Table G-13Behavioral counseling interventions for adults compared with usual care or with each other

Number of Studies; Number of SubjectsDomains Pertaining to Strength of EvidenceMagnitude of EffectStrength of Evidence
Risk of Bias; Design/ QualityConsistencyDirectnessPrecisionSummary Effect Size (95% CI)High, Moderate, Low, Insufficient
Mortality, all-cause mortality (person-years)
4; 2,006Low to medium;
RCTs/Fair and Good
InconsistentDirectImpreciseRate ratio 0.64 (95% CI, 0.24 to 1.7)aLow
Alcohol-related accidents
4; 1,117Medium;
RCTs/Fair and Good
ConsistentDirectImpreciseUnable to determine a magnitude of effectbInsufficient
Alcohol-related liver problems
0; 0NANANANANAInsufficient
Outpatient/primary care visitsc
5; 876Low;
RCTs/Fair and Good
InconsistentDirectImpreciseNo significant difference (WMD, 0.14 visits, 95% CI, -0.5 to 0.2)Low
Hospitalizations (hospital days)
3; 1,417Low;
RCTs/Fair and Good
InconsistentDirectImpreciseBest evidence found a significant difference in hospital days in the last 6 months for the intervention group than the control group at 6, 12, and 48 months (35 vs. 180, 91 vs. 146, and 420 vs. 664, p<0.001, p<0.001, and p<0.05, respectively).dLow
Emergency visits
2; 901Low;
RCTs/Fair and Good
ConsistentDirectImpreciseTrend favoring control, but not statistically significant. At 6, 12, and 48 months for intervention vs. control: 47 vs. 70, 60 vs. 62, and 302 vs. 376, p>0.10, p>0.10, and p<0.10, respectivelydLow
2; 901Low
RCTs/Fair and Good
InconsistentDirectImprecise12 months: average per subject benefit over $1,100 and benefit-cost ratio 5.6:1 (95% CI, 0.4 to 11.0).
48 months: cost per patient of $205, benefit per patient of $7,985, for a resulting benefit-cost ratio of 39 (95% CI, 5.4 to 72.5)d
Legal events: assault/battery/child abuse, resist/obstruct officer/disorderly conduct, criminal damage/property damage, theft/robbery, other arrests, controlled substance/liquor violations
1; 774Low
NA, single studyDirectImpreciseNo statistically significant differences reported except for controlled substance/liquor violations (2 vs. 11, p<0.05)aLow
Sick days and employment stability
0; 0NANANANANAInsufficient
Quality of Life
3; 353Medium
ConsistentDirectImpreciseNo differenceeLow

Analyses with the addition of the included studies in older adults (GOAL) and in younger adults14 trended further toward favoring behavioral interventions, but remained nonstatistically significant (rate ratio 0.52, 95% CI 0.22 to 1.2; 6 studies, 2,255 subjects).


Four studies reported data on accidents in adults. Studies were not designed or powered to detect differences in these outcomes. The best available evidence comes from Project TrEAT (N=774),15 which reported outcomes after 48 months of followup. The study found lower numbers of motor vehicle crashes with fatalities (0 vs. 2), motor vehicle crashes with nonfatal injuries (20 vs. 31), and motor vehicle crashes with property damage only (67 vs. 72), that were not statistically significantly different between the intervention and control groups. Two studies (Anderson 1992 and Scott 1990) reported accident scores (from an alcohol-related problems scale), both with endpoint scores numerically favoring the intervention group.16, 17 Neither found a significant change from baseline data for the intervention group or for the control group. One study conducted in Thailand reported alcohol-related accidents (1 in the intervention group and 4 in the control group) and alcohol-related traffic accidents (3 in the intervention group and 5 in the control group).18


List the actual outcome measures that were reported for primary care utilization.


Summary effect sizes and data are from Project TrEAT, as it provided the best evidence (due to design, sample size of 774, risk of bias, and duration of followup). For hospitalizations, two smaller studies of shorter duration reported no statistically significant difference between groups for hospitalization outcomes, but Senft and colleagues (N=516) reported a slightly lower percentage of subjects hospitalized in the intervention group than the control group at 24 months that was not significant (21.2% vs. 22.0%, p=0.81) and a trend toward fewer mean hospital days for those hospitalized (4.7 vs. 6.6, p=0.37); Lock and colleagues (N=127) reported no significant difference between groups for hospital inpatient stays. 48-month cost data are from the societal perspective.15, 19


Two 12-month studies reported no difference in change in mean life quality scores between the intervention and control groups (0 vs. 0 and -0.3 vs. -0.3).16, 17 A nurse-led intervention (N=127) reported no significant differences between the intervention and control groups at 6 or 12 months for change in SF-12 physical or mental health scores.20

Abbreviations: CI = confidence interval; NA = not applicable; RCT = randomized controlled trial; vs. = versus; WMD = weighted mean difference.

From: Appendix G, Strength of Evidence Tables

Cover of Screening, Behavioral Counseling, and Referral in Primary Care To Reduce Alcohol Misuse
Screening, Behavioral Counseling, and Referral in Primary Care To Reduce Alcohol Misuse [Internet].
Comparative Effectiveness Reviews, No. 64.
Jonas DE, Garbutt JC, Brown JM, et al.

NCBI Bookshelf. A service of the National Library of Medicine, National Institutes of Health.