Table 2.

Coefficients of multiple logistic regression models with and without covariates.

Model and term		Coefficient estimates (beta)	SE	P value	Odds ratio and 95% CI
Model 1: without covariates
	Intercept	−.950a	0.226	<.001	0.387 (0.246-0.609)b
	Year	.233	0.101	.03	1.263 (1.030-1.547)
	Year2	−.022	0.011	.04	0.978 (0.957-0.999)
	PPRc	−.462	0.097	<.001	0.630 (0.518-0.766)
	Year×PPR	−.060	0.025	.02	0.941 (0.896-0.989)
Model 2: with covariatesd
	Intercept	−1.763	0.58	.004	0.172 (0.053-0.551)
	Year	.253	0.104	.02	1.288 (1.044-1.588)
	Year2	−.024	0.011	.04	0.977 (0.955-0.999)
	PPR	−.286	0.113	.01	0.751 (0.598-0.943)
	Year×PPR	−.068	0.026	.01	0.934 (0.887-0.984)

^aThe estimate for the intercept is the baseline log-odds when year is in 2011 and PPR is 0.

^bThe odds ratio and 95% CI for intercept are the baseline odds and 95% CI when the year is in 2011 and patient-physician relationship is 0.

^cPPR: patient-physician relationship.

^dCovariates include gender, education, race, urbanity, age group, occupation, census division, born in the United States, general health status, provider maintains electronic medical record, depression, trust doctors, ever had cancer, and frequency to visit providers.