Table 3.

Baseline Practice and Practice Member Variables Associated with Greater Self-Management Support Activities, from 36 Primary Care Practices in Colorado and Northern California; Data Collected from 12/2013 to 9/2015

VariableUnadjusted Bivariate Associations Coef (P value)Bivariate Associations, Adjusted for Clustering Coef (SE)Multivariable Model 1* Coef (SE)Multivariable Model 2* Coef (SE)Multivariable Model 3* Coef (SE)
Intercept50.47 (2.78)30.67 (5.59)51.28 (5.61)
Specialty (FM vs multispecialty/other)0.0799 (.0330)3.45 (2.54)
Rural location0.1197 (.0014)6.23 (2.54)
Number of diabetic patients0.0732 (.0508)0.0043 (0.0034)
Care manager0.1615 (<.0001)6.42 (2.33)5.22 (1.68)5.24 (1.99)5.63 (1.89)
Percent Medicaid0.1014 (.0068)0.1234 (0.0599)
Percent uninsured0.1262 (.0007)0.1479 (0.0640)
Level of PCMH0.1925 (<.0001)3.73 (0.89)2.46 (0.81)3.20 (0.90)2.74 (0.87)
Underserved designation0.1665 (<.0001)9.50 (3.01)5.81 (2.62)6.66 (2.90)5.97 (2.81)
PCA change0.3426 (<.0001)0.4305 (0.0385)0.3500 (0.0455)
PCA chaos−0.2858 (<.0001)−0.3420 (0.0401)−0.1440 (0.0459)−0.2632 (0.0471)
PCA work0.2264 (<.0001)0.2943 (0.0318)0.1429 (0.0480)
  • Coef, Coefficient; FM, family medicine; PCMH, Patient-Centered Medical Home; PCA, Practice Culture Assessment; SE, standard error.

  • * Model 1 includes basic practice characteristics, model 2 added one combination of PCA scale (Change culture and chaos), and model 3 tested the other combination of PCA scales (Work culture and chaos). Change culture and work culture were tested in separate models to reduce multicollinearity; SE, standard error.

  • P < .05.

  • P < .01.