Table 2. Predictors of practitioner behavior in relation to asking about patient smoking status.
| Univariate model | Multivariate model | ||
|---|---|---|---|
| Variable | p value | p value | Adjusted OR (95% CI) |
| Own smoking status | 0.383 | ||
| Taken dietary supplements in past year | 0.017* | 0.090 | |
| Own annual expenditure on dietary supplements | 0.282 | ||
| Gender | 0.171 | ||
| Age | 0.034* | 0.016* | 1.30 (1.05 to 1.61) |
| Therapeutic endorsement status | 0.230 | ||
| Years in practice | 0.116 | ||
| Practice location | 0.632 | ||
* indicates a statistically significant difference (p < 0.05).
Only those variables found to be statistically significant by multiple univariate modeling were included in the subsequent multivariate modeling. Smoking status was collapsed to become dichotomous, i.e., ever or never.