Table 2.
Characteristics associated with incomplete retention among individuals contributing ≥1 HIV-lab between 2000 and 2008. Odds ratios (OR) from beta-binomial regression identify factors associated with increased (OR>1) or decreased (OR<1) likelihood of incomplete retention.
| Factor | Beta-binomial 2000–2004a | Beta-binomial 2005–2008b | ||
|---|---|---|---|---|
| OR | (95% CI) | OR | (95% CI) | |
| Female Sex | 0.82 | (0.70, 0.95) | 1.11 | (0.91, 1.36) |
| Age (per 10 y) | 0.78 | (0.74, 0.83) | 0.80 | (0.74, 0.86) |
| Non-Hispanic Black (vs. Non-Hispanic White) | 1.31 | (1.16, 1.49) | 1.36 | (1.15, 1.62) |
| IDU Risk (vs. non-IDU) | 1.68 | (1.49, 1.89) | 1.34 | (1.11, 1.61) |
| ART (at baseline) | 0.61 | (0.54, 0.68) | 0.73 | (0.63, 0.84) |
| Time Since Enrollment (per year) | 1.09 | (1.07, 1.11) | 1.22 | (1.16, 1.29) |
|
| ||||
| ρ | (95% CI) | ρ | (95% CI) | |
|
| ||||
| Intra-class Correlation Coefficient | 0.24 | (0.22, 0.26) | 0.20 | (0.17, 0.23) |
a
Model includes individuals entering care in 2000 contributing data 2000–2004, N=5,718
b
Model includes individuals entering care in 2005 contributing data 2005–2008, N=5,071
Time since enrollment: years under clinical observation in NA-ACCORD
Non-overlapping periods of 5 & 4 years used to eliminate artifacts of truncation in separate beta-binomial models
All models adjusted for contributing cohort and additional race/ethnicity categories of “Hispanic” and “Other/Unknown”