TABLE 2

Heckman Selection Model Predicting Use of Any Services and Level of Services Use

Any Services Use, Unstandardized β (SE)Level of Services Use, Unstandardized β (SE)
Regression intercept1.83 (0.03)*132 (2.53)*
Age0.01 (0)−0.93 (0.03)*
Male sex−0.10 (0.01)*−5.80 (0.50)*
Race (reference: white)
 Black−0.17 (0.02)*−6.12 (1.62)*
 A/PI0.19 (0.04)*−19.8 (3.51)*
 Hispanic−0.02 (0.03)−16.0 (3.31)*
 AIAN0.03 (0.08)−9.46 (7.95)
Part D subsidy (reference: no subsidy)
 Partial subsidy0.03 (0.02)65.4 (1.35)*
 Full subsidy0.22 (0.01)*15.2 (0.70)*
Rural location0.07 (0.01)*−37.0 (0.79)*
Diabetes and combinations of MCCs (reference: diabetes)
 Diabetes + CHF0.05 (0.01)*31.3 (1.25)*
 Diabetes + cancer0.11 (1.01)*20.8 (1.43)*
 Diabetes + COPD0.13 (1.01)*23.3 (1.20)*
 Diabetes + stroke0.09 (0.01)*30.6 (1.78)*
 Diabetes + CKD0.21 (2.01)*30.6 (1.26)*
 Diabetes + CHF/COPD0.17 (1.01)*62.8 (1.32)*
 Diabetes + CHF/CKD0.25 (2.01)*85.9 (1.38)*
 Diabetes + CHF/COPD/CKD0.42 (4.01)*152 (1.21)*
 Diabetes + CHF/COPD/stroke/CKD0.55 (5.01)*195 (1.56)*
 Diabetes + all other combinations of MCCs0.39 (3.01)*105 (0.77)*
Interactions
 Black ×COPD0.12 (0.03)9.35 (3.56)
 Black ×stroke−0.12 (0.04)§0.85 (5.40)
 Black × CHF/COPD0.11 (0.03)4.22 (3.53)
 Black × CHF/CKD−0.02 (0.03)8.74 (3.48)§
 Hispanic × CHF/CKD0.01 (0.09)19.8 (8.76)§
 Black × CHF/COPD/CKD0.02 (0.03)29.1 (3.09)*
 Black × CHF/COPD/stroke/CKD0.10 (0.04)§86.5 (3.53)*
 A/PI × CHF/COPD/stroke/CKD0.01 (0.24)96.9 (14.0)*
 Hispanic × CHF/COPD/stroke/CKD0.01 (0.14)51.8 (10.3)*
 Black × all other combinations of MCCs0.11 (0.01)*51.7 (1.95)*
 Hispanic × all other combinations of MCCs0.004 (0.05)20.0 (4.69)*
 AIAN × all other combinations of MCCs0.04 (0.12)−24.1 (10.9)§
  • * P <0.001.

  • Only significant interactions are included; readers are welcome to contact the corresponding author for the complete table.

  • P <0.01.

  • § P <0.05.