. Author manuscript; available in PMC: 2013 May 1.

Published in final edited form as: Addict Behav. 2012 Feb 7;37(5):651–656. doi: 10.1016/j.addbeh.2012.01.021

Table 3.

Diversion sources as a function of nonmedical use of prescription opioids and screening for substance use disorders based on CRAFFT^a

	Bought prescription opioids from any source		Multiple sources of diversion		Family member gave for free only

Nonmedical use and substance abuse status	%	AOR^b (95% CI)	%	AOR^b (95% CI)	%	AOR^b (95% CI)

Past-year nonmedical use (−CRAFFT)	2.3	Reference	10.2	Reference	27.9	Reference
Past-year nonmedical use (+CRAFFT)	28.8	14.6 (2.7 – 80.5)^***	34.6	5.2 (1.8 – 15.2)^**	7.7	0.2 (0.1, 0.7)^**
Chi-square (DF), p-value	21.01(1), p < 0.001		12.5(1), p < 0.001		8.2(1), p < 0.01

p < 0.05,

^**

p < 0.01,

^***

p < 0.001 based on logistic regression.

Positive (+) and negative (−) screen for substance abuse and dependence based on CRAFFT (car, relax, alone, forget, friends, trouble).

Odds ratios are adjusted for gender, race/ethnicity, school, and grade level (odds ratios for these variables are not shown).

Due to missing values and “rather not say” responses, sample sizes were 138 for the bought prescription opioids from any source model, 140 for the multiple sources of diversion model and 138 for the family members for free only model.