Table 5.

Hierarchical logistic regression estimates of childhood maltreatment predicting early age of onset of cannabis dependence at 21-year, Brisbane, Australia.

Childhood maltreatment	Category	Unadjusted OR (95%CI)	Adjusted OR (95%CI)
		Model 1	Model 2	Model 3	Model 4
Any maltreatment	No	1	1	1
Any maltreatmenta,b	Yes	4.21 (1.83–9.69)***	3.72 (1.59–8.75)**	3.72 (1.58–8.73)**	–
Sexual abuse	Yes	–	–	–	–
Physical abuse	Yes	5.59 (1.89–16.54)**	4.86 (1.62–14.61)**	4.83 (1.60–14.55)**	5.09 (1.69–15.39)**
Emotional abuse	Yes	3.61 (1.07–12.19)*	3.16 (0.92–10.81)	3.14 (0.92–10.77)	3.39 (0.99–11.68)
Neglect	Yes	5.14 (1.49–17.69)***	4.54 (1.29–15.90)*	4.56 (1.30–15.96)*	4.92 (1.40–17.25)*

^aAny childhood maltreatment included one or more combination of neglect, sexual, physical or emotional abuse.

^bModel did not adjust for other forms of childhood maltreatment.

Model 1: Unadjusted child maltreatment and cannabis dependence.

Model 2: Model 1 + gender at birth, maternal age at pregnancy, maternal smoking at pregnancy and 6 months postpartum and family poverty over the first 5 years of age.

Model 3: Adjusted for Models 1 + 2 + educational level and marital status at 21 years.

Model 4: Adjusted for Models 1 + 2 + 3 + any substantiated childhood maltreatment excluding the variable of interest to avoid the effect of overlapping maltreatment.

^*p <0.05;

^**p <0.01;

^****p <0.0001.