Table 2.
Multivariable Associations of High Patient Comfort*
| Correlates | Base Model Adjusted Bivariate OR (95% CI)  | 
P Value | Final Model OR (95% CI)  | 
P Value | 
|---|---|---|---|---|
| Patient variables | ||||
| Gender | ||||
| Male | 1.0 | .192 | 1.0 | .018 | 
| Female | 1.38 (0.85–2.25) | 2.05 (1.13–3.72) | ||
| Race | ||||
| White | 1.0 | 1.0 | ||
| African American | 1.40 (0.88–2.24) | .480 | 1.29 (0.67–2.42) | .735 | 
| Hispanic | 1.30 (0.53–3.19) | 1.25 (0.41–3.76) | ||
| Other | 2.03 (0.44–9.39) | 2.54 (0.36–17.96) | ||
| Education Less than high school diploma  | 
0.67 (0.37–1.24) | .202 | 0.45 (0.22–.93) | .031 | 
| Depression tertile | ||||
| 1 | 1.0 | .005 | 1.0 | .201 | 
| 2 | 0.45 (0.27–0.73) | 0.65 (0.39–1.07) | ||
| 3 | 0.51 (0.29–0.91) | 0.89 (0.49–1.61) | ||
| On antiretroviral therapy (Reference=no) | 2.15 (1.36–3.40) | .001 | 2.33 (1.18–4.6) | .015 | 
| Current problematic alcohol use (Reference=no) | 0.32 (0.15–0.66) | .002 | 0.33 (0.15–0.72) | .006 | 
| Current drug use (Reference=no) | 0.46 (0.28–0.77) | .003 | 0.33 (0.15–0.72) | .015 | 
| High communication Self-efficacy  | 
2.90 (1.72–4.89) | <.001 | 1.96 (1.08–3.54) | .025 | 
| PAM tertile | ||||
| 1 | 1.0 | <.001 | 1.0 | .037 | 
| 2 | 1.56 (0.95–2.57) | 1.55 (0.81–2.95) | ||
| 3 | 3.12 (1.84–5.32) | 2.66 (1.24–5.69) | ||
| Length of visit (Odds for a 10 minute increase)  | 
0.75 (0.61–.92) | .006 | 0.78 (0.64–0.99) | .015 | 
| Provider Variables | ||||
| Gender | ||||
| Male | 1.0 | .545 | 1.0 | .314 | 
| Female | 1.19 (.67–2.12) | 1.41 (.72–2.74) | ||
| Race | ||||
| White | 1.0 | .611 | 1.0 | .927 | 
| Non-white | 1.14 (0.69–1.90) | 1.03 (0.57–1.87) | ||
| Type of provider | ||||
| Physician | 1.0 | 0.70 | 1.0 | .087 | 
| Midlevel | 1.91 (0.95–3.84) | 2.17 (0.89–5.27) | ||
| Communication training | ||||
| No | 1.0 | 0.156 | 1.0 | .859 | 
| Yes | 1.49 (0.86–2.60) | 0.94 (0.50–1.78) | ||
Multivariable Model A reports the odds of high comfort for each independent variable while controlling for site and provider clustering. The Final Model includes all variables, as well as site and adjustment for provider clustering.