Table 2. Logistic regression analysis of current drug use and the current and prior status of unhealthy alcohol use, smoking and depression.

Four logistic regression models were fitted for four outcome variables (current stimulant use, current marijuana use, current prescription opioid use, and current heroin use), respectively, with the status of the corresponding outcome variable at baseline, and the status of unhealthy alcohol, smoking, and depression at the current and preceding surveys, adjusting for number of years from enrollment and HIV status

	Current Stimulant Use		Current Marijuana Use		Current Opioid Use		Current Heroin Use
	Model 1		Model 2		Model 3		Model 4
Parameter	ORa	p value	ORa	p	ORa	p	ORa	p
Number of years from enrollment	0.98	0.54	0.99	0.49	1.01	0.24	0.93	<.0001
HIV status	2.48	<.0001	1.39	<.0001	0.87	0.05	1.14	0.31
Stimulant use at baseline	18.34	<.0001	-	-	-	-	-	-
Marijuana use at baseline	-	-	15.57	<.0001	-	-	-	-
Opioid use at baseline	-	-	-	-	2.99	<.0001	-	-
Current unhealthy alcohol use	1.68	0.01	1.59	<.0001	0.82	0.02	1.22	0.09
Current unhealthy alcohol use at preceding surveyb	1.28	0.15	0.99	0.86	0.88	0.07	0.90	0.32
Current smoking	1.16	0.55	1.20	0.02	1.01	0.89	1.56	<.01
Current smoking at preceding surveyc	1.21	0.42	1.09	0.24	1.05	0.55	2.02	<.001
Current depression	2.06	<.0001	1.35	<.0001	1.55	<.0001	1.60	<.0001
Current depression at preceding surveyd	1.35	0.09	0.93	0.27	1.33	<.001	1.30	0.01

‘-’ indicates that predictor was not included in the model;

^*Baseline data for heroin and prescription opioids are combined.

^aAdjusted for HIV status and number of years from enrollment

^bControlling for current unhealthy alcohol use

^cControlling for current smoking

^dControlling for current depression