TABLE 2.
Model | Absolute Change in CD4 Percentage From Baseline (95% CI)
|
|||||
---|---|---|---|---|---|---|
<0–6 mo of cART | <6–12 mo of cART | <18–24 mo of cART | <30–36 mo of cART | <42–48 mo of cART | <54–60 mo of cART | |
1. Weighted, adjusted† (baseline confounders) | 6.8 (4.7–8.9) | 8.6 (7.0–10.2) | 12.7 (10.7–14.7) | 13.4 (10.9–15.9) | 15.9 (12.8–19.0) | 20.5 (16.1–24.9) |
2. Unweighted, unadjusted (no confounders) | 0.3 (−1.5–2.2) | 1.2 (−0.1–2.6) | 4.0 (2.3–5.6) | 3.4 (1.2–5.6) | 3.6 (0.7–6.5) | 3.7 (−0.5–7.8) |
3. Unweighted, adjusted (baseline confounders) | 4.2 (2.2–6.1) | 5.7 (4.1–7.3) | 9.9 (7.9–11.9) | 10.8 (8.2–13.4) | 12.6 (9.6–15.7) | 16.3 (12.0–20.5) |
4. Unweighted, adjusted (baseline and time-varying confounders) | 8.4 (6.7–10.2) | 8.4 (7.3–9.4) | 5.0 (3.8–6.1) | 3.6 (2.2–5.1) | 4.2 (5.8–4.2) | 4.8 (2.6–6.9) |
All estimates are derived from repeated measures linear models, fit with generalized estimating equations, that include time modeled as a restricted cubic spline with 4 knots.
MSM: Weighting appropriately accounts for time-varying confounders affected by prior exposure.