Table 4.
Reversion to smoking among switchers, stratified by duration of switching: risk perceptions and sociodemographic characteristics
| By Switch duration | Less than a year | A year or mored | |
|---|---|---|---|
|
Multivariable resultse aOR (95% CI) |
Multivariable resultse aOR (95% CI) |
||
| Time | Lineara | 0.89 (0.77 – 1.03) | 0.94 (0.70 – 1.27) |
| Risk perception | Less harmfulb | Reference | Reference |
| Equally/more harmful | 1.78 (1.04 – 3.06) | 1.30 (0.70 – 2.43) | |
| Race/ethnicity | NH White | Reference | Reference |
| NH Black | 0.59 (0.22 – 1.59) | 0.63 (0.19 – 2.09) | |
| Hispanic | 1.24 (0.75 – 2.04) | 2.14 (0.79 – 5.83) | |
| NH others | 1.31 (0.73 – 2.35) | 0.27 (0.05 – 1.62) | |
| Sex | Male | Reference | Reference |
| Female | 0.91 (0.65 – 1.27) | 0.87 (0.48 – 1.59) | |
| Age | 18 – 24 | Reference | Reference |
| 25 – 44 | 0.93 (0.54 – 1.58) | 0.25 (0.11 – 0.57) | |
| 45 – 54 | 0.80 (0.44 – 1.44) | 0.34 (0.11 – 1.03) | |
| 55 or older | 0.81 (0.41 – 1.59) | 0.23 (0.07 – 0.71) | |
| Marital statusc | Married | Reference | Reference |
| D/S/W | 0.80 (0.47 – 1.38) | 2.62 (1.08 – 6.38) | |
| Never married | 0.86 (0.57 – 1.31) | 1.48 (0.79 – 2.76) | |
| Household income | < $25 k | Reference | Reference |
| $25 k – $50 k | 0.86 (0.54 – 1.39) | 0.47 (0.19 – 1.18) | |
| > $50 k | 0.69 (0.45 – 1.04) | 0.80 (0.39 – 1.62) | |
| Educational attainment | HS/GED or less | Reference | Reference |
| Some college | 0.87 (0.57 – 1.34) | 2.12 (1.10 – 4.10) | |
| Bachelor or higher | 0.68 (0.37 – 1.27) | 1.02 (0.44 – 2.37) | |
Boldface represents statistically significant results
NH Non-Hispanic, HS High school, D/S/W Divorced/Separated/Widowed
aNo quadratic effects of time were found at either subgroup (OR 0.99, CI 0.84 – 1.16 among switchers with quit duration of less than a year; OR 1.09, CI 0.80 – 1.50 among switchers with quit duration of a year or more)
bFor reversion analysis, the “less harmful” level of risk perception was used as the referent category
cMarital status of Wave 1 has been extrapolated from the marital status of Wave 2 due to unavailability in the survey
dSwitching history of this subset is as follows (in years): Mean = 4.4; Median = 2.5; Interquartile range = 1.6 – 4.0
eThe subset with less than a year duration of switching consists of 982 observations from 771 unique participants. The subset of participants who have switched a year or more are based on 959 observations from 519 unique participants. Some participants are included in both subsets as their switching history increases over time. Multivariable results represent an analysis with all listed variables simultaneously in the model
The interaction term (risk perception*switch duration) was not statistically significant (P = .5229)