We selected articles for review based on information derived from the title, abstract, and keywords. If the title, abstract, and keywords did not yield enough information we then reviewed the full paper, We evaluated for inclusion all randomized studies of patients with known, suspected AUD, or alcohol-related injuries, assessing the effectiveness of brief ED-based interventions for the reduction of alcohol consumption, as well as the secondary goals of reducing alcohol-related negative consequences for both physical and social consequences of AUD,

Next, the articles that met inclusion criteria were appraised and assessed by two authors for their methodological quality, such as the method of randomization, blinding, allocation, description of withdrawals and dropouts, as well as loss to follow-up (Table).18 A third author reviewed the articles if there were any discrepancies in the grading. The reviewers were not blinded to the study hypothesis.

Given the lack of standardization across studies, including variations in patient populations, settings, screening techniques, and outcomes, data were analyzed descriptively. We focused on presenting trends and themes that emerged with regard to alcohol consumption and complications from continued alcohol use, such as injury. We also present the quality of studies that met our inclusion criteria.

The alcohol screening tools differed among the studies and included both self-reported questionnaires and biomarkers. Several structured questionnaires (Appendix B) were used to determine current and/or past alcohol use, and increased the sensitivity of self-report.56–76 Of the controlled randomized studies included in Appendix B, the self-reported screening instruments were as follows: one study used AUDIT (Alcohol Use Disorder Identification Test)-C42 and 12 studies used the full AUDIT.23, 24, 28, 31–33, 38, 39, 42, 50, 53, 54 Of these, 10 studies included all patients with a score of 8 or higher,28, 31–33, 38, 39, 42, 50, 53, 54 and two studies stratified the patients into three categories: low risk (0 to 6), at risk/moderate risk (7 to 18), and high risk (19 to 40). 3, 24 Authors mainly chose a lower cut-off score of 4 or greater for inclusion of adolescents.

In some studies, AUDIT was used with other alcohol screening tools such as the National Institute on Alcohol Abuse and Alcoholism (NIAAA) Guide,31 the CAGE questionnaire,23, 39 positive test for alcohol, and self-report of ingesting alcohol within six hours prior to the injury.50 Six studies21–24, 35, 39 used the CAGE questionnaire to screen injured patients for alcohol consumption; however, in one study it was followed by AUDIT to evaluate the quantity and frequency of alcohol consumption,24 while in another study the patients were initially screened using NIAAA followed by CAGE.35 Two studies, one in the United Kingdom and one in Australia, employed the Paddington Alcohol Test. (PAT)29, 34 *[PAT features a table of commonly encountered beverages coded in British units. Eight grams of alcohol are equivalent to one unit. The PAT allows for the different relative strengths of certain products, thus differentiating between a patient who may consume two pints (i.e., four “drinks”) of normal strength beer (four units) and the same amount of “strong” lager (10 units)].

Five studies used the NIAAA guide to screen patients.21, 22, 31, 35 In three of the five studies, NIAAA was used in conjunction with CAGE;21, 22, 35, in one additional study it was used with AUDIT, 31 and in another study it was used with Alcohol, Smoking and Substance Involvement Screening Test (ASSIST).43

Several instruments (Appendix C) were used to evaluate adolescent alcohol intake and consequences of drinking.77–85 Three studies involving adolescents 44, 48, 49 used Adolescent Drinking Questionnaire (ADQ) and Adolescent Drinking Index (ADI) instruments to evaluate alcohol consumption, and at follow-up used Adolescent Health Behavior Questionnaire and Short Michigan Alcoholism Screening Test (SMAST) to evaluate alcohol-related injuries. Eleven of the 35 studies used biomarkers (blood, breath or saliva tests) 25, 33, 36, 37, 39, 44–47, 48–50 as part of the screening tools.

Drinker’s Inventory of Lifetime Consequences (DrInC), a 45-item, self-report questionnaire about the negative consequences experienced from drinking that was validated on an alcohol treatment-seeking population of 1,728 inpatients and outpatients 86 and on Project MATCH, 87 was used by six studies 25, 27, 40, 50, 53, 55 to measure not only the physical but also the intrapersonal, social, interpersonal, and impulse control (e.g., driving while intoxicated, physical fights) consequences from drinking.

One author 50 used the Readiness to Change Contemplation Ladder 88 adapted for an ED treatment-seeking population of injured drinkers 69 to measure the subject’s attitude towards modifying alcohol-related behaviors with response categories ranging from 0 (no thought of changing) to 10 (taking action to change [e.g., cutting down]).

Brief interventions (BI) are designed to motivate reduction and cessation of substance use by exploring and highlighting individual risks and negative outcomes of problematic substance use. Though it is not intended to treat people with serious substance use disorders, it can be used to encourage those with more serious dependence to accept either more intensive treatment within the primary care setting or a referral to a specialized alcohol and drug treatment agency.

The most common behavioral therapies used in SBIRT programs are brief versions of cognitive behavioral therapy and motivational interviewing or some combination of the two. Brief interventions can be made more effective by using the technique of motivational interviewing. The principles of brief motivational interviewing (BMI) and asking for permission to discuss alcohol use; (2) providing feedback on current drinking and consequences; (3) assessing readiness to change; and (4) providing options to help with behavioral changes and assisting in obtaining appointments or placements if desired.91

ED-based brief interventions were performed by a variety of professionals and staff members including, physicians, medical students, mid-level providers,21, 31 nurses,35 social workers, psychologists,23, 36, community outreach workers and “health promotion advocates.”24 ED staff nurses trained to conduct SBIRT were less fully engaged with SBIRT implementation when the ED was extremely busy. 35 The training required to prepare staff for delivering BI included reading review of materials about the assessment of adverse consequences of alcohol abuse, 89 as well as structured sessions to teach and practice the principles and techniques of SBIRT. 20

All studies used reduction of alcohol consumption as the primary outcome measure. Thirteen studies (37%) enrolling a total of 5,261 participants reported significant differences between control and intervention groups defined by the number of drink days and number of units per drink day.21–23, 28, 29, 36, 37, 39, 40, 44, 45, 50, 51 Sixteen studies (46%) showed a reduction in alcohol consumption in both the control and intervention groups. 25, 26, 31– 37, 42, 43, 47, 49, 50, 52, 55 Nine25, 26, 35–37, 42, 49, 50, 55 out of these 16 studies showed greater improvement in the BI group as compared to the control group as follows: a higher reduction in the overall consumption of alcohol;25, 26, 35, 49 reduction in the concomitant use of marijuana and alcohol;55 and fewer injuries.35, 36, 50 However, the effectiveness of the interventions in reducing at-risk drinking was weakened at six- and 12-month follow-up points.22, 29–33, 41, 46, 48, 49

Seventeen out of 35 studies failed to demonstrate an intervention effect for the primary outcome of alcohol consumption reduction.21, 24–27, 30–35, 38, 41–43, 46–49, 51–55 However, 11 of those 17 studies (65%) enrolling a total of 4,706 participants observed some significant differences between BI and control conditions, at least for specific subgroups or secondary outcome criteria.21–26, 38, 41, 42, 47, 48, 53–55 For example, one author found24 statistically significant changes in “trying to be careful while drinking” in the intervention group in patients 18–21 years old with low AUDIT scores, and another26 reported decrease in drinking, drinking days per week, maximum drinks per occasion and negative consequences of drinking in injured patients older than 18 years old. Among adolescents, a subgroup with a history of previous drinking and driving, the intervention group showed a beneficial effect in the reduction of drinking and driving.41 Additionally, Segatto47 found in adolescent and young adult patients a decrease in the following outcomes: days of alcohol use; days with moderate and heavy use; and negative consequences. Spirito et al.48 found that the subgroup of adolescents who screened positive for problematic alcohol use at baseline reported significantly more improvement with fewer drinking days as well as fewer high-volume drinking days. Focusing on women, for example, the subgroup age 22 years, a reduction was found on the Drinker Inventory of Consequences (DrInC) in the intervention group.25 Havard et al.38 also found that women in the intervention group engaged in heavy drinking at one third of the frequency as the control group.

In some studies BI was shown to have an effect only on low or moderate drinkers 23 and not on high-risk or dependent drinkers (defined as an AUDIT score >15 or >18, respectively). However, Mello et al.42 found the subgroup of participants with AUDIT scores>15 in the BI group had a lower three-month impaired score. If the participants attributed their injury to alcohol, Walton et al.53 demonstrated lower levels of average alcohol consumption and less-frequent heavy drinking in the BI group. In addition, Wang et al.54 found a significant increase in readiness to change in the BI group (in excessive alcohol users, AUDIT 2+ for men and 1+ for women), but not in the control group. Woolard et al.55 showed binge drinking and concomitant marijuana use decreased for the BI group.

A study by Stein 50 looked at pretreatment readiness to reduce drinking as a mediator of BMI effectiveness on alcohol-related consequences and found positive effects only on those highly motivated to change prior to the intervention but not for those with low pre-intervention motivation.

In the United Kingdom study by Crawford, 29 the patients were screened in the ED and then referred for outpatient follow-up with an alcohol health worker (AHW) for about 30 minutes of assessment and discussion of current and previous drinking. Of those referred, 65.8% followed up with an AHW. Alcohol consumption in patients who followed up with an AHW decreased to a mean of 59.7 units* per week as compared with 83.1 units in patients in the control group (t –2.4, p=0.02). At 12 months, those who pursued follow-up were drinking 57.2 units per week compared with 70.8 in controls (t –1.7, p=0.09). This study also showed that the patients followed by the AHW had a mean of 0.5 fewer visits to the ED over the following 12 months (1.2 compared with 1.7, t –2.0, p=0.046).