Univariable and Multivariable Poisson Regression Models of HIV Incidence��Presented Separately for Males and Females

Seroconversions and person years may not add up across different attributes because of missing dataa Adjusted for age and attributes significant at p �� 0.1 (Wald test)b Exposure between baseline and follow-up (approximately 3 y)c Not significant but shown to demonstrate the direction of effectd Exposure in the 1 y prior to follow-upCI, confidence interval; PYAR, person years at risk