and population over 65 years old as a surrogate for demand. Addresses were geocoded to create coordinates for spatial analysis. US Census Centers of Population files were used to assign population data to spatial coordinates. Isochrone lines for driving time around population centers were generated using an internet road maps programming interface. Enhanced two-step floating catchment area (E2SFCA) analysis was performed with catchment area defined as a 120-minute drive from the population centers. A distance decay factor was applied. Spatial access ratios (SPARs), defined as the ratio of a given county's access to care against the mean across all counties, were calculated for each of the 3143 US counties, where a value <1 represents underserved and >1 represents overserved areas. A Gini coefficient was calculated to create a value representing spatial access equality (SPAE) for the current geographic distribution of supply and demand across the US. The incremental impact on SPAE was determined by iteratively simulating the addition of a single radiation oncologist in each of the 3143 US counties, and subsequently used to quantifiably rank areas by magnitude and direction of impact. Olmsted County, MN and the surrounding counties represent 8 of the 10 most overserved in the US (SPAR: 4.03 - 5.84). There are 1713 underserved counties with SPARs <1 and 130 counties with a SPAR of 0. Thomas County, KS would have greatest benefit on SPAE from the addition of a single radiation oncologist as it would introduce access to numerous neighboring underserved counties. Conversely, introduction of a single radiation oncologist negatively impacted SPAE in 173 of the previously underserved counties by shifting SPAR to overserved without significant positive effect on neighboring counties due to having a small population and isolated location. E2SFCA and SPAE provide a method for evaluating spatial access and a mathematical framework for assessing areas with the greatest potential benefit from new resource allocation. In some instances, resource allocation to a previously underserved area paradoxically has a negative impact on equal access to care. Better solutions such as travel/lodging programs and appropriate hypofractionation would do more to improve issues of access inequality for these areas.