TABLE 3.
Summary of Sequential Logistic Regression Model Predicting Adherence (n = 1809)
| Variable | Step 1
|
Step 2
|
||||||
|---|---|---|---|---|---|---|---|---|
| OR | SE | 95% CI | P | OR | SE | 95% CI | P | |
| Step 1: demographics | ||||||||
| Race/ethnicity | Omnibus: χ2 (21) = 119.09 | <0.001 | Omnibus: χ2 (21) = 122.10 | <0.001 | ||||
| African American vs. White | 0.62 | 0.05 | 0.53 to 0.71 | <0.001 | 0.60 | 0.05 | 0.52 to 0.70 | <0.001 |
| Latino vs. White | 0.85 | 0.13 | 0.62 to 1.16 | 0.30 | 0.84 | 0.14 | 0.61 to 1.15 | 0.27 |
| Age in yrs | 1.01 | 0.004 | 1.00 to 1.02 | 0.01 | 1.01 | 0.004 | 1.00 to 1.02 | 0.01 |
| Male gender | 1.05 | 0.08 | 0.91 to 1.21 | 0.53 | 1.04 | 0.08 | 0.90 to 1.20 | 0.62 |
| Annual household income >$10k | 1.09 | 0.07 | 0.96 to 1.24 | 0.19 | 1.06 | 0.07 | 0.93 to 1.20 | 0.33 |
| Education | Omnibus: χ2 (2) = 6.97 | 0.03 | Omnibus: χ2 (2) = 8.08 | 0.02 | ||||
| HS diploma vs. <HS diploma | 1.22 | 0.09 | 1.05 to 1.42 | 0.01 | 1.24 | 0.10 | 1.06 to 1.44 | 0.01 |
| College degree vs. <HS diploma | 1.23 | 0.15 | 0.97 to 1.56 | 0.09 | 1.29 | 0.16 | 1.01 to 1.64 | 0.04 |
| Step 2: substance use and depression | ||||||||
| Substance use | — | — | — | — | Omnibus: χ2 (11) = 39.43 | <0.001 | ||
| Yes vs. no | — | — | — | — | 0.74 | 0.08 | 0.60 to 0.92 | 0.01 |
| Missing vs. no | — | — | — | — | 0.57 | 0.17 | 0.32 to 1.02 | 0.06 |
| Depression | — | — | — | — | Omnibus: χ2(13) = 15.92 | 0.25 | ||
| Mild to moderate vs. normal | — | — | — | — | 0.87 | 0.08 | 0.72 to 1.05 | 0.15 |
| Severe vs. normal | — | — | — | — | 0.76 | 0.12 | 0.56 to 1.03 | 0.07 |
| Missing vs. normal | — | — | — | — | 0.79 | 0.09 | 0.62 to 0.99 | 0.04 |
The model controlled for the main effect of site and the interaction of site with all variables, except for education. All parameter estimates are weighted by site. χ2 (36) = 172.32, P < 0.0001, for step 1; χ2 (24) = 61.94, P < 0.0001, for step 2.