Document Type


Original Publication Date


Journal/Book/Conference Title






DOI of Original Publication



Originally published at

Date of Submission

November 2015



Financial and demographic pressures in US require an understanding of the most efficient distribution of physicians to maximize population-level health benefits. Prior work has assumed a constant negative relationship between physician supply and mortality outcomes throughout the US and has not addressed regional variation.


In this ecological analysis, geographically weighted regression was used to identify spatially varying relationships between local urologist density and prostate cancer mortality at the county level. Data from 1,492 counties in 30 eastern and southern states from 2006–2010 were analyzed.


The ordinary least squares (OLS) regression found that, on average, increasing urologist density by 1 urologist per 100,000 people resulted in an expected decrease in prostate cancer mortality of -0.499 deaths per 100,000 men (95% CI -0.709 to -0.289, p-value < 0.001), or a 1.5% decrease. Geographic weighted regression demonstrated that the addition of one urologist per 100,000 people in counties in the southern Mississippi River states of Arkansas, Mississippi, and Louisiana, as well as parts of Illinois, Indiana, and Wisconsin is associated with decrease of 0.411 to 0.916 in prostate cancer mortality per 100,000 men (1.6–3.6%). In contrast, the urologist density was not significantly associated with the prostate state mortality in the new England region.


The strength of association between urologist density and prostate cancer mortality varied regionally. Those areas with the highest potential for effects could be targeted for increasing the supply of urologists, as it associated with the largest predicted improvement in prostate cancer mortality.


Yao, N., Foltz, S. M., & Odisho, A. Y., et al. Geographic Analysis of Urologist Density and Prostate Cancer Mortality in the United States. PLoS ONE, 10, e0131578. Copyright © 2015 Yao et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.

Is Part Of

VCU Healthcare Policy and Research Publications

S1_Fig.pdf (265 kB)
Residuals from the OLS model.

S1_File.pdf (516 kB)
GWR and FDR methods.

S1_Table.pdf (64 kB)
Descriptive Statistics of Outcome and Explanatory Variables (n = 1492 counties).

S2_Fig.pdf (418 kB)
Local Indicators of Spatial Association of residuals from the OLS Model.

S3_Fig.pdf (194 kB)
Counties with complete prostate cancer incidence/mortality data.

S4_Fig.pdf (242 kB)
Urologist density by county in the study region: 2006–2010.

S5_Fig.pdf (215 kB)
GWR collinearity diagnostics.