Table 2.

Multivariable prediction models for HIV acquisition among MSM developed in Public Health—Seattle & King County repeat HIV testers, 2001–2008.

Variable	Hazard Ratio	P value	β	β*10	Weight*
Full Model
Socio-demographic Characteristics
Non-white, non-Asian Pacific Islander race/ethnicity	1.48	0.096	0.39	3.9	4
< 40 years of age	1.88	0.050	0.63	6.3	6
Sexually Transmitted Infection
Diagnosis or history of bacterial sexually transmitted infection† at baseline	1.65	0.024	0.50	5.0	5
Substance Use
Methamphetamine or inhaled nitrites, prior 6 months	3.00	<0.001	1.10	11.0	11
Sexual Risk
≥ 10 male sex partners, prior year	1.47	0.096	0.39	3.9	4
Receptive non-concordant unprotected anal sex, prior year	1.09	0.711	0.09	0.9	1
Simple Model
Sexually Transmitted Infection
Diagnosis or history of bacterial sexually transmitted infection† at baseline	1.57	0.039	0.45	4.5	4
Substance Use
Methamphetamine or inhaled nitrites, prior 6 months	2.94	<0.001	1.07	10.7	11
Sexual Risk
≥ 10 male sex partners, prior year	1.33	0.207	0.27	2.9	3
Non-concordant unprotected anal sex, prior year	1.16	0.536	0.14	1.4	1

^*Weights are based on the coefficients (β) of the Cox proportional hazards regression. We calculated these weights by multiplying the model coefficient by 10 and rounding to the nearest whole integer. The weights rank the risk predictors in relative importance and dictate how one assigns integer point values for each risk predictor for a given individual. The assigned points are then summed to compute that individual’s risk score.

^†Gonorrhea, chlamydia, early syphilis (primary, secondary, early latent).