. Author manuscript; available in PMC: 2016 Apr 30.

Published in final edited form as: Stat Med. 2015 Feb 4;34(9):1605–1620. doi: 10.1002/sim.6440

Table 2.

Results of ML Analysis of Main/Internal Validation Study Data on 904 Women at the 4^th HERS Visit (n_m=690, n_v=214).

Variable	β̂ (StdErr)	$\hat{O R}$ (95% CI)
Assuming dependent and differential misclassification^a (AIC=1771.5)
Trichomoniasis (+ vs. −)	0.64 (0.41)	1.90 (0.38, 3.42)
Age (Years)	−0.05 (0.02)	0.95 (0.92, 0.98)
Race (Black vs. Others)	0.79 (0.24)	2.19 (1.16, 3.23)
HIV Risk Cohort (IDU vs. Sexual)	0.28 (0.26)	1.32 (0.64, 2.00)
HIV Status (+ vs. −)	0.23 (0.27)	1.26 (0.58, 1.94)
Assuming independent and differential misclassification^b (AIC=1777.7)
Trichomoniasis (+ vs. −)	1.06 (0.38)	2.88 (0.73, 5.02)
Age (Years)	−0.05 (0.02)	0.95 (0.92, 0.98)
Race (Black vs. Others)	0.73 (0.24)	2.07 (1.09, 3.05)
HIV Risk Cohort (IDU vs. Sexual)	0.30 (0.26)	1.35 (0.65, 2.05)
HIV Status (+ vs. −)	0.21 (0.28)	1.23 (0.56, 1.90)
Assuming nondifferential misclassification^c (AIC=1809.8)
Trichomoniasis (+ vs. −)	1.38 (0.37)	3.99 (1.09, 6.90)
Age (Years)	−0.05 (0.02)	0.95 (0.92, 0.99)
Race (Black vs. Others)	0.73 (0.24)	2.08 (1.11, 3.04)
HIV Risk Cohort (IDU vs. Sexual)	0.85 (0.22)	2.35 (1.32, 3.37)
HIV Status (+ vs. −)	−0.14 (0.23)	0.87 (0.47, 1.26)

ML estimates of primary parameters are obtained by jointly modeling eqn.s (1) and (18)–(20).

WETTRICH is removed from eqn. (19) to indicate independence. The assumption is not supported by the data from the AIC value.

No covariates affect SE and SP of Y and X in eqn.s (19) and (20). The assumption is not supported by the data from the AIC value.