TABLE 3.
GRADE (Grading of Recommendations Assessment, Development, and Evaluation) approach to evaluate the quality of evidence of studies according to the primary outcomes of preventing alcohol return-to-drinking or promoting abstinence
| Certainty assessment | No. patients | Effect | |||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|
| No. studies | Study design | Risk of bias | Inconsistency | Indirectness | Imprecision | Other considerations | Psychosocial therapy | Standard of care | Relative (95% CI) | Absolute (95% CI) | Certainty |
| Alcohol return-to-drinking—nonrandomized studies | |||||||||||
| 7 | Nonrandomized studies | Serious a | Not serious | Not serious | Not serious | Strong association All plausible residual confounding would reduce the demonstrated effect |
73/330 (22.1%) | 163/437 (37.3%) | RR 0.52 (0.28– 0.99) | 179 fewer per 1000 (from 269 fewer to 4 fewer) | ⨁⨁⨁◯ Moderate a |
| Alcohol return-to-drinking—randomized controlled studies | |||||||||||
| 4 | Randomized trials | Not serious | Not serious | Not serious | Serious b | None | 52/168 (31.0%) | 45/146 (30.8%) | RR 0.99 (0.45–2.16) | 3 fewer per 1000 (from 170 fewer to 358 more) | ⨁⨁⨁◯ Moderate b |
| Abstinence—nonrandomized controlled studies | |||||||||||
| 5 | Nonrandomized studies | Very serious a | Not serious | Not serious | Serious b | All plausible residual confounding would reduce the demonstrated effect | 70/150 (46.7%) | 134/285 (47.0%) | RR 1.35 (0.62– 2.93) | 165 more per 1000 (from 179 fewer to 907 more) | ⨁◯◯◯ Very lowa,b |
| Abstinence—randomized controlled trials | |||||||||||
| 6 | Randomized trials | Serious c | Not serious | Not serious | Not serious | None | 71/238 (29.8%) | 59/218 (27.1%) | RR 1.17 (0.92– 1.48) | 46 more per 1000 (from 22 fewer to 130 more) | ⨁⨁⨁◯ Moderate c |
a
Significant bias as the comparator group often did not adjust for confounding factors.
b
Effect estimates with broad CIs.
c
Some trials with high ROB.
Abbreviations: ROB, Risk of Bias; RR, risk ratio.