. Author manuscript; available in PMC: 2015 Apr 27.

Published in final edited form as: Alcohol Clin Exp Res. 2015 Feb 23;39(3):532–539. doi: 10.1111/acer.12638

Table 2.

Interrelations Among Anxiety Disorder Symptoms, Anxiety Sensitivity, and High-Consequence Drinking

Outcome variable	Predictor variable		Mediation RCADS subscale → CASI → RAPI
Outcome variable	RCADS subscale^a b (95% CIs)	CASI^b b (95% CIs)	Indirect effect b (95% CIs)	Remaining direct effect b (95% CIs)
Models using generalized anxiety disorder symptoms subscale
CASI	4.00 (3.28 to 4.72)^***	–		–
RAPI cutoff status	0.75 (0.41 to 1.09)^***	0.05 (0.01 to 0.09)^*	0.19 (0.01 to 0.38)^*	0.57 (0.19 to 0.95)^**
Models using panic disorder symptoms subscale
CASI	5.84 (4.93 to 6.76)^***	–	–	–
RAPI cutoff status	0.88 (0.48 to 1.28)^***	0.05 (0.003 to 0.09)^*	0.28 (0.02 to 0.58)^*	0.62 (0.15 to 1.08)^**
Models using social phobia symptoms subscale
CASI	4.50 (3.83 to 5.17)^***	–
RAPI cutoff status	0.47 (0.14 to 0.80)^**	0.07 (0.03 to 0.11)^**	0.31 (0.10 to 0.54)^**	0.17 (30.21 to 0.56)
Models using generalized anxiety disorder symptoms subscale
CASI	4.00 (3.28 to 4.72)^***	–		–
RAPI continuous	2.62 (1.41 to 3.81)^***	0.24 (0.10 to 0.39)^*	0.97 (0.31 to 1.79)^**	1.65 (0.32 to 2.98)^*
Models using panic disorder symptoms subscale
CASI	5.84 (4.93 to 6.76)^***	–	–	–
RAPI continuous	4.55 (2.99 to 6.12)^***	0.18 (0.03 to 0.34)^*	1.07 (0.16 to 2.14)^*	3.49 (1.69 to 5.28)^***
Models using social phobia symptoms subscale
CASI	4.50 (3.83 to 5.17)^***	–
RAPI continuous	1.65 (0.46 ot 2.84)^**	0.33 (0.17 to 0.48)^***	1.48 (0.71 to 2.41)^***	0.18 (31.19 to 1.54)

CASI, Childhood Anxiety Sensitivity Index; RAPI, Rutgers Alcohol Problem Index; RCADS, Revised Children’s Anxiety and Depression Scale.

Coefficients reflect main effects of respective RCADS subscale on CASI and the total effect (i.e., combined indirect and direct) of the RCADS subscale on RAPI status.

Coefficients for CASI predicting RAPI cutoff score vary by analysis because they are adjusting for variance explained by the RCADS subscale in each analysis (i.e., the RCADS subscale relevant for each analysis is included in the model with CASI predicting RAPI cutoff and its variance is thus accounted for). Note that we include the continuous RAPI scores as dependent variables as a secondary way of presenting the data. These data were highly skewed and thus are reported only to show robustness across methods of analysis and should be interpreted with caution.

p < 0.05,

^**

p < 0.01,

^***

p < 0.001.

Ns range from 522 to 534 due to different patterns of missing data across the RCADS subscales. All analyses included gender as a covariate.