Estimated prevalence of T. vaginalis (TV) infection and odds ratios by gender and sociodemographic characteristics.

Estimates based on respondents who provided urine specimens for testing. Estimates are weighted to account for differing probabilities of selection and post-stratification adjustment to match Census (ACS) marginals for Baltimore, Maryland (see text). Base Ns are unweighted. Odds ratios (ORs) calculated using logistic regression. Gender-specific adjusted ORs (AOR) calculated from multivariable models controlling for race (Black; Non-Black), age (4 categories: 15–19, 20–24, 25–29, 30–35), and marital status (married vs. not married, including widowed, divorced, and separated). Black was omitted in calculation of adjusted ORs for males for education and employment since there were zero TV infections among non-Black respondents in the age range for these variables (20 to 35). P-values shown in column of prevalence estimates are for test of hypothesis that estimates of TV prevalence are independent across categories of the independent variable (e.g., marital status). Variance estimates and confidence intervals (CI) were calculated using statistical algorithms that take account of the complex sample design used in the MSSP surveys. P-values for ORs and adjusted ORs are for test of independence of Tv prevalence across categories of socio-demographic variables taking account of weighting, complex sample design, and covariates (for adjusted ORs).aPersons describing themselves as Hispanic are coded as Non-Black.bAges 20 to 35 only.cThe difference in infection prevalence among men and women is statistically significant (OR = 4.4, 95% CI = 2.4, 8.3, p