Table 3.
Model§ | OR | 95% CI | P value |
---|---|---|---|
Use of NRT | |||
Race/Ethnicity | |||
Caucasian | 1.0 | Reference | |
African American | 0.76 | 0.63, 0.91 | <.01 |
Latino | 0.76 | 0.54, 1.06 | .11 |
Asian American | 0.98 | 0.60-1.60 | .95 |
Gender | |||
Female | 1.0 | Reference | |
Male | 0.73 | 0.65-0.82 | <.001 |
Ever used other tobacco products | |||
No | 1.0 | Reference | |
Yes | 1.19 | 1.07, 1.35 | <.01 |
Time to first cigarette, when smoking the most | |||
> 30 minutes after waking | 1.0 | Reference | |
≤30 minutes after waking | 2.13 | 1.91, 2.38 | <.001 |
Cigarettes per day, when smoking the most** | 1.76 | 1.66, 1.86 | <.001 |
Age | 1.00 | 0.99-1.01 | .34 |
Education†† | 1.07 | 1.01, 1.13 | .02 |
NRT, nicotine replacement therapy; OR, odds ratio; CI, confidence interval
Logistic regression models controlled for site. All possible interactions for the variables selected for the multivariate model were tested, but were not statistically significant at the 0.05 level. A forced-entry simultaneous model of the significant covariates with race/ethnicity showed similar results.
Cigarettes per day treated as an ordinal variable.
Education treated as an ordinal variable.