Table 3:

Results of Multivariable Negative Binomial Regression Models With Total Number of Symptoms as the Outcome (N = 169)

	Total Number of Symptoms (Negative Binomial Regression)
	Model 1		Model 2		Model 3
	a.IRR^a	95% CI	a.IRR^a	95% CI	a.IRR^a	95% CI
Tobacco Use	1.043	(0.975-1.116)	1.026	(0.954-1.103)	1.032	(0.939-1.133)
Alcohol Use			1.034	(0.934-1.144)	1.024	(0.915-1.147)
Stimulant Use			1.163^**	(1.105-1.223)	1.113^†	(0.980-1.264)
Demographic Factors
Female (vs. Male)					0.977	(0.729-1.309)
Age					1.007	(0.998-1.016)
≥ College (vs. < College)					1.566	(0.781-3.139)
Income
Just enough (vs. not enough)					0.645^**	(0.586-0.709)
More than enough (vs. not enough)					0.601^**	(0.468-0.772)
Clinical Factors
Years with HIV					0.998^**	(0.997-0.999)
AIDS Diagnosis					1.018	(0.928-1.116)
CD4 Levels					0.943^**	(0.925-0.961)
On HIV Medication (vs. not on med)					0.798	(0.363-1.754)
Study waves	5.416^**	(5.117-5.731)	5.036^**	(4.81-5.273)	4.640^**	(4.151-5.187)
Constant	4.140^**	(3.748-4.573)	4.096^**	(3.533-4.75)	5.334^**	(3.103-9.168)

Model 1: Simple regression with control of study site; Model 2: Regression with control of alcohol and drug use and study site; Model 3: Regression with full control;

^a.

IRR: model-adjusted Incident Rate Ratio;

^†

< 0.1;

< 0.05;

^**

< 0.01