Substance Use Outcomes Among Sexual and Gender Minority Individuals Living with HIV Following Residential Substance Use Treatment in Washington, DC

Figures & data

Table 1. Baseline demographic and clinical characteristics for the overall sample and by SGM status.

Characteristic	Total sample (N = 60)	SGM (n = 21)	Non-SGM (n = 39)	p
Age, M (SD)	45.38 (7.81)	42.38 (8.32)	47.00 (7.15)	.02
% Female (n)^¶	53.3 (32)	52.4 (11)	53.9 (21)	.91
% African American (n)	96.7 (58)	90.5 (19)	100 (39)	.11
% Employed (n)	91.7 (55)	85.7 (18)	94.9 (37)	.33
% Married or cohabitating (n)	10.0 (6)	9.5 (2)	10.3 (4)	1.00
Years since HIV diagnosis, M (SD)	1.67 (7.68)^a	12.36 (6.82)^b	9.66 (8.10)^c	.24
CD4 count, M (SD)	420 (254)^d	422 (243)^e	419 (263)^f	.97
Prior residential treatment episodes, M (SD)	3.43 (3.11)^d	4.70 (3.71)^e	2.68 (2.45)^f	.01
SIP-AD, M (SD)	27.19 (13.07)^g	31.48 (12.47)	24.69 (12.92)^h	.06

Note. ^¶Sex assigned at birth (male, female, other). ^an = 38. ^bn = 18. ^cn = 30. ^dn = 54. ^en = 20. ^fn = 34. ^gn = 57. ^hn = 36. SIP-AD = Short Inventory of Problems – Alcohol and Drugs; SGM = sexual and gender minorities.

Figure 1. Survival function (i.e., those not using substances) by SGM status from residential discharge to 12-month follow-up.

Figure 2. Hazard function (i.e., risk of substance use) by SGM status from residential discharge to 12-month follow-up.

Table 2. Discrete time survival analysis from residential discharge to 12-months follow-up.

Parameter	Log odds estimate [95% CI]	SE	p value	Hazard ratio^¶
Intercept	−3.26 [−4.51, −2.01]	.64	<.001	–
SGM group	.54 [−.13, 1.21]	.34	.11	71.8
Residential discharge (reference)	-
Post-treatment	.90 [−.55, 2.35]	.74	.22	145.5
1-month follow-up	1.24 [−.17, 2.65]	.72	.08	244.6
3-month follow-up	1.21 [−.34, 2.76]	.79	.12	235.5
6-month follow-up	1.17 [−.67, 3.01]	.94	.21	221.9
12-month follow-up	.25 [−1.97, 2.46]	1.13	.82	27.8
Assessment period length	.01 [−.007, .03]	.01	.24	1.04
Assessment period length squared	<.001 [−.00005, .00003]	<.001	.71	–

Note. ^¶Interpreted as the percent increase in the risk of reusing substances relative to the comparator. SGM = sexual and gender minorities.

Table 3. Log odds model of TLFB frequency over time from residential discharge to 12-months follow-up.

Predictor	SGM only model Log odds (SE)	Final multivariable model Log odds (SE)
Intercept	−7.94 (1.23)***	−7.25 (1.21)***
Time (per week)	.03 (.03)	.03 (.03)
SGM	3.62 (1.53)*	2.46 (1.63)
Age		.12 (.08)
Prior residential treatment		.42 (.22)^†
Time × SGM	−.02 (.03)	.001 (.04)
Time × Age		−.002 (.002)
Time × Prior residential treatment		−.009 (.005)^†

Note. ^†p < .10. *p < .05. **p < .01. ***p < .001. TLFB = Timeline Follow Back; SGM = sexual and gender minorities.

Figure 3. Mean score on substance related problems using the SIP by SGM status from study entry to 12-months follow-up.

Table 4. Quadratic model of the SIP total (substance-related problems) over time from baseline (study intake) to 12-month follow-up.

Predictor	SGM only model B (SE)	Final multivariable model B (SE)
Intercept	22.64 (2.16)***	23.60 (2.34)***
Time (per week)	−.81 (.18)***	−.76 (.20)***
SGM	5.76 (3.48)	3.37 (4.09)
Age		.07 (.24)
Prior residential treatment		1.36 (.62)*
Time × SGM	−.24 (.26)	−.24 (.31)
Time × Age		−.003 (.02)
Time × Prior residential treatment		−.05 (.05)
Time × Time (quadratic effect)	.008 (.003)**	.007 (.003)*
Time × Time × SGM	.003 (.004)	.003 (.005)
Time × Time × Age		< −.001 (<.001)
Time × Time × Prior residential treatment		.001 (.001)

Note. ^†p < .10. *p < .05. **p < .01. ***p < .001. B = unstandardized beta; SGM = sexual and gender minorities.