Table 4. Odd ratios for each condition improving at second follow-up under different assumptions about the causal structure.

Condition improves	Analysis	Odd ratios (95% CIs)
Past-year unhealthy alcohol use	Fig 1A *	2.13 (1.03, 4.39)
	Fig 1B **	2.85 (1.43, 5.68)
	Fig 1C ***	2.18 (1.05, 4.53)
Past-year cannabis use	Fig 1A *	1.67 (0.94, 2.97)
	Fig 1B **	2.17 (1.27, 3.72)
	Fig 1C ***	1.72 (0.98, 3.02)
Past-year illicit opioid use	Fig 1A *	1.09 (0.57, 2.07)
	Fig 1B **	1.28 (0.70, 2.33)
	Fig 1C ***	1.03 (0.54, 1.94)
Past-year cocaine use	Fig 1A *	2.12 (1.11, 4.03)
	Fig 1B **	2.95 (1.61, 5.41)
	Fig 1C ***	1.89 (0.99, 3.62)
Current depressive symptoms	Fig 1A *	0.74 (0.41, 1.34)
	Fig 1B **	0.87 (0.50, 1.51)
	Fig 1C ***	0.72 (0.40, 1.31)
Current anxiety symptoms	Fig 1A *	0.98 (0.60, 1.60)
	Fig 1B **	0.87 (0.50, 1.51)
	Fig 1C ***	0.72 (0.40, 1.31)
Current moderate or severe pain symptoms	Fig 1A *	1.24 (0.78, 1.95)
	Fig 1B **	1.30 (0.29, 2.01)
	Fig 1C ***	1.30 (0.82, 2.05)

*Primary weighted analysis reported in Table 2. Adjusted for HIV status, race, age, education, income, baseline conditions (AUDIT score, depressive symptoms, anxiety symptoms, moderate or severe pain symptoms, current smoking, and past-year cannabis, cocaine, other stimulant, and illicit opioid use), and some time-varying covariates measured at first follow-up (past-year unhealthy alcohol, cannabis, cocaine, other stimulant, and illicit opioid use).

**All time-varying covariates measured at first follow-up (past-year cannabis, cocaine, other stimulant, and illicit opioid use, unhealthy alcohol use, depressive, anxiety, and moderate or severe pain symptoms) excluded from the model for the weights.

***All time-varying covariates measured at first follow-up included in the model for the weights.