Table 4.
Summary of regression results.
| Severe CUD (n = 147) | Depression (n = 274) | Anxiety (n = 274) | Psychotic-like symptoms (n = 273) | |
|---|---|---|---|---|
| User-group | NA | No | No | Yes (b = 7.121, p = 0.001) |
| Age-group | Yes (OR = 4.462, p < 0.001) | Yes (b = 3.766, p < 0.001) | Yes (b = 4.627, p < 0.011) | Yes (b = 3.130, p = 0.001) |
| User-group X age-group | NA | No | No | No |
| Adjusted user-group | NA | No | No | Yes (b = 6.004, p = 0.014) |
| Adjusted age-group | Yes (OR = 3.474, p = 0.004) | Yes (b = 3.915, p < 0.001) | Yes (b = 4.528, p < 0.001) | Yes (b = 5.509, p = 0.015) |
Do the exposure variables (user-group and age-group, and their interaction) significantly predict the outcome variables? Severe CUD models were run only in users (n = 147, adjusted models n = 143). Bold text highlights if the user-group or age-group main effects or interaction were significant for the four outcome variables.
Depression and anxiety n = 274, adjusted models n = 268. Psychotic-like symptoms n = 273, adjusted model n = 267. Adjusted terms are from models including pre-defined covariates: gender, SES, RT-18, daily smoking, AUDIT and other drug use. The best models never included the interaction term; hence, there are no adjusted interaction terms.
AUDIT: alcohol use disorders identification test; b: unstandardised beta; CUD: cannabis use disorder; NA: not applicable; OR: odds ratio; RT-18: Risk-Taking 18.