You Are Here: U.S. Preventive Services Task Force > Topic Index > Screening: Alcohol Misuse > Summary of the Evidence > Appendix Table 3
Screening: Alcohol Misuse
Appendix Table 3. Evidence Tablea
| Study, Year (Reference) |
Participant Selection | Behavioral Intervention | Outcomes at 12 Months | Generalizability | Study Summary |
|---|---|---|---|---|---|
Richmond et al., 199561 Controlled clinical trial in 40 primary care practices involving 119 general practitioners. Australia Standard drink = 10 g ETOH |
378 adults (age 18-70 y) attending primary care visits who drank >35 drinks/wk (men) or >21 drinks/wk (women) Mean age: 39 y Baseline alcohol consumption: mean 38.5 drinks/wk Alcohol assessment: 2-step alcohol assessment in the waiting room before a routine visit. Patients self-administered 3 min Health & Fitness Questionnaire assessing alcohol, smoking, exercise, and weight; if results were "positive," 15-minute interview assessment by research assistant including drinking diary for past week. |
IG1 (n = 93) had alcohol assessment results placed on the chart for their visit with their usual PCP. IG2 (n = 96) received results of the assessment and brief (5 min) within-visit physician advice and a self-help manual. IG3 (n = 96) received the same brief advice intervention with 4 additional 15-20 min provider visits at 1 wk, 1 mo, 3 mo, and 5 mo. Delivery: 51% got only single visit (IG2 protocol) CG (n = 93) assessment results not put on chart. Not followed at 12 mo. |
Note: For IG1and IG2 only, because intervention delivery inadequate for IG3. Mean drinks/wkb Binge/heavy drinking episodes: NR Not exceeding recommendations: ≤28 units for men; ≤14 units for women (calculated from intention-to-treat analysis) |
Broadly includes heavier drinkers (one third "moderately dependent") attending primary care. Excludes persons with severely dependent/severe alcohol-related problems, persons with previous or current alcohol treatment, or those for whom any alcohol consumption was contraindicated. Systems support: Usual care providers "trained." Receptionist or research assistant did patient screening and prompted physician. No incentives. |
FAIR QUALITY: Non-random assignment to study conditions could have allowed manipulation. True control condition followup not assessed. Possible contamination between IG1 and IG2. Delivery of IG3 inadequate to differentiate it from IG2. Baseline and followup non-comparability of groups on several measures, not controlled in all analyses. Very brief intervention (IG2) and assessment only (IG1) reduced consumption at 12 mo with no significant differences between conditions. |
WHO Brief Intervention Study Group, 199658 RCT in various outpatient medical settings. 8 countries including United States Standard drink = 1.5 cL ETOH (14 g or 0.5 oz) |
1,559 adults (age 18-70 y) who drank >50 g ETOH/d (men) or 32 g ETOH/d (women) OR 6 or more drinks/occasion. Mean age: NR Baseline alcohol consumption: NR Alcohol assessment: Alcohol assessed in 2-step process: initial screening interview followed by 20-min face-to-face health interview addressing alcohol and other lifestyle issues. |
IG1 (n = 503) received 5 min of health advice from a "health advisor" (46% RNs, 18% MDs, 35% other) as part of a routine primary care visit. Intervention included feedback, advice, goal-setting. Delivery: NR IG2 (n = 565) received 15 min of brief counseling from health advisor who also addressed behavioral techniques as part of the routine visit. Some sites offered 3 followup visits. Intervention included feedback, advice, goal-setting, assistance, followup (for some subsets). Delivery: NR CG (n = 491) received assessment only. |
Outcomes assessed at 6-19 mo (mean, 9 mo) Average cL of alcohol/d Average cL of alcohol per drinking occasion: Mean drinks/wk: NR Reporting drinking within recommended weekly limits (no more than 24 cL of ETOH/wk for men or 13.3 cL ETOH/wk for women) |
Broadly includes multi-cultural heavier drinking primary care patients, many of whom may have been help-seeking. Excludes known or suspected alcoholics or very high daily consumers, those with prior liver damage or alcohol dependence treatment, and those warned by MD or other health professional to abstain. Systems support: Some provider training reported. No incentives reported. |
FAIR QUALITY: Limited information to evaluate study quality regarding baseline comparability of groups and maintenance of comparable groups. Potential for contamination exists since different interventions were delivered by same interventionists. Very brief and brief interventions reduced daily alcohol consumption in men at an average of 9 mo followup compared with assessment only. Some interventions could have been brief multi-contact. Among women, all groups significantly reduced consumption at followup without between-group differences. |
Anderson & Scott, 199254 RCT conducted in 8 community-based primary care group practices. England Standard drink = 10 grams ETOH |
154 male patients age 17-69 y registered with practices who exceeded 35 drinks/wk. Mean age: 45.1 y Baseline alcohol consumption: 52 drinks/wk Alcohol assessment: 2 steps: self-administered Health Survey Questionnaire by mail or in waiting room. If participants drank > 35 drinks/wk, they were invited to structured assessment interview of alcohol use with research staff outside clinic. |
IG (n = 80) received 10-min face-to-face visit with usual PCP at special visit scheduled after assessment. Intervention included advice and feedback. Delivery: NR CG (n = 74) received no intervention after assessment unless requested. |
Change in mean drinks/wk: Not bingeing: Percent attained low-risk drinking as measured by 22 or fewer DR/week: |
Broadly includes heavier drinking male (up to 105 drinks/wk) primary care patients. 41% of patients hsd abnormal dependence scores. Excludes those drinking >105 drinks/wk and those who received advice to cut down in previous year. Systems support: Provider training (15-30 minutes). Research staff did alcohol assessment entirely outside clinic. No incentives. |
GOOD QUALITY: Relatively high attrition levels (IG, 31% and CG, 39%) but these were addressed by replacing missing values with baseline consumption levels. Otherwise overall good quality criteria met. This brief intervention showed improved low-risk drinking, improved bingeing, and nearly significant changes in mean drinks/wk. |
Maisto et al., 200160 RCT in 12 primary care clinics. United States Standard drink = 0.6 oz ETOH |
301 patients of primary care practices aged ≥ 21 y with AUDIT score ≥ 8 OR ≥ 16 drinks/wk (men) or ≥ 12 Drinks/wk (women). Mean age: 45.5 y Baseline consumption: 5.5 drinks/drinking day Alcohol Assessment: Self-administered AUDIT embedded in lifestyle questionnaire. If "positive," face-to-face structured 30-day TLFB alcohol assessment interview including AUDIT and Q/F questions, labs, tests, and blood pressure. Assessment results for all participants forwarded to PCP. |
IG1 (n =100) immediately after assessment received 10-15 min "brief advice" from research staff, which intentionally limited patient input. Intervention included feedback, advice, goal-setting. Delivery: 93% got brief advice session IG2 (n = 101) received 30-45 min "motivational enhancement" session from research interventionist and 2 15-20 min booster sessions. Intervention included feedback, advice, goal-setting, assistance, and followup. Delivery: 91% ≥ 1 and 69% all 3 sessions. CG (n = 100) had assessment results forwarded to PCP. |
Change in mean drinks/drinking day: Change in mean drinks/wk: Binge/heavy episodes: NR Reporting benefit: NR |
Broadly includes primary care patients with risky/harmful drinking. Excludes those with acute alcoholic symptoms/recent substance abuse treatment. Not clearly applicable to primary care, because there were no definite or clear provider/clinical staff roles. Systems support: Research staff provided all of assessment and intervention. No provider training reported. Participants were paid for all assessments except the initial one. |
FAIR QUALITY: Fairly high loss to followup (23%) with intention-to-treat analysis of complete cases only (no replacement of missing values). Unclear blinding of participants and outcomes. Potential contamination between levels of IG (since IG1 could have gotten more intensive intervention) and between IG and CG (since all participants' doctors received assessment results, but unclear how or if these were acted upon. Two intensities of motivational interviewing-based interventions by non-clinical staff showed null effects with comparable reductions in alcohol consumption among interventions and control. |
Nilssen, 199157 RCT conducted within the Tromso Study. Norway Standard drink: NR |
338 community-dwelling adults who met high-risk alcohol use criteria (drinking ≥1 bottle of wine or equivalent per occasion 1-2 times per month OR drinking alcohol 2-3 times weekly) AND elevated GGT levels (45-200 u/L). Mean age: 42 y (approximately). Women: 14% Baseline consumption: NR Alcohol assessment: Population-based coronary heart disease risk factor screening of men aged 12-62 y and women aged 12-56 y included physical exam, lab tests, and questions about alcohol consumption along with other health behaviors. Risk group randomly assigned. |
IG1 (n = 113) invited by letter to re-exam for "elevated blood test"; received info on elevated GGT level (including alcohol) and GGT redrawn. Mailed repeat GGT results and invited to re-screen at 1 year. Interventions included feedback assistance and letter followup IG2 (n = 113) also invited by same letter to re-exam; intervention focused on further assessing and addressing alcohol consumption. GGT redrawn and repeat visits w/lab tests offered until GGT normalized. Interventions included: feedback assistance and letter followup Delivery: NR CG (n = 112) had no alcohol-related contact. |
Mean alcohol consumption, g/d: Bingeing: NR Reporting benefit: NR |
Targeted "early stage problem drinkers" (those with moderately increased GGT and self-reported increased alcohol intake) and did so among people already willing to participate in a heart disease risk assessment at outpatient clinic setting. Excluded known alcoholics. Systems support: Staff and training not clear. No incentives reported. |
FAIR QUALITY: Report inadequately covers allocation concealment or blinding for participant or outcome assessment. Comparability of groups at baseline or followup not clear. Not clear who delivered the interventions or the potential for contamination. Brief intervention and brief, multi-contact interventions among more severely affected problem drinkers reduced daily alcohol consumption compared with no intervention. |
Scott & Anderson, 199059 RCT in 8 community-based primary care practices. England Standard drink = 1 unit (10 grams ETOH) |
72 women (age 17-69 y) registered with the practices who had weekly consumption of 21-71 units of alcohol/wk. Mean age: 44 y Abnormal alcohol dependence scores IG, 73%; CG, 41% Baseline alcohol consumption: 35.3 mean drinks/wk Alcohol assessment: 2-step alcohol & lifestyle assessment by research staff: if findings were "positive" on self-administered survey, research staff conducted assessment interview, including one-week drinking diary. |
IG (n = 33) received 10-min face-to-face visit with usual PCP at special visit scheduled after assessment. Interventions included feedback and advice Delivery: NR CG (n = 39) received nothing after assessment unless requested. |
Change in mean drinks/wk: Not bingeing at followup (≥14 units on ≥2 occasions in previous 3 months): Attaining low-risk drinking (Health Survey Questionnaire = 22 drinks/wk): |
Broadly includes heavier drinking female (up to 71 drinks/wk) primary care patients, >50% of sample with abnormal dependence score. Excludes women consuming ≥ 71 units/wk or those who received advice to cut down alcohol use in previous year. Systems support: Provider training (15-30 minutes). Research staff conducted all alcohol assessment outside clinic. No incentives. |
FAIR QUALITY: Non-comparable groups at baseline for percentage with abnormal dependence scores. Unclear allocation concealment. Intervention delivery uncertain and control possibly contaminated. Inadequate power. Brief intervention in heavier drinking females showed null effects on all alcohol consumption and other outcome measures. Both groups comparably reduced alcohol consumption. |
Senft et al., 199756 RCT conducted in 3 large primary care HMO group practices (47 clinicians) Oregon and Washington Standard DR = 0.5 oz ETOH |
516 adults age >21 attending primary care visits with AUDIT score 8-21 OR 2 AUDIT Q/F item scores >5 OR ≥6 drinks/occasion at least weekly. Mean age: 41.9 y Baseline alcohol consumption: 16.5 mean drinks/wk Alcohol Assessment: Self-administered AUDIT-based alcohol use survey in waiting room. |
IG (n = 260) received 30 seconds of advice from their usual PC provider during the visit, immediately followed by a 15-min motivational interviewing-based session with a research health counselor. Intervention included advice, goal setting, and assistance. Delivery: 70% received advice & MI session. CG (n = 256) received usual care after assessment. |
Mean drinks/wk (calculated from total drinks in prior 3 mo): Women: Men: Reporting no binge drinking: Reporting no more than 3 drinks/d for men and 2 drinks/d for women: |
Broadly includes risky/harmful adult drinkers in primary care. Excludes dependent drinkers, those with AUDIT score >21. Systems support: Providers prompted with script to give advice only; research staff delivered assessment and most of intervention. No incentives. |
GOOD QUALITY: Although loss-to-followup of 20% overall (and differentially greater in IG), with dropouts less educated, missing values replaced in sensitivity analysis with no impact on reported results. Otherwise overall good quality criteria met. Brief intervention with no effects on average consumption or bingeing; modest intervention effects, primarily on total drinking days for women at 12 mo. Mean drinks were reduced at 6 mo (P = .04) but not at 12 mo (P = .13). IG tended toward more benefit (drinking within recommended limits) at 12 mo. Screening, recruitment, and intervention all occurred at a single primary care visit. |
Curry et al., 200350 RCT conducted in HMO-based primary care with patients of 23 clinicians. Washington Standard drink = 14 g ETOH |
307 adult patients with AUDIT score ≤ 15 and risky use in past month: ≥2 mean drinks/d OR ≥2 occasions of ≥5drinks OR driving after ≥3 drinks, who kept primary care appointments. Mean age: 48.2 y Baseline alcohol consumption: 14.9 mean drinks/wk Alcohol Assessment: Researchers assessed alcohol use in 10-15 minute general health telephone interview (including AUDIT, alcohol use questions addressing Q/F, binge, driving after alcohol use) prior to scheduled routine visit. |
IG (n = 151) received very brief (1-5 minute) motivational message from their PCP and self-help manual at routine visit, plus up to 3 telephone counseling calls from research health educator. Intervention included feedback, advice, goal-setting, assistance, tailoring, and followup contact. Delivery: 99% got provider intervention and materials; 87% got at least 1 call. CG (n = 156) received usual care after assessment. |
Mean drinks/wk: Reporting not bingeing: Reporting no at-risk drinking pattern (outcomes adjusted for missing data at followup): |
Includes broadly defined risky/harmful adult drinkers with advance primary care appointments. Excludes persons with AUDIT score >15 and known alcoholics. Systems support: Provider training (15-60 minutes); research staff put intervention materials on chart and conducted assessment and followup calls. No incentives. |
GOOD QUALITY: Although high differential loss to followup (IG 34% and CG 22%), replacement of missing values using multiple imputation procedures in analysis. Otherwise, met overall good quality criteria. Brief, multi-contact intervention with minimal provider burden and multiple followup contacts was clearly delivered and reduced at-risk drinking patterns at 12 months. No effects on average consumption. |
Fleming et al., 199753 RCT conducted in 17 community-based primary care practices (64 physicians) in practice-based research network. Wisconsin Standard drink = 12 g ETOH |
774 adult patients (age, 18-65 y) with routine primary care visits who met "problem drinking" criteria: ≥2/4 CAGE questions OR men >14 drinks/wk OR ≥5 drinks/occasion; women >11 drinks/wk or ≥4 drinks/occasion. Mean age: NR Baseline alcohol consumption: 19.1 mean drinks/wk Alcohol assessment: Alcohol assessed in self-administered Health Screening Survey (embedding CAGE and alcohol Q/F questions) in waiting room. If results were "positive," then 30-min face-to-face lifestyle interview (including 7-day TLFB alcohol review) by research personnel. |
IG (n = 392) had 2 brief visits scheduled 1 mo apart with usual PCP plus a call from clinic nurse 2 wk after each visit. Intervention included feedback, goal setting, assistance, & followup. Delivery: 76% completed the protocol and received both physician visits. CG (n = 382) received usual care after assessment |
Mean drinks/wk: Women: Men: No bingeing in past 30 days: Women: Men: Not drinking excessively: Women: Men: |
Broadly includes lower-level risky/harmful drinkers visiting primary care. Excludes heavier users (>50 drinks/week) and those with alcohol treatment or symptoms of withdrawal in previous year or who recently received MD advice to change alcohol use. Systems support: Provider training (1 hour followed by 2 30-min booster sessions); research staff did all assessment; clinic nurses provided followup calls. Providers were paid $300 to participate and patients were paid $50 to complete study procedures. |
GOOD QUALITY: Low levels (≤10%) slightly differential loss to followup, but intention-to-treat with replacement of missing values. All other good quality criteria met. Brief, multi-contact intervention by the usual care PCP reduced alcohol consumption by men and women and reduced proportions bingeing at 12 mo compared with no intervention. Women showed the greatest treatment effects. |
Fleming et al., 199951 RCT conducted in 24 community-based primary care practices with 43 MDs in practice-based research network Wisconsin Standard drink: 12-14 g ETOH |
158 adults age ≥65 y with scheduled visits who met hazardous drinking criteria: ≥2/4 CAGE questions OR men >11 drinks/wk or ≥4 drinks/occasion; women >8 drinks/wk or ≥3 drinks/occasion. Age 65-75 y Baseline alcohol consumption: 16 mean drinks/wk. Alcohol assessment: 2-step alcohol and lifestyle assessment by research staff: if results "positive" on self-administered modified Health Screening Survey (including CAGE and alcohol Q/F questions), then 30-min face-to-face lifestyle interview (including 7-d TLFB alcohol review). |
IG (n = 71) had 2 brief 10-15 min visits scheduled one month apart with usual PCP plus calls from clinic nurse 2 wk after each visit. Intervention included feedback, goal setting, assistance, and followup. Delivery: 94% received at least 1 physician visit. CG (n = 87) received a general health booklet after assessment. |
Mean drinks/wk at 12 mo: Binge episodes in previous 30 days: Not bingeing: Not drinking excessively: |
Broadly includes lower-level risky/harmful seniors (age ≥65 y) visiting primary care. Excludes heavier users (>50 drinks/week) and those with alcohol treatment or symptoms of withdrawal in previous year or who recently received MD advice to change alcohol use. Systems support: Provider training provided; research staff did all assessment; clinic nurses provided followup calls. Providers were paid $250 to participate and patients were paid $70 to complete study procedures. |
GOOD QUALITY: Met overall good quality criteria. Brief, multi-contact intervention among fairly stable (75% married) adults ≥65 y reduced risky/harmful alcohol use at 12 mo for all alcohol consumption measures, including those relating to binge use. Effects were even greater than those seen with comparable intensity interventions in younger adults and occurred by 3 mo. Self-reported alcohol use was corroborated by family members. |
Ockene et al., 199952 RCT conducted in 4 primary care academic medical sites with 46 MDs and 47 NPs. Massachusetts Standard drink = 12.8 g ETOH |
530 adults seeking routine primary care who screened as "high-risk drinker" (≥2/4 CAGE questions OR men >12 drinks/week OR ≥5 drinks/occasion in past mo; women >9 drinks/wk OR ≥4 drinks/occasion in past mo), and who made a primary care visit. Age range: 21-70 y Baseline alcohol consumption: 18.9 mean drinks/wk. Alcohol assessment: 2-step alcohol and lifestyle assessment by research staff: if "positive" on self-administered or interview-based Health Habits Survey (including CAGE and alcohol Q/F questions), then 20-35 minute lifestyle interview (including 7-day TLFB alcohol review). |
IG (n = 274) received brief (5-10 min) face-to-face intervention tailored to patients' problem alcohol use from usual MD/NP at routine visit and asked to make a followup appointment. Intervention included advice, goal setting, assistance, tailoring, and followup. Delivery: 99% reported provider discussion and 59% had followup visit within 6 mo. CG (n = 256) received general health pamphlet after assessment. |
6 month outcomes only: Change in mean drinks/wk: Women: Men: Not bingeing at 6 months (calculated): Reporting safe weekly and non-binge drinking at 6 months: |
Includes broadly defined risky/harmful adult drinkers who have recently used primary care. Excludes those already in alcohol intervention program. Systems support: Provider training (2½ h); research staff put intervention materials on chart and provided assessment. No incentives. |
GOOD QUALITY: Met overall good quality criteria. Brief, multi-contact intervention with followup visit showed significant reductions in change in mean drinks/wk at 6 mo, even after adjustment for age, gender, and baseline drinking levels, and significantly improved proportion drinking safely. Binge use insignificantly improved. |
Wallace et al., 198855 RCT conducted in 47 group practices in research network. England & Scotland Standard drink = 1 unit (not further defined) |
909 adults (age 17-69 y) registered primary care patients with self-assessed drinking problems OR ≥2/4 CAGE questions OR drank >35 units/wk (men) or >21 units/wk (women). Mean age: 42 y (approx.) Women: 29.1% Non-white: NR Smokers: NR Baseline mean alcohol consumption: Females: 35.1 drinks/wk Males: 62.2 drinks/wk Alcohol assessment: 2-step alcohol and lifestyle assessment by research staff: if "positive" on self-administered Health Survey questionnaire at visit, then face-to-face structured interview of alcohol use. |
IG (n = 450) contacted by PCP to schedule at least 1-2 visit(s) with up to 5 visits possible as needed. Intervention included feedback, advice, goal setting, assistance, and followup. Delivery: 83% of men and 92% of women completed ≥1 visit; 57% of men and 65% of women completed ≥2 visits. CG (n = 459) received general health booklet after assessment and no alcohol advice unless GGT level >150 International Units per Liter (IU/L) or requested by patient. |
Weekly consumption (units):
Women: Men: Binge/heavy episodes: NR Not drinking excessively: Men: |
Broadly includes heavier drinking adult primary care patients. Excludes those with recent medical advice about drinking or with GGT level >150 IU/l. Systems support: Provider training not reported. Research nurse did assessment. No incentives reported. |
GOOD QUALITY: At followup, IG lost 17% and CG lost 11%, so missing values were replaced with baseline values in analyses. Otherwise, overall good quality criteria met. This brief, multi-contact intervention by the PCP reduced alcohol consumption by men and women and the proportion drinking excessively at 12 mo compared with no intervention. |
a AUDIT = alcohol use disorders identification test - 10 item instrument for risky/harmful use; CAGE = 4-item screening questionnaire to detect alcoholism; CG = control group; ETOH = alcohol; GGT = serum gamma glutamyltransferase; HMO = health maintenance organization; IG = intervention group (numbered 1, 2 if >1 per study; MD = physician; MI = motivational interviewing; NR = not reported; NS = not statistically significant; P <.05; PCP = primary care provider; Q/F = questions addressing quantity and frequency of alcohol use; RCT = randomized controlled trial; RN = registered nurse; TLFB = time line follow back procedure.
b No significant group by time interactions based on repeated measures analysis.
USPSTF Program Office |
