Table 1. Summary of Logistic Regression Analysis Examining Joint Impact of Size and Rurality on Colocation of Primary Care Physicians with Behavioral Health Providers

Size and Location	Rurality Only	Rurality and Size	Rurality-Size Interaction
Urban	Reference	Reference	Reference
Large rural	0.552 (0.533–0.571)^*	0.758 (0.730–0.787)^*	0.921 (0.843–1.006)
Small rural	0.430 (0.410–0.452)^*	0.790 (0.750–0.832)^*	0.589 (0.515–0.672)^*
Isolated rural	0.439 (0.400–0.482)^*	1.063 (0.961–1.174)	0.807 (0.652–0.999)^{^†}
Frontier	0.537 (0.489–0.590)^*	1.289 (1.166–1.425)^*	1.118 (0.899–1.391)
ln(size)		2.676 (2.650–2.703)^*	2.678 (2.650–2.706)^*
Large rural^*ln(size)			0.907 (0.871–0.944)^*
Small rural^*ln(size)			1.185 (1.105–1.269)^*
Isolated rural^*ln(size)			1.214 (1.066–1.381)^*
Frontier rural^*ln(size)			1.106 (0.966–1.265)

Data are odds ratios (95% confidence intervals). Data are from the National Plan and Provider Enumeration System 2010 and the American Medical Association Masterfile 2010. The analysis is based on 207,955 primary care physicians in direct patient care. Ln(size) is the natural log of the number of primary care physicians located at the same location (street address). Colocation of primary care physicians and behavioral health providers also is based on the same, more precise location. The outcome variable for the three models is colocation of primary care physician with behavioral health providers.
↵* Significant at 1%.
↵† Significant at 5%.