Table 3.

Using club drugs and alcohol abuse to predict poor adherence (log-binomial analyses, n=401)

		Univariate		Adjusted for significant background variable^b
	Row %	PR(95%CI)	P-value	RR(95% CI)	P-value
Club drug abuse
Use of any of the club drugs in the last six months
No	27.2	1.00		1.00
Yes	32.4	1.19 (0.68, 1.85)	0.485	1.19 (0.68, 1.83)	0.490
Use of methamphetamine in the last six months
No	27.0	1.00		1.00
Yes	35.5	1.31 (0.74, 2.04)	0.289	1.33 (0.75, 2.06)	0.256
Use of Ma Gu in the last six months
No	27.5	1.00		1.00
Yes	37.5	1.36 (0.40, 2.66)	0.503	1.26 (0.37, 2.43)	0.614
Use of Triazolam Tablets in the last six months
No	27.7	1.00		1.00
Yes	25.0	0.90 (0.06, 2.66)	0.906	0.95 (0.06, 2.82)	0.955
Use of ketamine in the last six months
No	27.8	1.00		1.00
Yes	20.0	0.72 (0.05, 2.30)	0.715	0.57 (0.04, 1.87)	0.538
Use of ecstasy in the last six months
No	27.7	1.00		1.00
Yes	25.0	0.90 (0.06, 2.66)	0.906	0.78 (0.05, 2.29)	0.778
Alcohol abuse
Overdrinking
No	25.9	1.00		1.00
Yes	34.1	1.32 (0.91, 1.83)	0.118	1.34 (0.93, 1.86)	0.092
Drinking frequency
Never	28.6	1.00		1.00
At least once	26.7	0.93 (0.68, 1.28)	0.676	0.95 (0.69, 1.31)	0.764
Drinks on a typical day when drinking
1–2 drinks	23.6	1.00		1.00
≥3 drinks	32.3	1.37 (0.83, 2.19)	0.201	1.35 (0.83, 2.16)	0.214
Frequency of having 6 or more drinks on 1 occasion when drinking
Never	21.3	1.00		1.00
At least once	38.9	1.82 (1.13, 2.89)	0.011	1.87 (1.16, 2.95)	0.018

Log-binomial regression models adjusting for potential confounders (transfer treatment), which predicted poor adherence significantly at end of the study in a multivariate model (stepwise). RR: relative risk; CI: confidence interval.