TABLE 2.

Influential analyses using random effects models according to outcomes of preventing alcohol return-to-drinking or promoting abstinence stratified by nonrandomized controlled trials and randomized controlled trials

Omitted study or studies	RR	95% CI	p	I 2 (%)
Alcohol return to drinking non-RCT influential analysis
Erim et al 33	0.45	0.23–0.87	0.026	72.9
Carrique et al 35	0.53	0.25–1.15	0.089	77.2
López-Pelayo et al 36	0.45	0.23–0.89	0.031	67.4
Bjornsson et al 38	0.52	0.23–1.17	0.092	75.4
Addolorato et al 39	0.55	0.25–1.20	0.107	73.4
Peeraphatdit et al 40	0.59	0.31–1.13	0.09	72.4
Attilia et al 41	0.58	0.28–1.19	0.107	73.2
Erim et al 33 + Lopez-Pelayo et al 36	0.37	0.25–0.56	0.003	0.0
Alcohol return-to-drinking RCT influential analysis
Proeschold-Bell et al 23	0.89	0.25–3.14	0.72	50.5
Weinrieb et al 25	0.94	0.19–4.57	0.89	59.1
North et al 26	0.89	0.18–4.31	0.78	52.6
Shen et al 27	1.20	0.83–1.74	0.17	0.0
Abstinence non-RCT influential analysis
Erim et al 33	1.76	0.59–5.26	0.20	59.9
Mehta et al 32	0.94	0.62–1.44	0.69	15.3
Andersen et al 34	1.24	0.49–3.11	0.52	57.5
Alexandre et al 37	1.75	0.59–5.20	0.20	62.7
Dieperink et al 42	1.24	0.48–3.17	0.52	57.7
Abstinence RCT influential analysis
Proeschold-Bell et al 23	1.24	0.93–1.6379	0.11	0.0
Stein et al 24	1.21	0.96–1.5259	0.08	0.0
Dieperink et al 28	1.15	0.85–1.5501	0.28	0.0
Edelman et al 29	1.14	0.84–1.5273	0.30	0.0
Dhanda et al 30	1.19	0.89–1.5900	0.18	0.0
DeMartini et al 31	1.09	0.84–1.4082	0.44	0.0

Abbreviations: RCT, randomized controlled trial; RR, relative risk.