Table 7: Odds ratios from binary logistic regression predicting patient satisfaction.

Unadjusted Age moderated Adjusted Patient-centered interaction 1.76** 1.38* 1.27 Patient age over 65 0.47* 0.76 Patient-centered × age over 65 2.02* 1.41 Patient characteristics  Gender—male 0.71  Race—white 1.69†  Education—less than high school 0.24*  Education—high school degree 0.65†  Education—some college 0.68  Education—college degree 1.33  Income—less than $20,000 1.13  Income—$20–$29,999 0.89  Income—$30,000–$39,999 0.97  Income—$40,000–$49,999 1.12  SF12—physical health 0.98  SF12—mental health 1.06** Physician characteristics  Male 0.61†  White 1.46† Years of practice 0.99 Encounter characteristics  First visit 0.50†  Number of visits 1.12

Note: Coefficients are odds ratios. Standard errors are corrected using Huber-White sandwich matrix estimator that does not assume independence of cases within clusters. †P <  .10, *P <  .05, **P <  .01, ***P <  .001.