Screening and brief interventions for hazardous and harmful alcohol use among patients with active tuberculosis attending primary public care clinics in South Africa: results from a cluster randomized controlled trial

Table 2 Alcohol-related outcome measures at baseline, 3-month and 6-month follow-up and TB treatment outcome at 6-month follow-up

Variables	Time	Control	Intervention	OR* (95% CI)	P-value	ICC (SE)
				Adjusted for cluster
AUDIT total (M,SD)	Baseline	14.2 (6.1)	16.5 (6.5)	0.91 (0.78-1.07)	0.264	0.13 (0.04)
	3 months	4.0 (5.9)	5.0 (6.1)
	6 months	3.6 (6.2)	2.4 (4.8)
AUDIT (7–40) High risk or alcohol dependence	Baseline	455 (100)	741 (100)	0.70 (0.41-1.19)	0.186	0.11 (0.03)
	3 months	37 (21.0)	139 (34.8)
	6 months	57 (21.2)	98 (16.8)
AUDIT (7–19) High risk (N,%)	Baseline	367 (80.7)	507 (68.4)	0.64 (0.43-0.97)	0.035	0.07 (0.02)
	3 months	33 (18.8)	126 (31.5)
	6 months	45 (16.7)	91 (15.6)
AUDIT (20–40) Alcohol dependence (N,%)	Baseline	88 (19.3)	233 (31.4)	1.39 (0.75-2.56)	0.296	0.09 (0.03)
	3 months	4 (2.3)	13 (3.2)
	6 months	12 (4.5)	7 (1.2)
Heavy episodic drinking¹ (weekly+) (N,%)	Baseline	131 (29.0)	331 (45.2)	0.96 (0.46-2.02)	0.921	0.22 (0.06)
	3 months	11 (9.6)	23 (10.4)
	6 months	13 (11.6)	16 (10.6)
Daily or almost daily tobacco use (N,%)	Baseline	171 (39.0)	278 (40.8)	1.12 (0.67-1.89)	0.662	0.09 (0.03)
Daily or almost daily tobacco use (N,%)	6 months	71 (31.8)	93 (18.6)
TB treatment outcomes (N,%) ²	6 months	195 (53.6)	289 (45.9)
Cure		76 (20.9)	135 (21.5)
Complete		22 (6.0)	29 (4.6)
Failure		65 (17.9)	155 (24.6)
Default		6 (1.6)	21 (3.3)
Died		12 (3.2)	44 (6.5)
Transfer out
TB treatment cure or completion	6 months	271 (74.5)	424 (67.4)	0.93 (0.46-1.88)	0.840	0.15 (0.5)

¹For men 5 or more and for women 4 or more drinks on one occasion.
² TB treatment outcomes defined as from WHO classification [35] was not complete (N = 1049; 88%) due to various reasons including missing codebook of fieldworker lost, the incorrect recording of names, names could not be found in clinic register, and clinic register lost.
OR Odds Ratio, ICC Intracluster correlation coefficient, SE Standard Error.
*Odds ratios from regression models were adjusted for age, sex, and baseline alcohol use disorders identification test score.

ISSN: 1471-2458