Table 3.

Predictors of Any Alcohol Consumption, the Number of Alcohol Drinks, and Having Any Binge Drinking Days in the Past Week

Variables		95% Confidence Interval	p
Logistic Regression for Any Alcohol Consumption in the Past Week1
	Adjusted Odds Ratio
Male gender	1.45	1.02–2.07	.04
Used any illicit drug in the past month	2.11	1.26–3.55	.005
Network size	.96	.93–.99	.01
Used any injected drug in the past month	2.54	1.42–4.54	.002
Poisson Regression for the Number of Alcohol Drinks in the Past Week2
	Incidence Rate Ratio
Male gender	2.38	1.63–3.48	<.0001
Used any illicit drug in the past month	2.66	1.79–3.95	<.0001
Network size	.97	.96–.98	<.0001
Social Provisions Scale (SPS)	.97	.95–.99	.007
Married	.48	.30–.76	.002
Logistic Regression for Having Any Binge Drinking Days in the Past Week1
	Adjusted Odds Ratio
Male gender	1.85	1.11–3.07	.02
Used any illicit drug in the past month	3.01	1.76–5.17	.0001
Network size	.96	.94–.98	.0001
Condomless intercourse with a non-main partner in the past 3 months	2.49	1.25–4.98	.01

¹All p-values were calculated using logistic regression fitted by GEE to account for clustering within social networks. Stepwise backward variable selection was used to find the final parsimonious model

²All p-values were calculated using multiple Poisson regression fitted by GEE to account for clustering within social networks. Stepwise backward variable selection was used to find the final parsimonious model