AUDIT-C Scores as a Scaled Marker of Mean Daily

Drinking, Alcohol Use Disorder Severity, and Probability of

Alcohol Dependence in a U.S. General Population

Sample of Drinkers

Anna D. Rubinsky, Deborah A. Dawson, Emily C. Williams, Daniel R. Kivlahan, andKatharine A. Bradley

Background: Brief alcohol screening questionnaires are increasingly used to identify alcohol misusein routine care, but clinicians also need to assess the level of consumption and the severity of misuse sothat appropriate intervention can be offered. Information provided by a patient’s alcohol screeningscore might provide a practical tool for assessing the level of consumption and severity of misuse.

Methods: This post hoc analysis of data from the 2001 to 2002 National Epidemiologic Survey onAlcohol and Related Conditions (NESARC) included 26,546 U.S. adults who reported drinking in thepast year and answered additional questions about their consumption, including Alcohol Use DisordersIdentification Test—Consumption questionnaire (AUDIT-C) alcohol screening. Linear or logisticregression models and postestimation methods were used to estimate mean daily drinking, the numberof endorsed alcohol use disorder (AUD) criteria (“AUD severity”), and the probability of alcoholdependence associated with each individual AUDIT-C score (1 to 12), after testing for effect modifica-tion by gender and age.

Results: Among eligible past-year drinkers, mean daily drinking, AUD severity, and the probabilityof alcohol dependence increased exponentially across increasing AUDIT-C scores. Mean daily drinkingranged from < 0.1 to 18.0 drinks/d, AUD severity ranged from < 0.1 to 5.1 endorsed AUD criteria,and probability of alcohol dependence ranged from < 1 to 65% across scores 1 to 12. AUD severityincreased more steeply across AUDIT-C scores among women than men. Both AUD severity andmean daily drinking increased more steeply across AUDIT-C scores among younger versus older agegroups.

Conclusions: Results of this study could be used to estimate patient-specific consumption and sever-ity based on age, gender, and alcohol screening score. This information could be integrated into elec-tronic decision support systems to help providers estimate and provide feedback about patient-specificrisks and identify those patients most likely to benefit from further diagnostic assessment.

Key Words: AUDIT-C, Alcohol Screening, Mean Daily Drinking, Alcohol Misuse Severity,Alcohol Dependence.

ALCOHOL MISUSE SCREENING followed by briefcounseling for those who screen positive has been

deemed a top prevention priority for U.S. adults based onthe health impact of alcohol misuse and cost-effectiveness ofbrief alcohol interventions (Jonas et al., 2012; Solberg et al.,2008). Healthcare systems are increasingly implementingalcohol misuse screening and intervention in primary care(Williams et al., 2011). However, providers often lack thenecessary time and resources to assess the severity of alcoholmisuse (Nilsen, 2010), which ranges from drinking over rec-ommended limits to meeting diagnostic criteria for alcoholdependence.

A variety of preventive and treatment services are recom-mended for patients with alcohol misuse, depending on theseverity (Willenbring, 2009). Clinicians should considerwhere a patient falls on the spectrum of severity and be ableto respond accordingly with a range of appropriateinterventions (National Institute on Alcohol Abuse and

From the Health Services Research & Development (ADR, ECW,DRK, KAB), Veterans Affairs Puget Sound Health Care System, Seattle,Washington; Department of Health Services (ADR, ECW, KAB), Uni-versity of Washington, Seattle, Washington; Laboratory of Epidemiologyand Biometry, Division of Clinical and Biological Research (DAD),National Institute on Alcohol Abuse and Alcoholism, National Institutesof Health, Bethesda, Maryland; Kelly Government Services, Inc. (DAD),Bethesda, Maryland; Center of Excellence in Substance Abuse Treatmentand Education (DRK, KAB), Veterans Affairs Puget Sound Health CareSystem, Seattle, Washington; Department of Psychiatry and BehavioralSciences (DRK), University of Washington, Seattle, Washington;Department of Medicine (KAB), University of Washington, Seattle,Washington; and Group Health Research Institute (KAB), Seattle,Washington.

Received for publication June 14, 2012; accepted December 4, 2012.Reprints requests: Anna D. Rubinsky, PhD, MS, Health Services

Research & Development, Veterans Affairs Puget Sound Health CareSystem, 1660 S. ColumbianWay S-152, Seattle, WA 98108; Tel.:206-277-4156; Fax: 206-768-5343; E-mail: anna.rubinsky@va.gov

Copyright© 2013 by the Research Society on Alcoholism.

DOI: 10.1111/acer.12092

Alcohol Clin Exp Res,Vol **, No *, 2013: pp 1–11 1

ALCOHOLISM: CLINICAL AND EXPERIMENTAL RESEARCH Vol. **, No. *** 2013

Alcoholism, 2007; Willenbring, 2009). Although alcoholscreening questionnaires are typically used as dichotomoustests for misuse, the resulting scores capture additional infor-mation. If continuous alcohol screening scores werecalibrated in terms of level of consumption, severity of alco-hol use disorder (AUD), and likelihood of dependence,screening questionnaires might serve as valuable tools forassessing the severity of misuse. Specifically, providers couldutilize this information to provide the patient with individu-alized feedback about alcohol-related risks associated withtheir particular score and to make more informed decisionsabout need for further diagnostic assessment.

The 3-item Alcohol Use Disorders Identification Test—Consumption questionnaire (AUDIT-C) is a validatedscreen for alcohol misuse appropriate for routine screeningin primary care (Bradley et al., 2009; Bush et al., 1998). TheAUDIT-C comprises the consumption questions of the10-item AUDIT (Babor et al., 1989) and performs similarlyto the full AUDIT for identifying the spectrum of alcoholmisuse (Kriston et al., 2008; Reinert and Allen, 2007). TheAUDIT-C is often more practical than the AUDIT for usein routine care because of its brevity and is currently used forannual screening of all outpatients in the Veterans AffairsHealthcare System (Bradley et al., 2006).

AUDIT-C scores have been associated with several alco-hol-related health risks, including alcohol dependence(Rubinsky et al., 2010), severity of problem drinking (Bradleyet al., 2004), postoperative complications (Bradley et al.,2011), hospitalizations for gastrointestinal conditions (Lem-bke et al., 2011), trauma (Williams et al., 2012), and mortal-ity (Harris et al., 2010), generally in a dose–responsemanner. However, despite extensive literature establishingthe AUDIT-C as a scaled marker of alcohol-related risk, weare not aware of any study that has evaluated the level ofalcohol consumption across AUDIT-C scores or describedthe severity of AUDs associated with individual AUDIT-Cscores. The 3 questions of the AUDIT-C assess differentdimensions of consumption, with each question scored on adifferent scale (i.e., drinking d/wk, drinks/drinking day, andfrequency of heavy drinking). Therefore, the summed totalscore can reflect varying patterns of typical drinking andatypical heavy drinking, and increasing scores do not neces-sarily reflect linear increases in a given dimension of con-sumption. Further, reported consumption on simple globalquestions about typical drinking, such as first 2 items of theAUDIT-C, is often lower than estimated consumption basedon in-depth assessments that include beverage-specific infor-mation on the respondent’s usual drink size and brand(Bradley et al., 1998; Dawson, 1998).

The purpose of this study was to estimate mean dailydrinking, AUD severity, and probability of alcohol depen-dence, based on valid, reliable measures from structuredin-depth interviews, across individual AUDIT-C scores.Because the screening performance of the AUDIT-Cdepends on gender and age (Dawson et al., 2005b), each out-come was estimated separately for men and women and for

varying age groups after testing for effect modification bythese factors. Additionally, because low-level mean dailydrinking can reflect frequent drinking of small quantities oroccasional heavy drinking, secondary analyses investigatedparticular patterns of drinking across AUDIT-C scores.

MATERIALS ANDMETHODS

Data Source and Sample

This study represents a post hoc analysis of data from the 2001 to2002 National Epidemiologic Survey on Alcohol and RelatedConditions (NESARC) (Grant and Dawson, 2006; Grant et al.,2003a,b), which included 43,093 U.S. adults aged 18 years and older(overall response rate 81%). Structured interviews were administeredin respondents’ homes by trained lay interviewers using computer-assisted personal interviewing software that included built-in skippatterns and consistency safeguards. A final weighting factor wasassigned to each respondent to account for nonresponse and ensurethat the data were representative of the U.S. population in terms ofdemographics and geographic distribution.

The present cross-sectional analysis was based on a subsample of26,546 NESARC respondents who reported having at least 1 drinkin the previous 12 months and answered additional questions abouttheir past-year alcohol consumption, including items that corre-spond to the AUDIT-C.

Measures

Overview. All measures of alcohol consumption and AUDswere based on the Alcohol Use Disorder and Associated Disabili-ties Interview Schedule (AUDADIS-IV) (Grant et al., 2001), anin-depth assessment of past-year drinking and associated symp-toms with demonstrated reliability (Grant et al., 2003a) and valid-ity (Dawson et al., 2005a, 2008, 2010, 2012; Grant et al., 2003a,2007). Recommended drinking limits for healthy adults are nomore than 14 (men) or 7 (women) drinks/wk and no more than 4(men) or 3 (women) drinks in a single day (National Institute onAlcohol Abuse and Alcoholism, 2007). For each of 4 specific alco-holic beverage types individually (i.e., coolers, beer, wine, and spir-its), as well as all alcoholic beverages combined (total), theNESARC asked about the overall frequency of consumption, typi-cal and largest quantities consumed, frequency of drinking thelargest quantity, and frequency of drinking � 5 drinks in a singleday. Additionally, women were asked about the frequency ofdrinking � 4 total drinks in a single day. For each specific bever-age type, respondents’ typical drink size and brand were alsoascertained. To assess AUDs, the NESARC included 33 alcoholsymptom items that operationalized the 4 criteria of alcohol abuseand 7 criteria of alcohol dependence (11 total AUD criteria)according to Diagnostic and Statistical Manual of Mental Disor-ders—Fourth Edition (DSM-IV) specifications (American Psychi-atric Association, 1994).

AUDIT-C Alcohol Screening Score. As above, the AUDIT-Calcohol screening questionnaire comprises 3 questions about past-year consumption (Bush et al., 1998). Each question is scored 0 to 4points and summed for a total score ranging from 1 to 12 pointsamong drinkers. Scores � 4 (men) and � 3 (women) indicate posi-tive screens for alcohol misuse, based on the best balance of sensitiv-ity and specificity (Reinert and Allen, 2007).

NESARC AUDIT-C scores were derived from questions in thesequence about total consumption of all alcohol combined. TheNESARC AUDIT-C follows standard scoring conventions as clo-sely as possible, but the questions and response options varied

2 RUBINSKY ET AL.

somewhat from the most widely validated version (Table 1) (Daw-son et al., 2005b). Nonetheless, the NESARC AUDIT-C performswell for identifying alcohol misuse and AUDs (Dawson et al.,2005b), with areas under the receiver operating curve and recom-mended cut-points similar to those reported for other versions (Re-inert and Allen, 2007).

OutcomeMeasures

All outcome measures of consumption were based on the seriesof questions about past-year consumption of specific beveragetypes, as well as all alcohol combined, and made use of beverage-specific information on drink size and ethanol (EtOH) content ofmain brand. Based on the respondent’s typical brand and drink sizefor each specific beverage type, reported drinks of each beveragetype were adjusted for EtOH content and converted to U.S. stan-dard-sized drinks containing 0.60 ounces of EtOH.

Primary OutcomeMeasures

Mean Daily Drinking. Mean alcohol consumption in drinks perday was computed using a weighted formula incorporating typicaldrinking as well as episodic heavy drinking (National Institute onAlcohol Abuse and Alcoholism [NIAAA], 2004). Estimates werebased on the larger of (i) the sum of standard-sized drinks from the4 series of beverage-specific questions or (ii) reported total drinks ofall alcohol (unadjusted for drink strength or pour size). Values were

top-coded to 24 drinks/d to avoid undue influence of outliers(affecting < 0.5%).

Alcohol Use Disorder Severity. AUD severity was based on thecount (0 to 11) of endorsed DSM-IV AUD criteria for the prioryear. A simple count performs as well as dimensional severity mea-sures that are weighted for the severity of each endorsed criterion(Dawson et al., 2010).

Alcohol Dependence. To meet diagnostic criteria for alcoholdependence, at least 3 of the 7 DSM-IV dependence criteria had tobe endorsed for the prior year.

Secondary OutcomeMeasures—Patterns of Drinking

Mean daily drinking can obscure important variation in drinkingpatterns, including differences in maximum quantity consumed andfrequency of episodic heavy drinking.

Maximum Quantity of Drinks. Similar to mean daily drinking,maximum quantity of drinks consumed on any day in the past yearwas based on either (i) standard-sized drinks of any specific bever-age type or (ii) reported total drinks of all alcohol, whichever wasgreater, and values were top-coded to 24 drinks (affecting < 1%).Maximum quantity of drinks reflects the level of variability in anindividual’s consumption and has been suggested as an importantcomponent in understanding drinking patterns (Dawson et al.,

Table 1. Wording and Response Options from the NESARC’s AUDIT-C Compared with the Most Widely Validated Version

AUDIT-C question and wording NESARC response categories Score Standard AUDIT-C equivalent Score

Question 1NESARC AUDIT-C: During the last 12 months,about how often did you drink ANY alcoholicbeverage?

Standard AUDIT-C: How often did you have adrink containing alcohol in the past year?

Every day 4 4 or more times a week 4Nearly every day 43–4 times a week 3 2–4 times a month 32 times a week 3Once a week 2 2–3 times a week 22–3 times a month 2Once a month 1 Monthly or less 17–11 times in the last year 13–6 times in the last year 11–2 times in the last year 1Nevera 0 Never 0

Question 2NESARC AUDIT-C: Counting all types of alcoholcombined, howmany drinks did you USUALLYhave on days when you drank during the last12 months?

Standard AUDIT-C: Howmany drinks did you haveon a typical day when you were drinking in thepast year?

Open-ended responses groupedas follows1 or 2 0 1 or 2 03 or 4 1 3 or 4 15 or 6 2 5 or 6 27–9 3 7–9 310 or more 4 10 or more 4

Question 3NESARC AUDIT-C: During the last 12 months,about how often did you drink 5 or more drinksin a single day?

Standard AUDIT-C: How often did you have 6 ormore drinks on 1 occasion in the past year?

Every day 4 Daily or almost daily 4Nearly every day 43–4 times a week 3 Weekly 32 times a week 3Once a week 32–3 times a month 2 Monthly 2Once a month 27–11 times in the last year 1 Less than monthly 13–6 times in the last year 11 or 2 times in the last year 1Nevera 0 Never 0

NESARC, National Epidemiologic Survey on Alcohol and Related Conditions; AUDIT-C, Alcohol Use Disorders Identification Test—Consumptionquestionnaire.

aThis response category was determined from a separate screening question rather than by the endorsement of “never” on Question 1.Table adapted from Dawson and colleagues (2005b).

AUDIT-C AS A SCALEDMARKEROF SEVERITY 3

2005a; Greenfield et al., 2006; National Institute on Alcohol Abuseand Alcoholism [NIAAA], 2003).

Frequency of Heavy Drinking. The number of heavy drinkingdays per week was based on the highest frequency of either (i) drink-ing the largest quantity of any specific beverage type if the quantitywas � 5 (men) or � 4 (women) standard-sized drinks (based onstandard rounding procedures) or (ii) drinking � 5 (men) or � 4(women) reported total drinks in a single day. Frequency of heavydrinking is associated with several health consequences (Dawsonet al., 2008; Saha et al., 2007) and may add risk beyond that associ-ated with volume of mean daily consumption (Rehm et al., 2010a).

Exceeding Recommended Drinking Limits. Recommended dailydrinking limits were based on National Institute on Alcohol Abuseand Alcoholism (NIAAA) guidelines for healthy adults aged65 years or less (National Institute on Alcohol Abuse and Alcohol-ism, 2007). Exceeding recommended maximum daily drinking limitswas defined as drinking � 5 (men) or � 4 (women) drinks (i.e.,heavy drinking) at least once in the past year. Exceeding recom-mended average daily drinking limits was defined as mean dailydrinking >2 (men) or >1 (women) drinks/d. Although NIAAA sug-gests lower average daily limits (� 1 drink/d) for men above65 years because of potentially increased sensitivity to the effects ofalcohol (National Institute on Alcohol Abuse and Alcoholism,2007), we used the same threshold for all men so that the measurewould be comparable across age groups. Further, recent studiesindicate that the risk of mortality, cognitive and physical disability,depression/anxiety, and diminished self-reported health are nogreater for older adults who drink >1 drink daily than for those whodrink � 1 drink daily (Kirchner et al., 2007; Lang et al., 2007), call-ing into question the lower suggested limits for older men.

Statistical Analyses

Analyses were conducted using Stata/MP 12 software (Stata-Corp, 2011), and all estimates of variance accounted for NESARC’scomplex multistage sampling design and final weighting usingTaylor series linearization. Pearson chi-square tests of independencewere adjusted for the survey design and weighting based onRao–Scott correction F-statistics.

Sociodemographic characteristics of men and women in the studysample were described overall and across each AUDIT-C score (1to 12). Chi-square tests (for categorical variables) and analysis ofvariance (for continuous variables) were used to evaluate associa-tions between sociodemographic characteristics and AUDIT-Cscores.

Linear regression models and postestimation methods were usedto estimate average values of mean daily drinking and AUD severityat each AUDIT-C score, after testing for effect modification by gen-der and age group (18 to 29, 30 to 44, 45 to 64, 65+ years). Logisticregression models and postestimation methods were used to esti-mate the average predicted probability of alcohol dependence ateach AUDIT-C score, after testing for effect modification by genderand age. Each outcome was estimated across individual AUDIT-Cscores separately for men and women and for each age group, withdifferences by gender and age evaluated at each score based on post-estimation Wald tests. To allow for nonlinear patterns acrossAUDIT-C scores, individual scores were modeled as separate cate-gorical variables.

Secondary analyses investigated patterns of drinking acrossAUDIT-C scores using the same statistical approach as for mainanalyses. Average values of maximum quantity consumed andfrequency of heavy drinking were estimated and compared based onlinear regression methods. Average predicted probabilities ofexceeding recommended daily drinking limits were estimated andcompared based on logistic regression methods.

RESULTS

Sample Characteristics

Of the 26,546 (12,823 men and 13,723 women) respon-dents who reported drinking in the past year and met allstudy inclusion criteria, most had AUDIT-C scoresconsistent with a negative screen for alcohol misuse (scores 1to 3 for men and 1 to 2 for women), and the number of menand women with each score generally decreased as scoresincreased (Table 2). Individuals with higher AUDIT-Cscores were younger, less likely to be married, and morelikely to have low family incomes (Table 2).

Mean Daily Drinking

Mean daily drinking increased exponentially acrossincreasing AUDIT-C scores, from < 0.1 drink/d at a scoreof 1 to 18 drinks/d at a score of 12 (Table 3). AcrossAUDIT-C scores 2 to 7, each 1-point increase in score wasassociated with an increase in mean daily drinking of approx-imately 0.5 drinks. This magnitude of increase doubled toapproximately 1 drink/d for each additional point acrossscores 7 to 9, was 3 drinks/d for each additional point acrossscores 9 to 11, and reached over 6 drinks/d with a change inscore from 11 to 12. The relationship between AUDIT-Cscore and mean daily drinking was not modified by gender,and estimates for men and women were similar at every score(Fig. 1). Age, however, did modify the relationship(p < 0.00005), with the oldest age group (65+ years) havingthe highest mean daily drinking at AUDIT-C scores 4 to 6but the lowest consumption at scores 10 to 12 (Fig. 1).Although the oldest group differed most from the other agegroups, post hoc exploratory analyses revealed an interactionalso among the 3 age groups < 65 years (p < 0.00005),whereby those aged 45 to 64 had increasingly higher con-sumption compared with younger individuals acrossAUDIT-C scores 3 to 9, but not at higher or lower scores.

Alcohol Use Disorder Severity

Only 24% of drinkers endorsed at least 1 AUD criterion,and of these, nearly half endorsed a single criterion. MeanAUD severity ranged from < 0.1 to 0.5 across AUDIT-Cscores 1 to 4, increased from 1.0 to 3.3 across scores 5 to 11,and spiked to 5.1 at a score of 12 (Table 3). The relationshipbetween AUDIT-C score and AUD severity was modified byboth gender and age group (p-values < 0.00005). Womenhad increasingly higher AUD severity compared with menacross AUDIT-C scores 4 to 7 (Fig. 1), but gender differ-ences were not significant at higher scores, likely due to thesmall numbers of women with high scores. The interactionbetween AUDIT-C scores and age was relatively complex.AUD severity increased most steeply across AUDIT-Cscores among the youngest individuals (18 to 29 years) andleast steeply among the oldest individuals (65+ years)(Fig. 1). However, whereas AUD severity increased expo-

4 RUBINSKY ET AL.

nentially across AUDIT-C scores among those aged 18 to 29,30 to 44, and 45 to 64, reaching 5.9, 5.0, and 4.3 at a score of12, respectively, it reached a maximum of only 1.3 amongthose aged 65 years and above. Post hoc exploratory analy-ses revealed that an age interaction persisted among individu-als under age 65 whereby AUD severity was similar for thoseaged 30 to 44 and 45 to 64, but higher for those aged 18 to 29,with differences generally increasing across increasing scores.

Probability of Alcohol Dependence

The probability of past-year alcohol dependence followeda similar pattern to AUD severity across AUDIT-C scores,even as a function of gender and age (Fig. 1), but was notsignificantly modified by either factor, likely because of smallnumbers of individuals (especially women and older adults)with dependence at any given score. The probability ofdependence was < 5% among individuals with AUDIT-Cscores 1 to 4, increased to approximately 40% at scores 9 to11, and reached 65% at a score of 12 (Table 3). The proba-bility of dependence was significantly higher in women thanmen at scores 4 to 7, and in the youngest age group com-pared with the oldest age group at most scores (Fig. 1).Among the oldest individuals (65+ years), only 21 of 3,304

(0.6%) had alcohol dependence, and the highest probabilityat any score was < 20%.

Secondary Outcomes: Patterns of Drinking

Secondary analyses investigating patterns of drinkingacross AUDIT-C scores revealed that the maximum quantityof drinks consumed in a single day in the past year rangedfrom < 2 to 20 across increasing AUDIT-C scores (Fig. 2).The frequency of heavy drinking (i.e., drinking at levels thatexceed recommended maximum daily limits) ranged fromapproximately 0 d/wk at scores 1 to 2 to nearly 7 d/wk atscores 11 to 12 (Fig. 2). Furthermore, more than 90% ofindividuals with a score of 5 and all individuals with scores� 6 drank at this level at least once in the past year (Fig. 3).Exceeding recommended average daily limits was less com-mon, especially among men (Fig. 3). More than 50% ofwomen with scores � 4 and nearly 100% with scores � 8drank at levels that exceeded recommended average dailylimits. In contrast, fewer than 25% of men with scores 4 to 5drank at levels that exceeded average daily drinking limits,and only at scores � 10 did all men drink at these levels.

Both gender and age modified the association betweenAUDIT-C scores and each secondary outcome (p-values

Table 2. Probability (%) of Selected Sociodemographic Characteristics Among Past-Year Drinkers, by AUDIT-C Alcohol Screening Scorea

Total (alldrinkers)

AUDIT-C score

1 2 3 4 5 6 7 8 9 10 11 12

n % Weighted column%b

Men (n) 12,823 (2,994) (2,017) (1,936) (1,778) (973) (710) (709) (797) (263) (451) (82) (113)Age18–29 2,863 24 20 19 20 18 27 34 30 41 54 38 35 3730–44 4,449 34 31 36 33 31 39 36 38 37 28 35 25 3645–64 3,902 31 34 33 35 33 28 24 28 19 16 24 37 2565+ 1,609 11 14 13 11 19 6 5 4 4 2 3 3 2

Race/EthnicityWhite 7,855 74 72 72 75 76 75 77 73 75 68 73 74 79Black 1,821 9 9 10 10 9 7 6 9 9 6 8 7 10Hispanic/Latino 2,582 12 11 11 10 11 13 14 15 12 19 12 13 10Other 565 6 8 6 6 4 4 3 4 4 6 6 6 1

Married/widowed 7,160 62 70 69 66 69 59 52 52 41 34 38 40 36Family income <$20,000 2,465 17 16 15 13 15 15 17 22 23 29 30 35 43

Women (n) 13,723 (6,269) (2,692) (1,858) (1,406) (508) (297) (250) (254) (59) (90) (14) (26)Age18–29 3,198 24 22 22 22 22 42 42 35 48 66 41 47 3430–44 4,886 34 34 38 35 28 37 36 37 30 21 34 42 4745–64 3,923 30 32 30 32 34 17 19 27 20 13 21 5 1865+ 1,716 11 13 10 12 16 4 3 1 3 – 4 6 –

Race/EthnicityWhite 8,644 77 74 78 81 82 80 83 81 80 63 68 59 48Black 2,293 9 10 10 7 8 7 8 6 8 10 9 38 18Hispanic/Latino 2,289 9 11 8 8 6 10 6 8 8 18 13 2 9Other 497 4 5 4 4 3 3 4 5 4 9 10 – 24

Married/widowed 7,497 63 67 65 64 63 46 39 42 31 26 33 33 25Family income <$20,000 3,476 21 21 17 18 22 30 27 26 34 51 40 46 46

NESARC, National Epidemiologic Survey on Alcohol and Related Conditions; AUDIT-C, Alcohol Use Disorders Identification Test—Consumptionquestionnaire.

aBecause nondrinkers were not administered the AUDIT-C questions, the study sample includes no individuals with an AUDIT-C score of 0; associa-tions between all characteristics and AUDIT-C score were significant at the p < 0.00005 level.

bWeighted for NESARC sampling design and nonresponse.

AUDIT-C AS A SCALEDMARKEROF SEVERITY 5

< 0.00005). Maximum quantity of drinks increased more stee-ply across increasing scores among men than women, andamong younger versus older age groups. In contrast, thefrequency of heavy drinking increased more steeply acrossAUDIT-C scores among women than men, and among olderversus younger age groups. At a given score, women weregenerally more likely than men to exceed both recommendedmaximum and average daily drinking limits, whereasyounger individuals were more likely to exceed maximumdaily drinking limits, and older individuals were more likelyto exceed average daily drinking limits.

DISCUSSION

This study builds on extensive evidence that suggestsscores from the same brief alcohol screening questionnaireused for routine screening in primary care may also serve asan excellent marker of alcohol misuse severity. Previous stud-ies have demonstrated associations between AUDIT-Cscores and alcohol-related problems (Bradley et al., 2004)and probability of alcohol dependence (Rubinsky et al.,2010), but this is the first study to describe patterns ofconsumption across AUDIT-C scores and the first study toestimate AUD severity and the probability of alcoholdependence associated with each individual AUDIT-C score.Further, the large sample size of the present study facilitatedinvestigation of the impact of gender and age. Consistentwith prior research, alcohol consumption, AUD severity,and probability of dependence were found to increasedramatically as AUDIT-C scores increased. Across scores 1to 12, mean daily drinking ranged from < 0.1 to 18.0 drinks/d, AUD severity ranged from < 0.1 to 5.1 endorsed AUDcriteria, and the probability of alcohol dependence rangedfrom < 1 to 65%. However, at many scores, estimates werefound to differ by gender and age.

Mean daily drinking followed an exponential patternacross the range of AUDIT-C scores, increasing with a lowerslope across scores 1 to 7 and most steeply across scores 9 to12. These results could help inform the interpretation ofstudies that evaluate the risk of health outcomes acrossAUDIT-C scores. Health risks that increase exponentiallyacross AUDIT-C scores may be sensitive to increases inmean daily drinking in a dose–response manner, whereasrisks that increase more linearly or plateau at higherAUDIT-C scores may reflect a threshold effect as consump-tion increases. Additionally, findings extend the utility ofmeta-analyses that have compared the risk of disease acrossseveral levels of mean daily drinking (Corrao et al., 2004;Rehm et al., 2010a). Such studies have evaluated the risk ofseveral conditions, including cancers, hypertension, pneumo-nia, and injuries, at thresholds of � 2, 3.5, and 7 U.S. stan-dard-sized drinks per day versus < 2 drinks/d (Corrao et al.,2004; Rehm et al., 2010a). Findings of this study suggest thatthe risks described at these levels of consumption correspondto AUDIT-C scores � 7, 8, and 10 versus < 6, respectively.Healthcare systems that use electronic decision support

Table

3.MeanDaily

Drinking,AUDSeve

rity,andAlcoholD

ependence

Across

AUDIT-C

Sco

resAmongPast-YearDrinke

rs(N

=26,546)

Total

AUDIT-C

score

12

34

56

78

910

11

12

Meandrinks/d

a

(95%

CI)

1.1

(1.0,1.1)

0.0

b(0.0,0.0)

0.2

(0.2,0.2)

0.6

(0.6,0.6)

1.3

(1.2,1.3)

1.5

(1.4,1.6)

1.7

(1.6,1.7)

2.2

(2.0,2.3)

3.7

(3.5,3.9)

4.4

(4.0,4.8)

8.4

(7.9,8.9)

11.6

(10.8,12.5)

18.0

(16.9,19.0)

AUDse

verity

c

(95%

CI)

0.6

(0.5,0.6)

0.0

(0.0,0.0)

0.2

(0.2,0.2)

0.3

(0.3,0.4)

0.5

(0.5,0.6)

1.0

(0.9,1.1)

1.4

(1.3,1.5)

1.5

(1.3,1.6)

2.2

(2.1,2.4)

2.8

(2.4,3.1)

3.2

(2.9,3.5)

3.3

(2.6,4.0)

5.1

(4.3,5.8)

%Alcohol

dependentd

(95%

CI)

6(5,6)

0(0,0)

1(1,1)

2(1,2)

4(3,5)

9(7,10)

14(12,17)

16(13,19)

26(23,29)

35(29,41)

44(38,50)

39(26,51)

65(55,76)

AUD,alcoholuse

disorder;AUDIT-C

,AlcoholU

seDisorders

Identifica

tionTest—Consu

mptio

nquestionnaire;DSM-IV,DiagnosticandStatisticalM

anualofMentalD

isorders—FourthEdition.

aEstim

atesbase

donmaximum

from

(1)beve

rage-specific

consu

mptio

nquestionsor(2)questionsaboutallalcoholco

mbined;top-codedto

24drinks;andadjustedforsu

rveysa

mplingdesign

andnonresp

onse

.bEstim

atesof0.0

refle

ctva

lues<0.05;estim

atesof0%

refle

ctva

lues<0.5%.

cCountofendorsedDSM-IVAUDcrite

ria,rangingfrom

0to

11;estim

atesadjustedforsu

rveysa

mplingdesignandnonresp

onse

.dCurrentDSM-IValcoholdependence

;estim

atesadjustedforsu

rveysa

mplingdesignandnonresp

onse

.

6 RUBINSKY ET AL.

systems to prompt and score the AUDIT-C could integratethis information so that it would be provided automaticallyto the clinician alongside the patient’s screening score. Thisinformation could then be used by clinicians to estimate andcommunicate a patient’s alcohol-related health risks basedon their AUDIT-C score.

Gender modified the association of AUDIT-C scores withAUD severity, but not mean daily drinking or probability of

alcohol dependence. AUD severity was generally higher inwomen than men, with differences increasing across increas-ing AUDIT-C scores. This is consistent with previous find-ings that women develop AUD and alcohol-related medicalconditions such as liver disease and cardiomyopathy at lowerlevels of alcohol consumption than men (Fernandez-Solaand Nicolas-Arfelis, 2002; Hasin et al., 2007; Rehm et al.,2010b; Urbano-Marquez et al., 1995), as well as with U.S.

05

1015

20D

rinks

Per

Day

1 2 3 4 5 6 7 8 9 10 11 12AUDIT-C Score

MenWomen

By Gender

05

1015

20D

rinks

Per

Day

1 2 3 4 5 6 7 8 9 10 11 12AUDIT-C Score

18-2930-4445-6465+

By AgeMean Daily Drinking (with 95% CIs)

02

46

8M

ean

No.

of A

UD

Crit

eria

1 2 3 4 5 6 7 8 9 10 11 12AUDIT-C Score

MenWomen

By Gender

02

46

8M

ean

No.

of A

UD

Crit

eria

1 2 3 4 5 6 7 8 9 10 11 12AUDIT-C Score

18-2930-4445-6465+

By AgeAlcohol Use Disorder Severity (with 95% CIs)

0.2

.4.6

.81

Pro

babi

lity

of D

epen

denc

e

1 2 3 4 5 6 7 8 9 10 11 12AUDIT-C Score

MenWomen

By Gender

0.2

.4.6

.81

Pro

babi

lity

of D

epen

denc

e

1 2 3 4 5 6 7 8 9 10 11 12AUDIT-C Score

18-2930-4445-6465+

By AgeProbability of Alcohol Dependence (with 95% CIs)

Fig. 1. (Top) Mean daily drinking, (middle) alcohol use disorder severity, and (bottom) probability of alcohol dependence across AUDIT-C scores.AUDIT-C, Alcohol Use Disorders Identification Test—Consumption questionnaire.

AUDIT-C AS A SCALEDMARKEROF SEVERITY 7

drinking guidelines that recommend lower thresholds foridentifying risky drinking in women (National Institute onAlcohol Abuse and Alcoholism, 2007; U.S. Department ofAgriculture and U.S. Department of Health and HumanServices, 2010).

Although, at a given AUDIT-C score, women generallyexperienced greater AUD severity and were more likely tohave alcohol dependence compared with men, women wereless likely than men to have higher AUDIT-C scores. Incontrast, at a given AUDIT-C score, younger age groupsgenerally experienced greater AUD severity and higherprobability of dependence than older age groups and, inaddition, were more likely to have higher AUDIT-C scores.Younger individuals also consumed larger maximum quanti-ties of drinks and were more likely to exceed recommendedmaximum daily drinking limits compared with olderindividuals with the same AUDIT-C score. However, at agiven AUDIT-C score, older rather than younger age wasassociated with greater probability of exceeding recom-mended average daily limits and more frequent heavy drink-ing, suggesting that at a given score, older individuals may beengaging in more frequent drinking of less extreme quantitiescompared with younger individuals. This is consistent with a

recent study of heavy drinking among U.S. adults that foundthe highest prevalence and intensity among the youngestindividuals but the highest frequency among the oldestindividuals (MMWR, 2012).

The oldest age group (65+ years) endorsed far fewer AUDcriteria compared with younger groups and had relativelylow probability of dependence at every AUDIT-C score.This finding was not attributable to low mean daily drinkinggiven that this group drank as much as or more than youngerindividuals except at the highest scores (AUDIT-C 10 to 12),where they drank less than younger individuals but stillconsumed large quantities (7 to 12 drinks/d). One possibleexplanation is that the diagnostic criteria for DSM-IV AUDsmay be less sensitive to the alcohol-related sequelae of olderadults because of differences in life circumstances, physiol-ogy, and/or reporting. Several criteria for alcohol abuse anddependence ask about the impact of drinking on activitiesand responsibilities related to work and family, which maybe less applicable to older individuals. Further, the tendencyto develop alcohol tolerance diminishes with age (Milleret al., 1991), and endorsement of the tolerance criteriondecreased across increasing age groups. Variation across agegroups in the severity associated with expression of several

05

1015

2025

No.

of D

rinks

1 2 3 4 5 6 7 8 9 10 11 12AUDIT-C Score

MenWomen

By Gender

05

1015

2025

No.

of D

rinks

1 2 3 4 5 6 7 8 9 10 11 12AUDIT-C Score

18-2930-4445-6465+

By AgeMaximum Drinks Consumed in a Single Day in the Past Year (with 95% CIs)

02

46

8M

ean

Day

s pe

r Wee

k

1 2 3 4 5 6 7 8 9 10 11 12AUDIT-C Score

MenWomen

By Gender

02

46

8M

ean

Day

s pe

r Wee

k

1 2 3 4 5 6 7 8 9 10 11 12AUDIT-C Score

18-2930-4445-6465+

By AgeFrequency of Heavy Drinking (with 95% CIs)

Fig. 2. Maximum quantity of drinks and frequency of heavy drinking across AUDIT-C scores. AUDIT-C, Alcohol Use Disorders Identification Test—Consumption questionnaire.

8 RUBINSKY ET AL.

criteria has also been observed (Saha et al., 2006), suggestingage differences in either the experience or reporting of thesecriteria. Alternatively, differential survival favoring heavydrinkers who are less susceptible to the negative conse-quences of drinking could explain these findings. Finally,although the AUDIT-C has been found to perform similarlyfor identifying drinking above recommended limits, anyAUDs and alcohol dependence among older and youngerindividuals, at any given screening threshold, it is less sensi-

tive (and more specific) in older age groups (Dawson et al.,2005b).

A number of limitations of this study should be noted.The NESARC AUDIT-C differed from other versions(Dawson et al., 2005b) and followed a lengthy sequence ofquestions about consumption of specific alcoholic beveragetypes. Additionally, several other well-known alcohol screen-ing questionnaires, including the AUDIT, could not beevaluated using NESARC data. Because data are based on

0.5

1P

ropo

rtion

1 2 3 4 5 6 7 8 9 10 11 12AUDIT-C Score

MenWomen

By Gender

0.5

1P

ropo

rtion

1 2 3 4 5 6 7 8 9 10 11 12AUDIT-C Score

18-2930-4445-6465+

By AgeExceeded Recommended Maximum Daily Drinking Limits (with 95% CIs)

0.5

1P

ropo

rtion

1 2 3 4 5 6 7 8 9 10 11 12AUDIT-C Score

MenWomen

By Gender

0.5

1P

ropo

rtion

1 2 3 4 5 6 7 8 9 10 11 12AUDIT-C Score

18-2930-4445-6465+

By Age

Exceeded Recommended Average Daily Drinking Limits (with 95% CIs)

0.5

1P

ropo

rtion

1 2 3 4 5 6 7 8 9 10 11 12AUDIT-C Score

MenWomen

By Gender

0.5

1P

ropo

rtion

1 2 3 4 5 6 7 8 9 10 11 12AUDIT-C Score

18-2930-4445-6465+

By AgeExceeded Recommended Maximum and/or Average Daily Drinking Limits (with 95% CIs)

Fig. 3. Proportion of drinkers who exceeded recommended daily drinking limits across AUDIT-C scores. AUDIT-C, Alcohol Use Disorders Identifica-tion Test—Consumption questionnaire.

AUDIT-C AS A SCALEDMARKEROF SEVERITY 9

self-report, results are subject to a range of reporting biases,which could differ by gender and age. Estimates are also sub-ject to biases that might arise because the consumption out-comes and AUDIT-C score could have been determinedjointly to some degree in some cases. Despite the large sam-ple size, numbers of women and older individuals with highAUDIT-C scores were relatively small, resulting in wide con-fidence intervals for some estimates. Further, this study didnot have adequate numbers of individuals with highAUDIT-C scores to stratify analyses by groups defined byboth gender and age. Finally, study findings have practicalapplicability primarily in clinical settings, but the studysample was drawn from the general population. The generalpopulation could differ from clinical populations inimportant ways, although the screening performance of theAUDIT-C has been found to be similar in the NESARCsample and in clinical studies (Reinert and Allen, 2007).

This study also has several important strengths. TheNESARC’s unprecedentedly large sample size facilitated thisfirst-ever evaluation of the level of mean daily drinking,AUD severity, and probability of dependence associatedwith individual AUDIT-C scores. Further, because of thelarge sample size, analyses could be conducted separately inmen and women as well as across age groups. All outcomemeasures of alcohol consumption and AUDs were based onstructured in-depth interviews and have demonstratedreliability and predictive validity (Dawson et al., 2005a,2008, 2010, 2012; Grant et al., 2003a, 2007). Theconsumption outcomes were based on complex series ofquestions that made use of beverage-specific information onthe respondent’s typical drink size and the EtOH content oftheir main brand to achieve the most valid estimates ofalcohol intake (Dawson, 2003; Greenfield and Kerr, 2008).Additionally, the random sampling and high recruitmentrate by the NESARC limit bias in the study sample.

Results of this research could be used to improve assess-ment and management of the spectrum of alcohol misuse.Estimates, by gender and age, of the increasing level of con-sumption, AUD severity, and probability of dependenceassociated with each higher AUDIT-C score could be usedto assist healthcare providers in interpreting the implica-tions of a patient’s particular score. Specifically, healthcaresystems using electronic systems to guide screening andintervention for alcohol misuse could integrate this infor-mation so that it was provided to the clinician at the timeof screening. Information on risks for medical conditionsassociated with various levels of consumption could also beintegrated and provided alongside the patient’s screeningscore and associated level of consumption. Clinicians coulduse this information to offer patients personalized feedbackabout their risk, help engage the patient in ongoing dialog,and prompt diagnostic assessment (or referral for furtherassessment) with those who have scores associated withhigh AUD severity and/or likelihood of dependence.

To summarize, scores from the 3-item AUDIT-C alcoholscreening questionnaire captured meaningful variation in

several dimensions of consumption, as well as in AUD sever-ity and probability of dependence. Specifically, this study ofpast-year drinkers found that mean daily drinking increasedfrom < 0.1 to 18.0 drinks/d, AUD severity increased from< 0.1 to 5.1 criteria, and the probability of dependenceincreased from < 1 to 65% across the range of AUDIT-Cscores. Further, at many scores, patterns of drinking andseverity differed by gender and especially age. As healthcaresystems continue to implement alcohol screening usingelectronic decision support, it will become increasingly easierto tailor interventions based on the patient’s screeningscore and demographic characteristics. In this way, theinformation provided by this study could be used to informclinical care for the spectrum of alcohol misuse.

ACKNOWLEDGMENTS

The research reported here was supported by the U.S.Department of Veterans Affairs, Office of Research andDevelopment, Health Services Research and Development,and the Substance Use Disorders Quality EnhancementResearch Initiative. ADR was also supported by an Agencyfor Healthcare Research and Quality (AHRQ) NationalResearch Services Award (NRSA) at the University ofWash-ington (T32 HS013853). The views expressed in this articleare those of the authors and do not necessarily reflect the posi-tion or policy of the Department of Veterans Affairs, the Uni-ted StatesGovernment, or any of the authors’ institutions.

REFERENCES

American Psychiatric Association (1994)Diagnostic and StatisticalManual of

Mental Disorders. 4th ed. American Psychiatric Association, Washington,

DC.

Babor TF, De La Fuente JR, Saunders JB, Grant M (1989) AUDIT: The

Alcohol Use Disorders Identification Test. Guidelines for Use in Primary

Health Care. World Health Organization, Geneva.

Bradley KA, Kivlahan DR, Williams EC (2009) Brief approaches to alcohol

screening: practical alternatives for primary care. J Gen Intern Med 24:

881–883.Bradley KA, Kivlahan DR, Zhou XH, Sporleder JL, Epler AJ, McCormick

KA, Merrill JO, McDonell MB, Fihn SD (2004) Using alcohol screening

results and treatment history to assess the severity of at-risk drinking in

Veterans Affairs primary care patients. Alcohol Clin Exp Res 28:

448–455.Bradley KA, McDonell MB, Bush K, Kivlahan DR, Diehr P, Fihn SD

(1998) The AUDIT alcohol consumption questions: reliability, validity,

and responsiveness to change in older male primary care patients. Alcohol

Clin Exp Res 22:1842–1849.Bradley KA, Rubinsky AD, Sun H, Bryson CL, Bishop MJ, Blough DK,

Henderson WG, Maynard C, Hawn MT, Tonnesen H, Hughes G, Beste

LA, Harris AH, Hawkins EJ, Houston TK, Kivlahan DR (2011)

Alcohol screening and risk of postoperative complications in male VA

patients undergoing major non-cardiac surgery. J Gen Intern Med

26:162–169.Bradley KA, Williams EC, Achtmeyer CE, Volpp B, Collins BJ, Kivlahan

DR (2006) Implementation of evidence-based alcohol screening in the

Veterans Health Administration. Am JManag Care 12:597–606.Bush K, Kivlahan DR, McDonell MB, Fihn SD, Bradley KA (1998) The

AUDIT alcohol consumption questions (AUDIT-C): an effective brief

10 RUBINSKY ET AL.

screening test for problem drinking. Ambulatory Care Quality Improve-

ment Project (ACQUIP). Alcohol Use Disorders Identification Test. Arch

InternMed 158:1789–1795.Corrao G, Bagnardi V, Zambon A, La Vecchia C (2004) A meta-analysis of

alcohol consumption and the risk of 15 diseases. PrevMed 38:613–619.Dawson DA (1998) Volume of ethanol consumption: effects of different

approaches to measurement. J Stud Alcohol 59:191–197.Dawson DA (2003) Methodological issues in measuring alcohol use. Alcohol

Res Health 27:18–29.Dawson DA, Grant BF, Li TK (2005a) Quantifying the risks associated with

exceeding recommended drinking limits. Alcohol Clin Exp Res 29:902–908.

Dawson DA, Grant BF, Stinson FS, Zhou Y (2005b) Effectiveness of the

derived Alcohol Use Disorders Identification Test (AUDIT-C) in

screening for alcohol use disorders and risk drinking in the US general

population. Alcohol Clin Exp Res 29:844–854.Dawson DA, Li TK, Grant BF (2008) A prospective study of risk drinking:

at risk for what? Drug Alcohol Depend 95:62–72.Dawson DA, Saha TD, Grant BF (2010) A multidimensional assessment of

the validity and utility of alcohol use disorder severity as determined by

item response theory models. Drug Alcohol Depend 107:31–38.Dawson DA, Smith SM, Pickering RP, Grant BF (2012) An empirical

approach to evaluating the validity of alternative low-risk drinking guide-

lines. Drug Alcohol Rev 31:141–150.Fernandez-Sola J, Nicolas-Arfelis JM (2002) Gender differences in alcoholic

cardiomyopathy. J Gend Specif Med 5:41–47.Grant BF, Dawson DA (2006) Introduction to the National Epidemiologic

Survey on Alcohol and Related Conditions. Alcohol Res Health 29:74–78.Grant BF, Dawson DA, Hasin DS (2001) The Alcohol Use Disorder and

Associated Disabilities Interview Schedule—DSM-IV Version. National

Institute on Alcohol Abuse and Alcoholism, Bethesda, MD.

Grant BF, Dawson DA, Stinson FS, Chou PS, Kay W, Pickering R (2003a)

The Alcohol Use Disorder and Associated Disabilities Interview Schedule-

IV (AUDADIS-IV): reliability of alcohol consumption, tobacco use, fam-

ily history of depression and psychiatric diagnostic modules in a general

population sample. Drug Alcohol Depend 71:7–16.Grant BF, Harford TC, Muthen BO, Yi HY, Hasin DS, Stinson FS (2007)

DSM-IV alcohol dependence and abuse: further evidence of validity in the

general population. Drug Alcohol Depend 86:154–166.Grant BF, Kaplan K, Shepard J, Moore T (2003b) Source and Accuracy

Statement for Wave 1 of the 2001–2002 National Epidemiologic Survey on

Alcohol and Related Conditions. NIAAA, Bethesda,MD.

Greenfield TK, Kerr WC (2008) Alcohol measurement methodology in

epidemiology: recent advances and opportunities. Addiction 103:1082–1099.Greenfield TK, Nayak MB, Bond J, Ye Y, Midanik LT (2006) Maximum

quantity consumed and alcohol-related problems: assessing the most alco-

hol drunk with two measures. Alcohol Clin Exp Res 30:1576–1582.Harris AH, Bradley KA, Bowe T, Henderson P,Moos R (2010) Associations

between AUDIT-C and mortality vary by age and sex. Popul Health

Manag 13:263–268.Hasin DS, Stinson FS, Ogburn E, Grant BF (2007) Prevalence, correlates,

disability, and comorbidity of DSM-IV alcohol abuse and dependence in

the United States: results from the National Epidemiologic Survey on

Alcohol and Related Conditions. Arch Gen Psychiatry 64:830–842.Jonas DE, Garbutt JC, Amick HR, Brown JM, Brownley KA, Council CL,

Viera AJ, Wilkins TM, Schwartz CJ, Richmond EM, Yeatts J, Evans TS,

Wood SD, Harris RP (2012) Behavioral counseling after screening for

alcohol misuse in primary care: a systematic review and meta-analysis for

the U.S. Preventive Services Task Force. Ann InternMed 157:645–654.Kirchner JE, Zubritsky C, Cody M, Coakley E, Chen H, Ware JH, Oslin

DW, Sanchez HA, Durai UN, Miles KM, Llorente MD, Costantino G,

Levkoff S (2007) Alcohol consumption among older adults in primary

care. J Gen InternMed 22:92–97.Kriston L, H€olzel L, Weiser AK, Berner MM, H€arter M (2008) Meta-

analysis: are 3 questions enough to detect unhealthy alcohol use? Ann

InternMed 149:879–888.

Lang I, Guralnik J, Wallace RB, Melzer D (2007) What level of alcohol con-

sumption is hazardous for older people? Functioning and mortality in

U.S. and English national cohorts. J AmGeriatr Soc 55:49–57.Lembke A, Bradley KA, Henderson P, Moos R, Harris AH (2011) Alcohol

screening scores and the risk of new-onset gastrointestinal illness or related

hospitalization. J Gen InternMed 26:777–782.Miller NS, Belkin BM, Gold MS (1991) Alcohol and drug dependence

among the elderly: epidemiology, diagnosis, and treatment. Compr Psychi-

atry 32:153–165.MMWR (2012) Vital signs: binge drinking prevalence, frequency, and inten-

sity among adults—United States, 2010. MMWR Morb Mortal Wkly

Rep 61:14–19.National Institute onAlcohol Abuse andAlcoholism (2007) Helping Patients

Who Drink Too Much: A Clinician’s Guide (Updated 2005 Edition).

National Institutes of Health, U.S. Department of Health and Human

Services, Washington, DC.

National Institute on Alcohol Abuse and Alcoholism (NIAAA) (2003) Rec-

ommended Sets of Alcohol Consumption Questions, NIAAA Task Force

on Recommended Questions of the National Council on Alcohol Abuse

and Alcoholism. NIAAA, Bethesda,MD.

National Institute on Alcohol Abuse and Alcoholism (NIAAA) (2004) Wave

1 NESARCData Notes. NIAAA, Bethesda,MD.

Nilsen P (2010) Brief alcohol intervention—where to from here? Challenges

remain for research and practice. Addiction 105:954–959.Rehm J, Baliunas D, Borges GL, Graham K, Irving H, Kehoe T, Parry

CD, Patra J, Popova S, Poznyak V, Roerecke M, Room R, Samokhva-

lov AV, Taylor B (2010a) The relation between different dimensions of

alcohol consumption and burden of disease: an overview. Addiction

105:817–843.Rehm J, Taylor B, Mohapatra S, Irving H, Baliunas D, Patra J, Roerecke M

(2010b) Alcohol as a risk factor for liver cirrhosis: a systematic review and

meta-analysis. Drug Alcohol Rev 29:437–445.Reinert DF, Allen JP (2007) The alcohol use disorders identification test: an

update of research findings. Alcohol Clin Exp Res 31:185–199.Rubinsky AD, Kivlahan DR, Volk RJ, Maynard C, Bradley KA (2010) Esti-

mating risk of alcohol dependence using alcohol screening scores. Drug

Alcohol Depend 108:29–36.Saha TD, Chou SP, Grant BF (2006) Toward an alcohol use disorder contin-

uum using item response theory: results from the National Epidemiologic

Survey on Alcohol and Related Conditions. PsycholMed 36:931–941.Saha TD, Stinson FS, Grant BF (2007) The role of alcohol consumption in

future classifications of alcohol use disorders. Drug Alcohol Depend 89:

82–92.Solberg LI, Maciosek MV, Edwards NM (2008) Primary care intervention

to reduce alcohol misuse ranking its health impact and cost effectiveness.

Am J PrevMed 34:143–152.StataCorp (2011) Stata Statistical Software: Release 12. StataCorp LP, Col-

lege Station, TX.

Urbano-Marquez A, Estruch R, Fernandez-Sola J, Nicolas JM, Pare JC,

Rubin E (1995) The greater risk of alcoholic cardiomyopathy and myopa-

thy in women compared with men. JAMA 274:149–154.U.S. Department of Agriculture andU.S. Department of Health and Human

Services (2010) Dietary Guidelines for Americans, 2010. 7th ed. U.S. Gov-

ernment Printing Office, Washington, DC.

Willenbring ML (2009) Treatment of heavy drinking and alcohol use disor-

ders, in Principles of Addiction Medicine, 4th ed (Ries RK, Fiellin DA,

Miller SC, Saitz R, eds), pp 335–348. Lippincott Williams &Wilkins, Phil-

adelphia, PA.

Williams EC, Bryson CL, Sun H, Chew RB, Chew LD, Blough DK, Au

DH, Bradley KA (2012) Association between alcohol screening results and

hospitalizations for trauma in Veterans Affairs outpatients. Am J Drug

Alcohol Abuse 38:73–80.Williams EC, Johnson ML, Lapham GT, Caldeiro RM, Chew L, Fletcher

GS, McCormick KA, Weppner WG, Bradley KA (2011) Strategies to

implement alcohol screening and brief intervention in primary care

settings: a structured literature review. Psychol Addict Behav 25:206–214.

AUDIT-C AS A SCALEDMARKEROF SEVERITY 11