Table 3. Estimates of the effect of antidepressant medication treatment on viral suppression, using conventional statistical adjustment vs. inverse probability of treatment and censoring (IPTC) weighting (N=158).

The naïve (unweighted) estimates do not have a causal interpretation and are shown here for comparison purposes only. The estimates from the marginal structural models have the following interpretation: if, contrary to fact, participants are exposed to antidepressant medication at or before the prior visit, then their average odds of achieving viral suppression would be the odds ratio given.

	Odds ratio (95% CI)	P-value
Unweighted, crude	1.55 (1.03–2.31)	0.034
Unweighted, adjusted	1.58 (1.07–2.31)	0.02
Weighted	2.03 (1.15–3.58)	0.025