Skip to main content
PMC home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2013 Dec 1.
Published in final edited form as: Drug Alcohol Depend. 2012 Jun 22;126(3):384–388. doi: 10.1016/j.drugalcdep.2012.05.029

Comparative Performance of the AUDIT-C in Screening for DSM-IV and DSM-5 Alcohol Use Disorders*

Deborah A Dawson a,b, Sharon M Smith a, Tulshi D Saha a, Anna D Rubinsky c,d, Bridget F Grant a
PMCID: PMC3469743  NIHMSID: NIHMS389692  PMID: 22728044

Abstract

Objective

Under the proposed DSM-5 revision to the criteria for alcohol use disorder (AUD), a substantial proportion of DSM-IV AUD cases will be lost or shifted in terms of severity, with some new cases added. Accordingly, the performance of the AUDIT-C in screening for DSM-IV AUD cannot be assumed to extend to DSM-5 AUD. The objective of this paper is to compare the AUDIT-C in screening for DSM-IV and DSM-5 AUD.

Methods

Using a broad range of performance metrics, the AUDIT-C was tested and contrasted as a screener for DSM-IV AUD (any AUD, abuse and dependence) and DSM-5 AUD (any AUD, moderate AUD and severe AUD) in a representative sample of U.S. adults aged 21 and older and among past-year drinkers.

Results

Optimal AUDIT-C cutpoints were identical for DSM-IV and DSM-5 AUD: ≥4 for any AUD, ≥3 or ≥4 for abuse/moderate AUD and ≥4 or ≥5 for dependence/severe AUD. Screening performance was slightly better for DSM-5 severe AUD than DSM-IV dependence but did not differ for other diagnoses. At optimal screening cutpoints, positive predictive values were slightly higher for DSM-5 overall AUD and moderate AUD than for their DSM-IV counterparts. Sensitivities were slightly higher for DSM-5 severe AUD than DSM-IV dependence. Optimal screening cutpoints shifted upwards for past-year drinkers but continued to be identical for DSM-IV and DSM-5 disorders.

Conclusions

Clinicians should not face any major overhaul of their current screening procedures as a result of the DSM-5 revision and should benefit from fewer false positive screening results.

Keywords: AUDIT-C, screening, alcohol use disorder, DSM-IV, DSM-5

1. Introduction

The AUDIT-C, comprising the first three questions of the Alcohol Use Disorders Identification Test (Saunders et al., 1993), is a brief screener for alcohol use disorder (AUD) commonly used in general medical and emergency department settings. Its screening utility has been demonstrated among patients in Veterans Affairs (Bush et al., 1998; Bradley et al., 2003), primary care (Aertgeerts et al., 2001; Gual et al., 2002; Gordon et al., 2001; Bradley et al., 2007), emergency department (Kelly et al., 2009), occupational care (Kaarne et al., 2010), and prenatal (Burns et al., 2010) settings, as well as in the general population (Dawson et al., 2005a, 2005b; Nayak et al., 2009; Rumpf et al., 2002). It also shows promise for telephone screening (McPherson et al., 2010), and its screening performance approaches that of the full AUDIT (Aeertgeerts et al., 2001; Gordon et al., 2001; Nayak et al., 2009; Rumpf et al., 2002).

No study to date has evaluated the AUDIT-C with respect to the proposed DSM-5 criteria for AUD (http://www.dsm5.org), which differ in several ways from the DSM-IV criteria (American Psychiatric Association, 1994). Under the proposed revision, the criterion of alcohol-related legal problems has been dropped, and a new craving criterion has been added. DSM-IV abuse and dependence are based on separate sets of diagnostic criteria, with endorsement of ≥1 of four abuse criteria required for an abuse diagnosis and ≥3 of seven dependence criteria required for a dependence diagnosis. In contrast, all 11 DSM-5 criteria apply towards a unitary construct of AUD, with endorsement of 2–3 criteria representing moderate AUD and ≥4 criteria representing severe AUD. Accordingly, cases of AUD have been gained, lost and shifted in terms of severity (Agrawal et al., 2011; Dawson et al, under review). As per Supplementary Table 11, 3.3% of individuals negative for any DSM-IV AUD qualify for DSM-5 moderate AUD; these comprise former diagnostic orphans (Hasin and Paykin, 1998) who endorsed two of the DSM-IV dependence criteria (or one plus craving). Only 58% of individuals with DSM-IV abuse are classified with the corresponding DSM-5 diagnosis of moderate AUD; 36% who endorsed a single DSM-IV abuse criterion (or two including legal) do not qualify for any DSM-5 AUD, and 6.0% whose DSM-5 criteria count reaches 4+ by combining former abuse and dependence criteria and craving qualify for DSM-5 severe AUD. Finally, 19.5% of individuals with DSM-IV dependence, those with just three dependence criteria, no abuse criteria (or legal only) and no craving, are downshifted into DSM-5 moderate rather than severe AUD. As a result, the performance of the AUDIT-C in screening for DSM-IV AUD cannot be assumed to extend to DSM-5 diagnoses.

In light of the widespread use of the AUDIT-C, it is of critical importance to clinicians to document its screening ability for DSM-5 AUD. This analysis was designed to provide this missing information. It utilizes data from a representative sample of the U.S. adult population to compare the relationship of the AUDIT-C with DSM-IV and DSM-5 AUD in the total population and among past-year drinkers who consumed any alcohol in the year preceding alcohol screening.

2. Methods

2.1 Sample

This study uses data from Wave 2 of the National Epidemiologic Survey on Alcohol and Related Conditions (NESARC), the 3-year follow-up of a nationally representative sample of U.S. adults 18 and older living in households and noninstitutional group quarters, originally interviewed in 2001–2002. At the 2004–2005 Wave 2 follow-up, 34,653 of the 43,093 Wave 1 respondents were reinterviewed, 86.7% of those eligible for reinterview, for a cumulative response rate of 70.2% (Grant et al., 2003a, 2007). Informed consent was obtained after potential respondents were informed in writing about the nature of the survey, uses of the survey data, voluntary nature of participation and confidentiality of identifiable survey information. The research protocol received full ethical review and approval.

The overlap of NESARC respondents positive for both DSM-IV and DSM-5 AUD necessitated a split-sample design to compare AUDIT-C screening performance for DSM-IV and DSM-5 AUD using statistical tests designed for independent samples. After excluding individuals with missing data for AUDIT-C items (<0.5% of the total population), the screener was evaluated with respect to DSM-IV AUD among respondents with even case identification numbers (n = 17,225 including 10,944 past-year drinkers) and DSM-5 AUD among respondents with odd case identification numbers (n = 17,311 including 11,116 past-year drinkers).

2.2 Measures

All AUD diagnoses were based on the Alcohol Use Disorder and Associated Disabilities Interview Schedule (Grant et al., 2001). Past-year DSM-IV dependence required endorsement of ≥3 of seven dependence criteria (tolerance, withdrawal/relief of withdrawal, drinking more/longer than intended, persistent desire/unsuccessful attempts to reduce drinking, excessive time spent drinking, important activities given up, continued drinking despite physical/psychological problems); abuse required endorsement of ≥1 of four abuse criteria (failure to fulfill role obligations, recurrent hazardous drinking, recurrent alcohol-related legal problems, continued drinking despite interpersonal problems) and the absence of dependence. Any DSM-IV AUD refers to abuse or dependence. The DSM-IV AUD diagnoses are highly reliable, kappa = .74 for past-year AUD (Grant et al., 2003b). The 11 DSM-5 criteria for AUD consisted of craving and all of the DSM-IV abuse and dependence criteria listed above other than legal problems. Past-year DSM-5 moderate AUD required endorsement of 2–3 criteria; severe AUD required ≥4 criteria. Any DSM-5 AUD includes moderate and severe AUD.

The AUDIT-C score ( 0 to 12 in the total population and 1 to 12 among past-year drinkers) was calculated from responses to three questions regarding past-year consumption for all types of alcohol combined: 1) overall frequency of drinking, 2) usual quantity of drinks and 3) frequency of drinking 5+ drinks. The scoring of these questions and their differences relative to standard AUDIT-C items are presented in detail elsewhere (Dawson et al., 2005b). Briefly, although the AUDIT-C scoring conventions were followed as closely as possible, a few differences in the NESARC version were unavoidable: a) the NESARC identified nondrinkers from a separate screening question rather than a “never” response to question 1; b) the NESARC offered more frequency response categories and ordered them with higher frequencies at the top rather than bottom of the list; c) the NESARC’s combination of “3 to 4 times a week” in a single category meant that individuals drinking four times a week got a score of 3 rather than 4 points for question 1; and d) question 3 asked frequency of drinking 5+ rather than 6+ drinks.

2.3 Analysis

Sensitivity, specificity and positive predictive values (PPV) were based on weighted crosstabulations of diagnoses with AUDIT-C score cutpoints (≥3, ≥4, etc.) using SUDAAN (Research Triangle Institute, 2008), which uses Taylor series linearization to adjust variance estimates for complex, multistage sample designs. At any given cutpoint, sensitivity is the proportion of individuals with positive outcomes whose screener scores were ≥ the cutpoint, and specificity is the proportion of individuals with negative outcomes whose screener scores were < the cutpoint. PPV is the proportion of individuals with screener scores ≥ the cutpoint who had positive outcomes. Positive likelihood ratios (PLR), i.e., sensitivity divided by 1 minus specificity, were calculated for each cutpoint (Simel et al., (1991). SAS Version 9.2 (SAS Institute, Inc., 2008) was used to estimate areas under the curve (AUC) and standard errors for plots of sensitivity versus 1 minus specificity at each screener cutpoint (Swets and Pickett, 1982). Differences in sensitivity, specificity, PPV and AUC were tested for significance (p<.05 and p<.005) using t-tests for independent samples; since the confidence intervals for PLRs are symmetrical around their natural logs, we used t-tests for differences of their logged values (Simel et al., 1991).

3. Results

AUDIT-C screening performance was similar for DSM-IV and DSM-5 AUD in the total population. A cutpoint of ≥4 optimized the combined values of sensitivity and specificity for the diagnosis of any AUD under both the DSM-IV and DSM-5, and AUC values were virtually identical, 0.914 and 0.915 (left panel of Table 1). However, at this optimal cutpoint, the PPV was significantly higher for DSM-5 than DSM-IV AUD, 36.9%, versus 33.7%. When comparing DSM-IV abuse and DSM-5 moderate AUD (middle panel ), a cutpoint of ≥3 maximized combined sensitivity and specificity for both diagnoses, but at specificity levels that fell short of 70%; thus a cutpoint of ≥4 might be considered preferable. AUC values were comparable, 0.870 and 0.871, but PPV were greater with respect to the DSM-5 diagnosis at cutpoints of both ≥3 and ≥4. For DSM-IV dependence and DSM-5 severe AUD (right panel), cutpoints of ≥4 and ≥5 both performed well, the former favoring sensitivity and the latter favoring specificity. At both cutpoints, sensitivity was significantly higher for severe AUD than dependence; however, at the lower cutpoint of ≥4, the PPV was higher for dependence. The overall screening performance was superior for severe AUD, with an AUC of 0.935 compared to 0.923 for dependence.

Table 1.

Performance of the AUDIT-C in screening for DSM-IV and DSM-5 alcohol use disorders: U.S. adults 21 years of age and older

Any DSM-IV AUD DSM-IV Abuse DSM-IV Dependence
% Sens. % Spec. %PPVa PLRb % Sens. % Spec. %PPVa PLRb % Sens. % Spec. %PPVa PLRb
N of cases 1,534 15,691 Variablec 17,225 817 16,408 Variablec 17,225 717 16,508 Variablec 17,225
AUDIT-C score
 ≤1 100.0 (0.0) 37.8 (0.7) 14.7 (0.4) 1.6 100.0 (0.0) 36.0 (0.7) 7.8 (0.3) 1.6 100.0 (0.0) 35.8 (0.7)  6.9 (0.3) 1.6
 ≥2 98.2 (0.4) 60.5 (0.7) 21.0 (0.6) 2.5 97.3 (0.7) 57.6 (0.6) 11.1 (0.5) 2.3 99.1 (0.3) 57.4 (0.7)  9.9 (0.5) 2.3
 ≥3 92.8 (0.8) 72.2 (0.6) 26.4 (0.8) 3.3 90.4 (1.3) 69.0 (0.6) 13.7 (0.6) 2.9 95.4 (0.9) 68.8 (0.6) 12.7 (0.6) 3.1
 ≥4 82.6 (1.1) 82.6 (0.4) 33.7 (1.0) 4.7 77.7 (1.7) 79.2 (0.5) 16.9 (0.7) 3.7 88.3 (1.4) 79.3 (0.5) 16.8 (0.8) 4.3
 ≥5 70.2 (1.3) 90.8 (0.3) 45.0 (1.3) 7.6 62.7 (1.8) 87.5 (0.4) 21.4 (1.0) 5.0 78.9 (1.9) 87.9 (0.4) 23.6 (1.2) 6.5
 ≥6 59.0 (1.5) 94.0 (0.3) 51.3 (1.5) 9.8 49.9 (1.9) 91.0 (0.3) 23.1 (1.2) 5.5 69.3 (2.1) 91.6 (0.3) 28.2 (1.3) 8.3
 ≥7 45.1 (1.6) 96.0 (0.2) 54.7 (1.7) 11.3 35.0 (2.1) 93.5 (0.3) 22.6 (1.5) 5.4 56.5 (2.4) 94.3 (0.3) 32.0 (1.6) 9.9
 ≥8 33.8 (1.5) 97.8 (0.2) 62.4 (2.1) 15.4 24.5 (1.9) 95.8 (0.2) 24.0 (1.9) 5.8 44.5 (2.4) 96.6 (0.2) 38.4 (2.0) 13.1
 ≥9 18.0 (1.2) 99.2 (0.1) 69.4 (3.0) 22.5 9.6 (1.3) 97.9 (0.2) 19.8 (2.5) 4.6 27.5 (2.1) 98.7 (0.1) 49.6 (3.1) 21.2
 ≥10 13.7 (1.2) 99.4 (0.1) 70.8 (3.5) 22.8 6.8 (1.2) 98.4 (0.1) 18.7 (2.8) 4.3 21.5 (2.1) 99.1 (0.1) 52.1 (3.6) 23.9
 ≥11 4.1 (0.6) 99.9 (0.0) 82.2 (5.4) 41.0 1.5 (0.5) 99.6 (0.1) 15.8 (5.1) 3.8 7.1 (1.2) 99.8 (0.0) 66.5 (6.9) 35.5
AUC .914 (.003) .870 (.004) .923 (.004)
Any DSM-5 AUD DSM-5 Moderate AUD DSM-5 Severe AUD
% Sens. % Spec. %PPVa PLRb % Sens. % Spec. % PPVa PLRb % Sens. % Spec. % PPVa PLRb
N of cases 1,776 15,535 Variablec 17,311 1,149 16,162 Variablec 17,311 627 16,684 Variablec 17,311
AUDIT-C score
 ≤1 100.0 (0.0) 37.8 (0.8) 16.3 (0.5)* 1.6 100.0 (0.0) 36.3 (0.8) 10.7 (0.4)** 1.6 100.0 (0.0) 35.0 (0.7)   5.7 (0.3)** 1.5
 ≥2 98.4 (0.4) 60.1 (0.7) 23.0 (0.7)* 2.5 97.6 (0.6) 57.7 (0.7) 14.9 (0.5)** 2.3 99.8 (0.2) 55.9 (0.7)   8.1 (0.4)** 2.3
 ≥3 93.3 (0.8) 72.1 (0.6) 28.9 (0.8)* 3.3 90.7 (1.1) 69.3 (0.6) 18.3 (0.7)** 3.0 98.4 (0.5)** 67.5 (0.6) 10.6 (0.5)* 3.0
 ≥4 84.3 (1.1) 82.5 (0.4) 36.9 (1.0)* 4.8 78.9 (1.4) 79.4 (0.5) 22.6 (0.8)** 3.8 94.6 (1.0)** 78.0 (0.5) 14.3 (0.7)* 4.3
 ≥5 69.1 (1.3) 91.2 (0.3) 48.7 (1.3)* 7.9 60.9 (1.7) 88.1 (0.3) 28.0 (1.1)** 5.1 84.6 (1.9)* 87.3 (0.3) 20.7 (1.1) 6.7
 ≥6 56.0 (1.4) 94.1 (0.2) 53.7 (1.5) 9.5 46.4 (1.8) 91.4 (0.3) 29.1 (1.3)** 5.4 74.0 (2.1) 91.2 (0.3) 24.6 (1.3) 8.5
 ≥7 45.1 (1.4) 96.2 (0.2) 59.1 (1.8) 11.9 34.8 (1.7) 93.8 (0.3) 29.8 (1.6)** 5.6 64.6 (2.3)* 93.9 (0.3) 29.3 (1.6) 10.6
 ≥8 33.1 (1.4) 97.7 (0.2) 63.1 (2.0) 14.4 23.9 (1.6) 95.7 (0.2) 29.9 (2.0)* 5.6 50.3 (2.3) 96.1 (0.2) 33.3 (1.9) 13.4
 ≥9 16.9 (1.0) 99.1 (0.1) 69.7 (3.0) 18.8** 9.7 (1.2) 97.9 (0.2) 26.1 (2.9) 4.6 30.4 (2.0) 98.5 (0.1) 43.6 (3.2) 21.6
 ≥10 13.1 (0.9) 99.5 (0.1) 75.5 (3.5) 26.2* 6.8 (1.0) 98.5 (0.1) 25.6 (3.4) 4.5 25.0 (1.9) 99.0 (0.1) 49.9 (4.0) 24.1
 ≥11 4.7 (0.6) 99.9 (0.0) 87.0 (5.8) 47.0 2.2 (0.6) 99.5 (0.1) 26.8 (6.4) 4.4 9.5 (1.40 99.8 (0.1) 60.2 (7.3) 40.5
AUC .915 (.003) .871 (.004) .935 (.004)*
Open in a new tab

Note: Figures in parentheses are standard errors of estimates. Values of 0.0 denote standard errors <.05. Shaded cells denote recommended cutpoints for outcome in question.

**

Significantly different (p<.005) from corresponding value for DSM-IV diagnosis

*

Significantly different (p<.05) from corresponding value for DSM-IV diagnosis

a

Positive predictive value

b

Positive likelihood ratio

c

N differs for each row, based on n of cases with AUDIT-C score in question (decreasing from 10,994 for scores ≤1 to 73 for scores ≥11)

Among past-year drinkers, specificity and AUC values for the AUDIT-C declined substantially relative to the total population (Table 2). Since AUD diagnoses and positive screens are limited to past-year drinkers regardless of whether assessed in the total population or not, there was no change in sensitivity and PPV. For any DSM-IV and DSM-5 AUD (left panel), cutpoints of ≥4 and ≥5 yielded similar combined sensitivity and specificity, the former favoring sensitivity and latter favoring specificity. These cutpoints also were optimal for DSM-IV abuse and DSM-5 moderate AUD. For DSM-IV dependence and DSM-5 severe AUD, a cutpoint of ≥5 optimized combined sensitivity and specificity. As in the total population, the AUC was significantly greater for DSM-5 severe AUD than DSM-IV dependence, 0.896 versus 0.876.

Table 2.

Performance of the AUDIT-C in screening for DSM-IV and DSM-5 alcohol use disorders: Past-year drinkers 21 years of age and older

Any DSM-IV AUD DSM-IV Abuse DSM-IV Dependence
% Sens. % Spec. %PPVa PLRb % Sens. % Spec. %PPVa PLRb % Sens. % Spec. %PPVa PLRb
N of cases 1,534 9,410 Variablec 10,944 817 10,127 Variablec 10,944 717 10,227 Variablec 10,944
AUDIT-C
score
 ≤1 100.0 (0.0) 0.0 (0.0) 14.7 (0.4) 1.0 100.0 (0.0) 0.0 (0.0) 7.8 (0.3) 1.0 100.0 (0.0) 0.0 (0.0) 6.9 (0.3) 1.0
 ≥2 98.2 (0.4) 36.5 (0.7) 21.0 (0.6) 1.5 97.3 (0.7) 33.8 (0.6) 11.1 (0.5) 1.5 99.1 (0.3) 33.6 (0.6) 9.9 (0.5) 1.5
 ≥3 92.8 (0.8) 55.3 (0.7) 26.4 (0.8) 2.1 90.4 (1.3) 51.5 (0.7) 13.7 (0.6) 1.9 95.4 (0.9) 51.5 (0.7) 12.7 (0.6) 2.0
 ≥4 82.6 (1.1) 72.0 (0.6) 33.7 (1.0) 3.0 77.7 (1.7) 67.5 (0.6) 16.9 (0.7) 2.4 88.3 (1.4) 67.8 (0.6) 16.8 (0.6) 2.7
 ≥5 70.2 (1.3) 85.2 (0.5) 45.0 (1.3) 4.7 62.7 (1.8) 80.4 (0.6) 21.4 (1.0) 3.2 78.9 (1.9) 81.2 (0.6) 23.6 (1.2) 4.2
 ≥6 59.0 (1.5) 90.4 (0.4) 51.3 (1.5) 6.1 49.9 (1.9) 85.9 (0.5) 23.1 (1.2) 3.5 69.3 (2.1) 87.0 (0.5) 28.2 (1.3) 5.3
 ≥7 45.1 (1.6) 93.6 (0.3) 54.7 (1.7) 7.0 35.0 (2.1) 89.8 (0.4) 22.6 (1.5) 3.4 56.5 (2.4) 91.2 (0.4) 32.0 (1.6) 6.4
 ≥8 33.8 (1.5) 96.5 (0.3) 62.4 (2.1) 9.7 24.5 (1.9) 93.4 (0.4) 24.0 (1.9) 3.7 44.5 (2.4) 94.7 (0.3) 38.4 (2.0) 8.4
 ≥9 18.0 (1.2) 98.6 (0.2) 69.4 (3.0) 12.9 9.6 (1.3) 96.7 (0.2) 19.8 (2.5) 2.9 27.5 (2.1) 97.9 (0.2) 49.6 (3.1) 13.1
 ≥10 13.7 (1.2) 99.0 (0.1) 70.8 (3.5) 13.7 6.8 (1.2) 97.5 (0.2) 18.7 (2.8) 2.7 21.5 (2.1) 98.5 (0.2) 52.1 (3.6) 14.3
 ≥11 4.1 (0.6) 99.8 (0.1) 82.2 (5.4) 20.5 1.5 (0.5) 99.3 (0.1) 15.8 (5.1) 2.1 7.1 (1.2) 99.7 (0.1) 66.5 (6.9) 23.7
AUC 0.857 (.005) 0.789 (.007) 0.876 (.006)
Any DSM-5 AUD DSM-5 Moderate AUD DSM-5 Severe AUD
% Sens. % Spec. %PPVa PLRb % Sens. % Spec. %PPVa PLRb % Sens. % Spec. %PPVa PLRb
N of cases 1,776 9,340 Variablec 11,116 1,149 9,967 Variablec 11,116 627 10,489 Variablec 11,116
AUDIT-C score
 ≤1 100.0 (0.0) 0.0 (0.0) 16.3 (0.5)* 1.0 100.0 (0.0) 0.0 (0.0) 10.7 (0.4)** 1.0 100.0 (0.0) 0.0 (0.0) 5.7 (0.3)** 1.0
 ≥2 98.4 (0.4) 35.8 (0.7) 23.0 (0.7)* 1.5 97.6 (0.6) 33.6 (0.7) 14.9 (0.5)** 1.5 99.8 (0.2) 32.1 (0.6) 8.1 (0.4)** 1.5
 ≥3 93.3 (0.8) 55.2 (0.7) 28.9 (0.8)* 2.1 90.7 (1.1) 51.8 (0.6) 18.3 (0.7)** 1.9 98.4 (0.5)** 50.0 (0.6) 10.6 (0.5)* 2.0
 ≥4 84.3 (1.1) 71.9 (0.6) 36.9 (1.0)* 3.0 78.9 (1.4) 67.7 (0.6) 22.6 (0.8)** 2.4 94.6 (1.0)** 66.1 (0.6)* 14.3 (0.7)* 2.7
 ≥5 69.1 (1.3) 85.8 (0.4) 48.7 (1.3)* 4.9 60.9 (1.7) 81.3 (0.5) 28.0 (1.1)** 3.3 84.6 (1.9)* 80.5 (0.5) 20.7 (1.1) 4.3
 ≥6 56.0 (1.4) 90.6 (0.4) 53.7 (1.5) 6.0 46.4 (1.8) 86.5 (0.4) 29.1 (1.3)** 3.4 74.0 (2.1) 86.4 (0.4) 24.6 (1.3) 5.4
 ≥7 45.1 (1.4) 93.9 (0.3) 59.1 (1.8) 7.4 34.8 (1.7) 90.2 (0.4) 29.8 (1.6)** 3.6 64.6 (2.3)* 90.7 (0.4) 29.3 (1.6) 6.9
 ≥8 33.1 (1.4) 96.2 (0.3) 63.1 (2.0) 8.7* 23.9 (1.6) 93.3 (0.4) 29.9 (2.0)* 3.6 50.3 (2.3) 93.9 (0.3) 33.3 (1.9) 8.2
 ≥9 16.9 (1.0) 98.6 (0.2) 69.7 (3.0) 12.1 9.7 (1.2) 96.7 (0.2) 26.1 (2.9) 2.9 30.4 (2.0) 97.6 (0.2) 43.6 (3.2) 12.7
 ≥10 13.1 (0.9) 99.2 (0.1) 75.5 (3.5) 16.4* 6.8 (1.0) 97.6 (0.2) 25.6 (3.4) 2.8 25.0 (1.9) 98.5 (0.2) 49.9 (4.0) 16.7 **
 ≥11 4.7 (0.6) 99.9 (0.1) 87.0 (5.8) 47.0** 2.2 (0.6) 99.3 (0.1) 26.8 (6.4) 3.1* 9.5 (1.4) 99.6 (0.1) 60.2 (7.3) 23.8
AUC 0.858 (0.005) 0.791 (0.006) 0.896 (0.006)*
Open in a new tab

Note: Figures in parentheses are standard errors of estimates. Values of 0.0 denote standard errors <.05. Shaded cells denote recommended cutpoints for outcome in question.

**

Significantly different (p<.005) from corresponding value for DSM-IV diagnosis

*

Significantly different (p<.05) from corresponding value for DSM-IV diagnosis

a

Positive predictive value

b

Positive likelihood ratio

c

N differs for each row, based on n of cases with AUDIT-C score in question (decreasing from 7,308 for scores ≤1 to 73 for scores ≥11)

4. Discussion

The AUDIT-C cutpoints that optimally screen for DSM-IV AUD will also optimize screening for DSM-5 AUD; thus, clinicians should not face any major overhaul of their current screening procedures as a result of the revision. However, the AUDIT-C will do a somewhat better job of screening for DSM-5 than DSM-IV disorders, yielding slightly higher PPV for any AUD and moderate AUD and slightly higher sensitivity and AUC for severe AUD. Thus, clinicians should benefit from fewer false positive screening results.

Why might screening performance differ for DSM-IV and DSM-5 AUD? It could reflect a difference in the severity of the disorders. Assuming the AUDIT-C is a scaled marker of latent AUD severity, then at any given cutpoint a more severe disorder increases the likelihood that affected individuals will screen positive (higher sensitivity), whereas a milder disorder is more likely to be present among individuals with positive screens (higher PPV). Thus the higher sensitivity for DSM-5 severe AUD could reflect its being slightly less prevalent (i.e., more severe) than DSM-IV dependence, 3.9% versus 4.4%, and the higher PPV for DSM-5 moderate and any AUD could reflect their being slightly more prevalent (i.e., milder) than their DSM-IV counterparts, 10.8% and 6.9% versus 9.7% and 5.3%. Alternatively, the pooling of abuse and dependence criteria under the DSM-5 may have yielded diagnoses more strongly associated with consumption, the basis of the AUDITC, as prior analyses have shown that the AUD criteria most strongly associated with drinking quantity and frequency comprised a mixture of abuse and dependence criteria (Dawson et al., 2010).

A limitation of this study is that the AUDIT-C questions were embedded in a larger series of consumption questions, potentially affecting recall. They also differed slightly from the original AUDIT-C in the range of response options and in asking about the frequency of drinking 5+ rather than 6+ drinks; however, 5+ U.S. drinks at 12–14g ethanol each have about the same ethanol content as 6+ drinks containing 10g each, the drink size assumed by the original AUDIT (Saunders et al, 1993). Also, NESARC responses were anonymous, which may yield more complete reporting of consumption that would be obtained in clinical settings. Although these limitations suggest the need for caution in extrapolating the optimal cutpoints to other populations, they should not affect the comparison of the AUDIT-C performance for screening DSM-5 versus DSM-IV AUD. Subgroup (e.g., gender-specific) analyses, although of considerable importance to clinicians, lay beyond the scope of this brief report. These will be addressed in forthcoming analyses.

Supplementary Material

01

Acknowledgments

Role of funding source: The study on which this paper is based, the Wave 2 National Epidemiologic Survey on Alcohol and Related Conditions (NESARC), is sponsored by the National Institute on Alcohol Abuse and Alcoholism, National Institutes of Health, U.S. Department of Health and Human Services, with supplemental support from the National Institute on Drug Abuse. This research was supported in part by the Intramural Program of the National Institutes of Health, National Institute on Alcohol Abuse and Alcoholism.

Footnotes

Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

*

Supplementary material can be found by accessing the online version of this paper at http://dx.doi.org and by entering doi:…

Contributors: Dr. Dawson planned the study, executed the SUDAAN and SAS statistical analyses and prepared the first draft of the manuscript. Dr. Smith calculated the positive likelihood ratios and provided revisions to the manuscript. Dr. Saha and Ms. Rubinsky worked on estimation of the AUROC values and provided revisions to the manuscript. Dr. Grant reviewed the study design and statistical approaches and provided revisions to the manuscript.

The views and opinions expressed in this paper are those of the authors and should not be construed to represent the views of any of the sponsoring organizations, agencies or the U.S. government.

Conflict of interest: None of the authors has any financial conflict of interest to report.

REFERENCES

  1. Aertgeerts B, Buntinx F, Ansoms S, Fevery J. Screening properties of questionnaires and laboratory tests for the detection of alcohol abuse or dependence in a general practice population. Br. J. Gen. Pract. 2002;51:206–217. [PMC free article] [PubMed] [Google Scholar]
  2. Agrawal A, Heath AC, Lynskey MT. DSM-IV to DSM-5: the impact of proposed revisions on diagnosis of alcohol use disorders. Addiction. 2011;106:1935–1943. doi: 10.1111/j.1360-0443.2011.03517.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
  3. American Psychiatric Association. The Diagnostic and Statistical Manual of Mental Disorders. 4th Edition. American Psychiatric Association; Washington, D.C.: 1994. [Google Scholar]
  4. Bradley KA, Bush KR, Epler AJ, Dobie DJ, Davis TM, Sporleder JL, Maynard C, Burman ML, Kivlahan DR. Two brief alcohol-screening tests from the Alcohol Use Disorders Identification Test (AUDIT): validation in a female veterans affairs patient population. Arch. Int. Med. 2003;163:821–829. doi: 10.1001/archinte.163.7.821. [DOI] [PubMed] [Google Scholar]
  5. Bradley KA, DeBenedetti AF, Volk RJ, Williams EC, Frank D, Kivlahan DR. AUDIT-C as a brief screen for alcohol misuse in primary care. Alcohol. Clin. Exp. Res. 2007;31:1208–1217. doi: 10.1111/j.1530-0277.2007.00403.x. [DOI] [PubMed] [Google Scholar]
  6. Burns E, Gray R, Smith LA. Brief screening questionnaires to identify problem drinking during pregnancy: a systematic review. Addiction. 2010;105:601–614. doi: 10.1111/j.1360-0443.2009.02842.x. [DOI] [PubMed] [Google Scholar]
  7. Bush K, Kivlahan DR, McConell MB, Fihn SD, Bradley KA. The AUDIT alcohol consumption questions (AUDIT-C): an effective brief screening test for problem drinking. Arch. Int. Med. 1998;158:1789–1795. doi: 10.1001/archinte.158.16.1789. [DOI] [PubMed] [Google Scholar]
  8. Dawson DA, Grant BF, Stinson FS, Zhou Y. 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. 2005a;29:844–854. doi: 10.1097/01.alc.0000164374.32229.a2. [DOI] [PubMed] [Google Scholar]
  9. Dawson DA, Grant BF, Stinson FS. The AUDIT-C: screening for alcohol use disorders and risk drinking in the presence of other psychiatric disorders. Compr. Psychiatry. 2005b;46:405–416. doi: 10.1016/j.comppsych.2005.01.006. [DOI] [PubMed] [Google Scholar]
  10. Dawson DA, Saha TD, Grant BF. A multidimensional assessment of the validity and utility of alcohol use disorder severity as determined by item response theory models. Drug Alcohol Depend. 2010;107:31–38. doi: 10.1016/j.drugalcdep.2009.08.019. [DOI] [PMC free article] [PubMed] [Google Scholar]
  11. Gordon AJ, Maisto SA, McNeil M, Kraemer KL, Conigliaro RL, Kelley ME, Conigliaro J. Three questions can detect hazardous drinkers. J. Fam. Pract. 2001;50:313–320. [PubMed] [Google Scholar]
  12. Grant BF, Dawson DA, Hasin DS. The Alcohol Use Disorder and Associated Disabilities Interview Schedule – DSM-IV Version (AUDADIS-IV) Bethesda, MD: National Institute on Alcohol Abuse and Alcoholism; 2001. [Google Scholar]
  13. Grant BF, Kaplan K, Shepard J, Moore T. Source and Accuracy Statement for Wave 1 of the 2001–2002 National Epidemiologic Survey on Alcohol and Related Conditions. Bethesda, MD: National Institute on Alcohol Abuse and Alcoholism; 2003a. [Google Scholar]
  14. Grant BF, Dawson DA, Stinson FS, Chou PS, Pickering RP. The Alcohol Use Disorder and Associated Disabilities Schedule (AUDADIS): reliability of alcohol and drug modules in a general population sample. Drug Alcohol Depend. 2003b;39:7–16. doi: 10.1016/0376-8716(95)01134-k. [DOI] [PubMed] [Google Scholar]
  15. Grant BF, Kaplan K, Moore T, Kimball J. 2004–2005 Wave 2 National Epidemiologic Survey on Alcohol and Related Conditions: Source and Accuracy statement. Bethesda, MD: National Institute on Alcohol Abuse and Alcoholism; 2007. [Google Scholar]
  16. Gual A, Segura L, Contel M, Heather N, Colom J. AUDIT-3 and AUDIT-4: effectiveness of two short forms of the Alcohol Use Disorders Identification Test. Alcohol Alcohol. 2002;3:591–596. doi: 10.1093/alcalc/37.6.591. [DOI] [PubMed] [Google Scholar]
  17. Hasin D, Paykin A. Dependence symptoms but no diagnosis: diagnostic ‘orphans’ in a community sample. Drug Alcohol Depend. 1998;50:19–26. doi: 10.1016/s0376-8716(98)00007-6. [DOI] [PubMed] [Google Scholar]
  18. Kaarne T, Aalto M, Kuokkanen M, Seppa K. AUDIT-C, AUDIT-3 and AUDIT-QF in screening risk drinking among Finnish occupational health-care patients. Drug Alcohol Rev. 2010;29:563–567. doi: 10.1111/j.1465-3362.2010.00172.x. [DOI] [PubMed] [Google Scholar]
  19. Kelly TM, Donovan JE, Chung T, Bukstein OG, Cornelius JR. Brief screens for detecting alcohol use disorder among 18–20 year old young adults in emergency departments: comparing AUDIT-C, CRAFFT, RAPS4-QF, FAST, RUFT-Cut, and DSM-IV 2-Item Scale. Addict. Behav. 2009;34:668–674. doi: 10.1016/j.addbeh.2009.03.038. [DOI] [PMC free article] [PubMed] [Google Scholar]
  20. McPherson TL, Goplerud E, Derr D, Mickenberg J, Courtemanche S. Telephonic screening and brief intervention for alcohol misuse among workers contacting the employee assistance program: a feasibility study. Drug Alcohol Rev. 2010;29:641–646. doi: 10.1111/j.1465-3362.2010.00249.x. [DOI] [PubMed] [Google Scholar]
  21. Nayak MB, Bond JC, Cherpitel C, Patel V, Greenfield TK. Detecting alcohol-related problems in developing countries: a comparison of 2 screening measures in India. Alcohol. Clin. Exp. Res. 2009;33:2057–2066. doi: 10.1111/j.1530-0277.2009.01045.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
  22. Research Triangle Institute. SUDAAN Language Manual, Release 10.0. Research Triangle Park, NC: Research Triangle Institute; 2008. [Google Scholar]
  23. Rumpf H-J, Hapke U, Meyer C, John U. Screening for alcohol use disorders and at-risk drinking in the general population: psychometric performance of three questionnaires. Alcohol Alcohol. 2002;37:261–268. doi: 10.1093/alcalc/37.3.261. [DOI] [PubMed] [Google Scholar]
  24. SAS Institute, Inc. SAS/STAT 9.2 User’s Guide. Cary, NC: SAS Institute, Inc.; 2008. [Google Scholar]
  25. Saunders JB, Aasland OG, Babor TF, de la Fuente JR, Grant M. Development of the Alcohol Use Disorders Identification Test (AUDIT): WHO collaborative project on early detection of persons with harmful alcohol consumption. II. Addiction. 1993;88:791–804. doi: 10.1111/j.1360-0443.1993.tb02093.x. [DOI] [PubMed] [Google Scholar]
  26. Simel DL, Samsa GP, Matchar DB. Likelihood ratios with confidence sample size estimation for diagnostic test studies. J. Clin. Epidemiol. 1991;44:763–770. doi: 10.1016/0895-4356(91)90128-v. [DOI] [PubMed] [Google Scholar]
  27. Swets JA, Pickett RM. Evaluation of Diagnostic Systems: Methods from Signal Detection Theory. New York: Academic Press; 1982. [Google Scholar]
  28. World Health Organization. The ICD10 Classification of Mental and Behavioural Disorders. Geneva: World Health Organization; 1992. [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

01
NIHMS389692-supplement-01.docx (17.4KB, docx)

RESOURCES