The Role of Society in Promoting or Deterring Alcohol Use and Addiction

Simplifying Alcohol Assessment: Two Questions to Identify

Alcohol Use Disorders

Daniel C. Vinson, Robin L. Kruse, and J. Paul Seale

Background: Previous work has validated a single question to screen for hazardous or harmful drinking, but identifying those patients who have an alcohol use disorder (AUD) among those who screen positive is still time consuming. We therefore sought to develop and validate a brief assessment instrument using DSM-IV criteria for use in primary care medical practice.

Methods: Four cross-sectional surveys of past-year drinkers. The developmental sample included patients presenting to emergency departments with an acute injury. The second sample, from the same study, was recruited by random-digit dialing. The third sample was recruited in 5 family medicine practices in Georgia. The fourth sample was the National Epidemiologic Survey on Alcohol and Related Conditions. Interviews with the first 3 samples used the Diagnostic Interview Schedule. The National Epidemiologic Survey on Alcohol and Related Conditions (NESARC) used the Alcohol Use Disorder and Associated Disabilities Interview Schedule.

Results: Two constructs with promising test characteristics were identified: recurrent drinking in hazardous situations and drinking more than intended. Among those who screened positive with the single question in the developmental sample (N 5 959), if either of the 2 items was positive, the sensitivity for current AUD was 95% and the specificity was 77%. In the second (N 5 494) and third (N 5 280) samples, the sensitivity was 94 and 95% and the specificity was 62 and 66%, respectively, among those with a positive screen. In the NESARC sample, including those with at least 1 occasion in the past year of drinking 5 or more drinks (N 5 7,890), the sensitivity and specificity were 77 and 86%, respectively.

Conclusions: The sensitivity and specificity of these 2 items across 4 samples suggest that they could be formulated into 2 questions, potentially providing busy primary care clinicians with an efficient, reasonably accurate assessment instrument to identify AUD among those patients who screen positive with the single screening question.

Key Words: Screening, Primary Healthcare.

HAZARDOUS AND HARMFUL drinking arecommon. Of adults in the United States, 16.5% have had at least 1 episode of drinking more than 5 drinks in the previous 30 days (Division of Adult and Community Health, 2005), and 8.4% have a diagnosable past-year alcohol use disorder (AUD) (Grant et al., 2004). They are also treatable. In efficacy trials, brief interventions by primary care physicians helped about 40% of hazardous drinkers reduce their drinking to safe levels, compared

with 20% in control groups (Fleming et al., 1997; Moyer et al., 2002; Wallace et al., 1988). In healthcare settings, however, most patients with

alcohol problems are not identified and do not receive brief interventions (Bradley et al., 1995; D’Amico et al., 2005; Deitz et al., 1994; McGlynn et al., 2003; Moyer and Finney, 2005; Nilsen et al., 2006; Rush et al., 2003; Spandorfer et al., 1999; Vinson et al., 2000). Even physi- cians who participated in a successful brief intervention trial did not continue with screening and intervention after the researchers withdrew (Beich et al., 2002). New approaches are required to integrate screening and brief intervention into routine primary medical care practice (Glasgow et al., 2003). In their systematic review, Whitlock et al. (2004) concluded, ‘‘Future research should focus on implementation strategies to facilitate adoption of these practices into routine health care.’’ Redesigning innovations can enhance their diffusion

into practice (Berwick, 2003; Rogers, 2003). Given the complexities of primary medical care office visits with their multiple competing demands (Crabtree et al., 2005; Ostbye et al., 2005; Sussman et al., 2006; Yarnall et al., 2003), integration of alcohol screening and intervention into routine practice might be facilitated by providing primary

From the Department of Family and Community Medicine, University of Missouri-Columbia, Columbia, Missouri (DCV, RLK); and the Department of Family Medicine, Mercer University School of Medicine, Macon, Georgia (JPS).

Received for publication December 12, 2006; accepted April 19, 2007. This study was funded by a grant from the National Institute on Alcohol

Abuse and Alcoholism (R01 AA11078). Additional support was provided by the Department of Family and Community Medicine at the University of Missouri-Columbia through the Opal Lewis Fund for alcohol research and the Center for Family Medicine Science, which was supported by a grant from the American Academy of Family Physicians.

Reprint requests: Daniel C. Vinson, MD, MSPH, Department of Fam- ily and Community Medicine, University of Missouri-Columbia, Columbia, MO 65212; Fax: 573-882-3184; E-mail: [email protected]

Copyright r 2007 by the Research Society on Alcoholism.

DOI: 10.1111/j.1530-0277.2007.00440.x

Alcohol Clin Exp Res, Vol 31, No 8, 2007: pp 1392–13981392

ALCOHOLISM: CLINICAL AND EXPERIMENTAL RESEARCH Vol. 31, No. 8 August 2007

care clinicians with quick and accurate screening and assessment instruments that could be used orally within the flow of an ordinary office visit. Many screening instruments are available that could be

used not only for screening but also initial assessment in a primary care setting. Compared with DSM criteria, the CAGE questions (Ewing, 1984) had an area under the receiver operating characteristic (ROC) curve of 0.87 in a recent meta-analysis (Aertgeerts et al., 2004). In a national sample, a CAGE score of 2 or more had a sensitivity of 67% and a specificity of 98% in detecting alcohol depen- dence, and 36 and 99%, respectively, in detecting alcohol abuse (Cherpitel, 2002). In a large probability sample in a primary care outpatient setting, the CAGE questions performed differently across race and gender groups (Volk et al., 1997). In a wide variety of settings, the 10-item Alcohol Use

Disorders Identification Test (AUDIT) has a sensitivity and specificity generally between 80 and 95%, with an area under ROC curve in most studies between 0.8 and 0.9 (Reinert and Allen, 2007). The test characteristics of abbreviated versions of the AUDIT are almost as good as the full AUDIT (Reinert and Allen, 2007). In the assessment of a patient who screens positive, the

AUDIT can be divided into zones, which can guide further treatment and referral decisions (Donovan et al., 2006). Probably the most accurate assessment is a skilled inter- view by the clinician using the Diagnostic and Statistical Manual of Mental Disorders, fourth edition (DSM-IV) criteria, as in the NIAAA Guide (National Institute on Alcohol Abuse and Alcoholism, 2005). Neither of these instruments, however, can be used in the limited time available when the clinician performs opportunistic screening. To reduce the time required for screening, we have

developed (Taj et al., 1998) and validated (Williams and Vinson, 2001) a single question to screen for hazardous drinking or AUD: ‘‘When was the last time you had more than X drinks in one day?’’ where X is 4 for women and 5 for men. With an answer of any time in the past 3 months, the sensitivity and specificity were both 86% in identifying past-month hazardous drinking and/or current (past-year) AUD (Williams and Vinson, 2001). In an independent study in primary care, its sensitivity and specificity were lower, 80 and 74%, respectively, similar to the AUDIT in that study (Seale et al., 2006). Even with a simple single-question screen, however, the

clinician must make a quick decision about the patient who screens positive—to omit some other service from this visit, make subsequent patients wait longer, schedule a follow-up visit specifically to address alcohol, or refer the patient to specialty care. At this decision point, having a simplified approach to triage assessment may be helpful. In addition, if thresholds for defining a positive single- question screen are lowered by 1 drink to be consistent with guidelines (National Institute on Alcohol Abuse and

Alcoholism, 2005), the number of positive screens would increase, increasing the burden on the clinician, and a quick triage assessment instrument could then be especially useful. Here, we report the development and validation of such

an instrument. The Missouri Injury Study, a case–control study of alcohol and injury (Vinson et al., 2003), provided developmental data. Validation used data from controls in that study, from family medicine practices in Georgia (Seale et al., 2006), and from the 2001–2002 National Epidemiologic Survey on Alcohol and Related Conditions (NESARC, http://niaaa.census.gov/).

METHODS

The Missouri Injury Study was a case–crossover and case–control study (Vinson et al., 2003). Cases were acute injury patients cared for in any of the 3 emergency departments in Boone County, Missouri, from February 1998 through March 2000. The interview included the alcohol questions from the Diagnostic Interview Schedule (DIS) (Robins et al., 1996), which were mapped to the diagnostic criteria of the DSM-IV (American Psychiatric Association, 1994) to identify current (past-year) AUD. The DIS is a validated, structured inter- view designed for use by trained lay interviewers (Robins et al., 1981). A total of 2,517 injured patients were interviewed, 86% of those approached. Information about gender and race is provided in Table 2. Age was skewed toward younger patients (mean 32, SD 13, interquartile range 21–39).

The second sample was composed of controls in the Missouri Injury Study, who were recruited by random-digit dialing and inter- viewed by telephone. The structured interview also used the DIS questions to identify current AUD. The response rate was 47%; 1,856 persons were interviewed. Ages were matched to cases by design (mean 34, SD 13, interquartile range 23–42).

The Georgia data were collected in the Vital Signs Screening Project in family medicine offices, 4 in 2 small urban communities and 1 in rural central Georgia (Seale et al., 2006). Only individuals who had consumed 6 or more drinks in the past 12 months were enrolled. Interviews were conducted in physician offices from July 2004 through July 2005. The mean age was 41 (SD 13, interquartile range 30–49).

The NESARC (Grant and Dawson, 2006) was a stratified random sample of the civilian, noninstitutionalized, adult U.S. population. Participants were interviewed in their homes between August 2001 and April 2002. The response rate was 81%. The Alcohol Use Disorder and Associated Disabilities Interview Schedule (AUDA- DIS), mapped to DSM-IV criteria, was used to identify lifetime and past-year AUD. The Alcohol Use Disorder and Associated Dis- abilities Interview Schedule is a validated, structured interview designed for administration by trained lay interviewers (Grant and Dawson, 2006). The mean age was 46 (SD 18, interquartile range 32–59). Table 1 provides the wording used in the 2 interview instruments for the items reported on here. The single screening question used in the Missouri and Georgia studies was not asked in NESARC, but participants were asked, ‘‘During the last 12 months, about how often did you drink five or more drinks in a single day?’’ with 11 answer options from ‘‘every day’’ to ‘‘never.’’

We used Stata (Stata Corporation, 2005) to perform receiver operating characteristic analyses using DIS-derived or AUDADIS- derived diagnoses of AUD as the criterion variable, positive versus negative on the 2-question assessment instrument as the predictor variable, and stratified into subgroup analyses by gender and race. To perform weighted analyses of the NESARC data, we used programming in Stata written by Roger Newson (2006).

1393SIMPLIFYING ALCOHOL ASSESSMENT

RESULTS

In developmental analyses with data from cases in the Missouri study, bivariate analyses identified 2 DSM-IV criteria that together showed 96% sensitivity and 85% specificity in identifying AUD: recurrent drinking in situations in which it is physically hazardous, and drink- ing in larger amounts or over a longer period than intended. We then used data from the Missouri Injury Study control group, the Georgia study, and NESARC to validate the utility of a combination of these 2 constructs in identifying AUD. Analyses included those who reported a total past-year consumption of 6 or more drinks in the Missouri and Georgia studies and 1 or more drinks in NESARC (Table 2). Including all past-year drinkers, the sensitivity in NESARC was lower than in the other 2 samples (72%), but the specificity was 95%.

We also performed analyses limited to those who re- ported heavy episodic drinking because this will be how the new 2-question instrument might be used (Table 3). With Missouri and Georgia data, we included those who screened positive with the single question. Among cases in the Missouri study who screened positive, if either or both of these DSM-IV criteria were met, the sensitivity for detecting a current AUD was 95% and the specificity was 77%. With NESARC data, we included participants who reported at least 1 occasion of drinking 5 or more drinks in the past year. ‘‘Drinking in larger amounts or over a longer period

than intended’’ is a DSM-IV diagnostic criterion derived from 2 constructs in the DIS and AUDADIS. In the Missouri and Georgia studies, the second question, drink- ing longer than intended, was asked only of respondents who denied drinking more than intended. Among cases in

Table 1. Questions From Structured Interviews That Mapped to the Identified DSM-IV Constructs

In the Diagnostic Interview Schedule (DIS) ‘‘In the past year, have you sometimes been under the influence of alcohol in situations where you could have caused an accident or gotten hurt?’’ ‘‘Have there often been times when you had a lot more to drink than you intended to have?’’

The following item was asked only of those who denied drinking more than intended ‘‘Were there many times when your drinking continued for much longer than you intended to?’’

In the Alcohol Use Disorder and Associated Disabilities Interview Schedule (AUDADIS) Each question was preceded by the stem ‘‘In your entire life, did you ever . . .’’

‘‘have a period when you ended up drinking more than you meant to?’’ ‘‘have a period when you kept on drinking for longer than you had intended to?’’ ‘‘more than once drive a car, motorcycle, truck, boat, or other vehicle after having too much to drink?’’ ‘‘get into situations while drinking or after drinking that increased your chances of getting hurt—like swimming, using machinery, or walking in a

dangerous area or around heavy traffic?’’ Respondents who answered positively to a given item were then asked if the item had occurred in the previous 12 months.

Table 2. Utility of the 2-Question Assessment in Detecting a Current Alcohol Use Disorder: Sensitivity, Specificity, Area Under the Receiver Operating Characteristic (ROC) Curve, and Likelihood Ratios

Data set Subgroup N Sensitivity

(%) Specificity

(%) Area under the ROC

curve (95% CI) Likelihood ratios: positive

and negative

Cases from the Missouri Injury Study All 1,522 96 85 0.90 (0.89–0.92) 6.4 0.05 Men 1,009 97 83 0.90 (0.88–0.92) 5.8 0.04 Women 543 92 88 0.90 (0.87–0.93) 7.7 0.09 Blacks 132 97 82 0.90 (0.85–0.95) 5.5 0.03 Whites 1,338 96 85 0.91 (0.89–0.92) 6.6 0.05

Controls from the Missouri Injury Study All 1,124 93 80 0.87 (0.85–0.89) 4.7 0.09 Men 644 93 80 0.86 (0.84–0.89) 4.6 0.09 Women 480 93 81 0.87 (0.84–0.90) 4.9 0.08 Blacks 63 100 66 0.83 (0.76–0.90) 2.9 0 Whites 1,000 93 81 0.87 (0.85–0.89) 4.9 0.09

Georgia Vital Signs Screening Project All 623 94 82 0.88 (0.85–0.91) 5.2 0.07 Men 287 95 77 0.86 (0.82–0.90) 4.1 0.06 Women 338 93 86 0.89 (0.85–0.93) 6.4 0.09 Blacks 238 94 77 0.86 (0.81–0.90) 4.2 0.08 Whites 377 94 85 0.89 (0.86–0.93) 6.1 0.07

NESARC All 26,946 72 95 0.84 (0.83–0.84) 16 0.30 Men 13,067 71 95 0.83 (0.82–0.84) 14 0.30 Women 13,879 73 96 0.84 (0.83–0.86) 18 0.28 Blacks 4,185 64 96 0.80 (0.78–0.82) 17 0.37 Whites 16,732 74 95 0.85 (0.84–0.86) 15 0.27 Hispanics 4,949 69 96 0.82 (0.80–0.84) 17 0.33

Including all past-year drinkers. NESARC, National Epidemiologic Survey on Alcohol and Related Conditions.

1394 VINSON ET AL.

the Missouri Injury Study, 441 reported drinking more than intended, but only 49 reported drinking longer than intended and not more than intended. Among controls, 314 reported drinking more than intended, and 35 reported only drinking longer than intended. In the Georgia study, 190 patients reported drinking more than intended, but only 4 reported only drinking longer than intended. The Alcohol Use Disorder and Associated Disabilities Interview Schedule asks all drinkers about both drinking more than intended and longer than intended. In NESARC, 2,401 reported drinking more than intended in the previous 12 months and 1,969 reported drinking longer than intended, but only 494 of those 1,969 did not also report drinking more than intended.

Race, Gender, and Age Subgroup Analyses

Tables 2 and 3 show sensitivity and specificity by gender, and for African Americans and Caucasians in all 4 data sets. The National Epidemiologic Survey on Alcohol and Related Conditions data also included sufficient Hispanics for subgroup estimates. These analyses used unweighted NESARC data because weighting the overall analysis changed these results minimally. In NESARC, the 2-ques- tion assessment was less sensitive among African Americans and Hispanics than among whites. Among cases in the Missouri study, the 2-question

instrument performed well across age groups. Among those in the oldest quartile (age439), the sensitivity and specificity were 97 and 93%, respectively. There were only

54 past-year drinkers age 65 or older, 2 of whom had AUD. The 2-question instrument identified both, and with 96% specificity. However, among the 8,205 NESARC participants age 65 and older, 109 of whom (1.3%) had an AUD, the sensitivity was much lower, 37.6%, while the specificity was 99.7%. The other samples had too few participants age 65 and older for reliable analyses, but showed the same pattern of lower sensitivity.

Lifetime AUD

The Missouri and Georgia studies asked only about past-year alcohol-related problems, but NESARC asked about both lifetime and past-year experiences. Using the same 2 DSM-IV constructs, but now focused on lifetime consequences and lifetime AUD, the sensitivity and specificity of the 2-item assessment instrument in NES- ARC data were 85 and 88%, respectively, among past-year drinkers, with an area under the ROC curve of 0.87 (95% CI 0.86–0.87). Including all those who reported ever drinking, the specificity was slightly better, 89%. Again, weighting the analysis changed these findings minimally.

DISCUSSION

Two of the 11 DSM-IV criteria for alcohol abuse or dependence together have promising sensitivity and specificity in identifying persons with either disorder. When preceded by a single-question screening instrument

Table 3. Utility of the 2-Question Assessment in Detecting an Alcohol Use Disorder: Sensitivity, Specificity, Area Under the Receiver Operating Characteristic (ROC) Curve, and Likelihood Ratios

Data set Subgroup N Sensitivity

(%) Specificity

(%) Area under the ROC

curve (95% CI) Likelihood ratios: positive

and negative

Cases from the Missouri Injury Study All 959 95 77 0.86 (0.84–0.88) 4.1 0.06 Men 662 97 75 0.86 (0.84–0.88) 3.9 0.04 Women 297 91 80 0.86 (0.82–0.90) 4.6 0.11 Blacks 78 96 75 0.86 (0.79–0.93) 3.8 0.05 Whites 833 96 77 0.87 (0.84–0.89) 4.2 0.06

Controls from the Missouri Injury Study All 494 94 62 0.78 (0.74–0.81) 2.4 0.10 Men 312 94 65 0.79 (0.75–0.83) 2.7 0.09 Women 182 94 57 0.75 (0.70–0.80) 2.2 0.11 Blacks 23 100 67 0.83 (0.69–0.97) 3.0 0 Whites 445 94 62 0.78 (0.74–0.81) 2.5 0.10

Georgia Vital Signs Screening Project All 280 95 66 0.81 (0.77–0.85) 2.8 0.08 Men 157 97 62 0.79 (0.74–0.85) 2.6 0.05 Women 123 92 71 0.81 (0.75–0.88) 3.2 0.12 Blacks 112 97 62 0.80 (0.74–0.86) 2.6 0.05 Whites 165 93 69 0.81 (0.76–0.87) 3.0 0.09

NESARC All 7,890 77 86 0.82 (0.81–0.83) 5.6 0.27 Men 5,453 75 88 0.82 (0.81–0.83) 6.3 0.28 Women 2,437 81 83 0.82 (0.80–0.83) 4.6 0.24 Blacks 871 71 89 0.80 (0.77–0.83) 6.2 0.32 Whites 5,028 80 85 0.82 (0.81–0.83) 5.2 0.24 Hispanics 1,683 71 90 0.80 (0.78–0.82) 7.0 0.33

Including only those who screened positive with the single screening question or, in NESARC, those in the past year with at least 1 occasion of drinking 5 or more drinks.

NESARC, National Epidemiologic Survey on Alcohol and Related Conditions.

1395SIMPLIFYING ALCOHOL ASSESSMENT

(Williams and Vinson, 2001), the 3 items could serve as a triage instrument for brief screening and initial assessment. The major strength of this study is the validation in 3

other data sets. Although the Missouri study participants were not representative of the national population, younger adults were oversampled because of the age distribution of acute injury, age groups in which the prev- alence of AUD is higher (Harford et al., 2005). They are balanced by the data from Georgia, which included more African Americans, and NESARC, which used a stratified sampling design to insure representation of the U.S. adult population. The major limitation of this study is that the 2 DSM-IV

constructs that we identified are derived from 3 questions in DIS and 4 in AUDADIS. Further research is needed to develop and validate 2 reworded questions. However, of respondents who reported drinking more or longer than intended, between 83 and 98% reported drinking more than intended in the 4 data sets. Therefore, the first 2 questions in Table 1, as phrased in the DIS, could serve as a useful starting point for further refinement of this approach to initial assessment. It is unclear why the sensitivity for current AUD in the

NESARC data set was lower than in the other 3 samples. The results obtained using the DIS questions in 3 different settings (patients in Missouri emergency departments, Missouri participants recruited by random digit dialing, and patients in Georgia primary care clinics) are strikingly similar, suggesting that setting was not a major factor. The National Epidemiologic Survey on Alcohol and Related Conditions participants were older than in the Missouri sam- ples, but when we restricted the analysis to the NESARC subgroup whose ages were within the interquartile range of the Missouri cases, the findings changed little. Perhaps the findings with the NESARC data were

different because the questions were phrased differently (Table 1) and presented differently. In the Missouri and Georgia studies, DIS questions addressed only alcohol- related problems in the previous 12 months. In NESARC, each AUDADIS item was asked first about lifetime experience, and then about the previous 12 months. The substantially higher sensitivity of the 2 constructs in the NESARC data when focused on lifetime AUD, com- pared with past-year AUD, suggests that this difference between focusing exclusively on past-year symptoms and focusing more broadly first on lifetime symptoms may have been a factor. The high level of sensitivity in the 2 clinical samples, which used DIS questions focused on the past 12 months, underscores the potential utility of this 2-question instrument in clinical practice. A more fundamental question remains: why develop a

new assessment tool? First, the approach presented in the current study is briefer than other available alcohol assessment instruments and therefore might be more easily integrated into the review of the patient’s history in routine primary care office visits. This contrasts with

other suggested approaches that require use of a longer instrument such as the AUDIT (Babor et al., 2006) or a DSM-IV checklist (National Institute on Alcohol Abuse and Alcoholism, 2005). Use of such instruments requires substantially more interviewer time or necessitates a shift to a paper-based questionnaire, interrupting the flow of con- versation of the office visit. Whether this new instrument can be used conversationally remains to be determined. Second, the approach presented here uses DSM-IV

diagnostic criteria. This congruence with DSM-IV criteria should make it easier for clinicians to identify patients with more advanced AUDs who may require more in-depth primary care medical management (Anton et al., 2006) or referral. Third, this approach can provide clinicians with an

answer to their question, ‘‘What do I do if a patient screens positive?’’ With the approach presented here, the clinician could screen with 1 question, and then ask 2 questions that determine with reasonable accuracy whether the patient has an AUD. If the 2-item assessment is negative, a brief discussion of alcohol issues during that same visit might suffice, similar to the very brief intervention used in the Cutting Back study (Babor et al., 2006). But if the patient answers 1 or both of these 2 items affirmatively, the clinician could invite the patient to return for a visit devoted to further discussion of alcohol use and its consequences. The agenda for that visit could include more detailed assessment, further intervention, initiation of medication (naltrexone, e.g. see Anton et al., 2006), or referral for specialty care as needed. This new approach is modeled on what primary care

clinicians currently do with other chronic illnesses that are incidentally discovered during a visit for some other problem (Watkins et al., 2003). For example, if a patient presents for care of a knee injury, and the blood pressure is unexpectedly elevated, the physician typically will focus the visit on the knee problem, addressing the blood pressure briefly with a request to gather more data and an invitation to return for a visit focused on the possible diagnosis of hypertension. In previous clinical trials of screening and brief intervention for alcohol problems, the initial screening and assessment was performed by a researcher, who scheduled the patient for the intervention visit with the physician. We believe that this approach has not been integrated into primary care practice because, in the absence of a researcher, the physician is expected to screen and somehow intervene at the same visit. The approach presented here may achieve this integration by better matching the usual patient-centered way in which physicians now address other chronic illnesses found by screening. At this point in the development of this 2-question

assessment instrument, it seems reasonable to us to seek to maximize sensitivity. As it turns out, the items we chose have excellent sensitivity and fair specificity, at least in the first 3 samples. Further research in practice-based research

1396 VINSON ET AL.

networks will tell us if the balance between sensitivity and specificity is at the right point or whether it needs to be shifted toward a more specific (and less sensitive) point. If the questions are too nonspecific, clinicians may get discouraged, and the screening–assessment follow-up approach we are proposing might not get integrated into routine practice.

REFERENCES

Aertgeerts B, Buntinx F, Kester A (2004) The value of the CAGE in screening for alcohol abuse and alcohol dependence in general clinical

populations: a diagnostic meta-analysis. J Clin Epidemiol 57:30–39.

American Psychiatric Association (1994) Substance-related disorders, in

Diagnostic and Statistical Manual of Mental Disorders. 4th ed., pp 175– 204. American Psychiatric Association, Washington, DC.

Anton RF, O’Malley SS, Ciraulo DA, Cisler RA, Couper D, Donovan

DM, Gastfriend DR, Hosking JD, Johnson BA, LoCastro JS,

Longabaugh R, Mason BJ, Mattson ME, Miller WR, Pettinati HM, Randall CL, Swift R, Weiss RD, Williams LD, Zweben A, COMBINE

Study Research Group (2006) Combined pharmacotherapies and

behavioral interventions for alcohol dependence: the COMBINE

study, a randomized controlled trial. JAMA 295:2003–2017. Babor TF, Higgins-Biddle JC, Dauser D, Burleson JA, Zarkin GA, Bray

J (2006) Brief interventions for at-risk drinking: patient outcomes and

cost-effectiveness in managed care organizations. Alcohol Alcohol 41:624–631.

Beich A, Gannik D, Malterud K (2002) Screening and brief intervention

for excessive alcohol use: qualitative interview study of the experiences

of general practitioners. BMJ 325:870. Berwick DM (2003) Disseminating innovations in health care. JAMA

289:1889.

Bradley KA, Curry SJ, Koepsell TD, Larson EB (1995) Primary and

secondary prevention of alcohol problems: U.S. internist attitudes and practices. J Gen Intern Med 10:67–72.

Cherpitel CJ (2002) Screening for alcohol problems in the U.S. general

population: comparison of the CAGE, RAPS4, and RAPS4-QF by gender, ethnicity, and service utilization. Alcohol Clin Exp Res

26:1686–1691.

Crabtree BF, Miller WL, Tallia AF, Cohen DJ, Diccio-Bloom B,

McIlvain HE, Aita VA, Scott JG, Gregory PB, Stange KC, McDaniel RR Jr (2005) Delivery of clinical preventive services in family medicine

offices. Ann Fam Med 3:430–435.

D’Amico EJ, Paddock SM, Burnam A, Kung FY (2005) Identification

of and guidance for problem drinking by general medical providers: results from a national survey. Med Care 43:229–236.

Deitz D, Rohde F, Bertolucci D, Dufour M (1994) Prevalence of screen-

ing for alcohol use by physicians during routine physical examinations. Alcohol Health Res World 18:162–168.

Division of Adult and Community Health NCfCDPaHP Behavioral Risk

Factor Surveillance System Online Prevalence Data, 1995–2003. 2005.

Available at: http://apps.nccd.cdc.gov/brfss/. Accessed May 16, 2007. Donovan DM, Kivlahan DR, Doyle SR, Longabaugh R, Greenfield SF

(2006) Concurrent validity of the Alcohol Use Disorders Identification

Test (AUDIT) and AUDIT zones in defining levels of severity among

out-patients with alcohol dependence in the COMBINE study. Addiction 101:1696–1704.

Ewing JA (1984) Detecting alcoholism: the CAGE questionnaire. JAMA

252:1905–1907.

Fleming MF, Barry KL, Manwell LB, Johnson K, London R (1997) Brief physician advice for problem alcohol drinkers: a randomized

controlled trial in community-based primary care practices. JAMA

277:1039–1044. Glasgow RE, Lichtenstein E, Marcus AC (2003) Why don’t we see

more translation of health promotion research to practice? Rethinking

the efficacy-to-effectiveness transition. Am J Public Health 93:

1261–1267. Grant BF, Dawson DA (2006) Introduction to the national epidemio-

logic survey on alcohol and related conditions. Alcohol Res Health

29:74–78. Grant BF, Dawson DA, Stinson FS, Chou SP, Dufour MC, Pickering

RP (2004) The 12-month prevalence and trends in DSM-IV alcohol

abuse and dependence: United States, 1991–1992 and 2001–2002.

Drug Alcohol Depend 74:223–234. Harford TC, Grant BF, Yi HY, Chen CM (2005) Patterns of DSM-IV

alcohol abuse and dependence criteria among adolescents and adults:

results from the 2001 national household survey on drug abuse.

Alcohol Clin Exp Res 29:810–828. McGlynn EA, Asch SM, Adams J, Keesey J, Hicks J, DeCristofaro A,

Kerr EA (2003) The quality of health care delivered to adults in the

United States. N Engl J Med 348:2635. Moyer A, Finney JW (2005) Brief interventions for alcohol problems:

factors that facilitate implementation. Alcohol Res Health 28:44–50.

Moyer A, Finney JW, Swearingen CE, Vergun P (2002) Brief inter-

ventions for alcohol problems: a meta-analytic review of controlled investigations in treatment-seeking and non-treatment-seeking

populations. Addiction 97:279–292.

National Institute on Alcohol Abuse and Alcoholism (2005) Helping

Patients Who Drink Too Much: A Clinician’s Guide. U.S. Department of Health and Human Services, Washington, DC.

Newson R (2006) Confidence intervals for rank statistics: Somers’ D and

extensions. Stata J 6:309–334.

Nilsen P, Aalto M, Bendtsen P, Seppa K (2006) Effectiveness of strategies to implement brief alcohol intervention in primary

healthcare. Scand J Prim Health Care 24:5–15.

Ostbye T, Yarnall KS, Krause KM, Pollak KI, Gradison M, Michener JL (2005) Is there time for management of patients with chronic

diseases in primary care? Ann Fam Med 3:209–214.

Reinert DF, Allen JP (2007) The alcohol use disorders identification test:

an update of research findings. Alcohol Clin Exp Res 31:185–199. Robins L, Cottler L, Bucholz K, Compton W (1996) Diagnostic Interview

Schedule for DSM-IV. Department of Psychiatry, Washington

University School of Medicine, St. Louis, MO.

Robins L, Helzer J, Croughan J, Ratclif K (1981) National institute of mental health diagnostic interview schedule—its history, characteris-

tics, and validity. Arch Gen Psychiatry 38:381–389.

Rogers EM (2003) Diffusion of Innovations. 5th ed. Free Press, New York.

Rush BR, Urbanoski KA, Allen BA (2003) Physicians’ enquiries into

their patients’ alcohol use: public views and recalled experiences.

Addiction 98:895–900. Seale JP, Boltri JM, Shellenberger S, Velasquez MM, Cornelius M,

Guyinn M, Okosun I, Sumner H (2006) Primary care validation of

a single screening question for drinkers. J Stud Alcohol 67 (suppl):

778–784. Spandorfer JM, Israel Y, Turner BJ (1999) Primary care physicians’

views on screening and management of alcohol abuse: inconsistencies

with national guidelines. J Fam Pract 48:899–902.

Stata Corporation (2005) Stata Statistical Software: Release. Stata Corporation, College Station, TX.

Sussman AL, Williams RL, Leverence R, Gloyd PW Jr., Crabtree BF

(2006) The art and complexity of primary care clinicians’ preventive counseling decisions: obesity as a case study. Ann Fam Med 4:

327–333.

Taj N, Devera-Sales A, Vinson DC (1998) Screening for problem

drinking: does a single question work? J Fam Pract 46:328–335. Vinson DC, Elder NC, Werner JJ, Vorel LA, Nutting PA (2000)

Alcohol-related discussions in primary care: a report from ASPN.

J Fam Pract 49:28–33.

Vinson DC, Maclure M, Reidinger C, Smith GS (2003) A population- based case–crossover and case–control study of alcohol and the risk

of injury. J Stud Alcohol 64:358–366.

1397SIMPLIFYING ALCOHOL ASSESSMENT

Volk RJ, Cantor SB, Steinbauer JR, Cass AR (1997) Item bias in the

CAGE screening test for alcohol use disorders. J Gen Intern Med 12:763–769.

Wallace P, Cutler S, Haines A (1988) Randomised controlled trial of

general practitioner intervention in patients with excessive alcohol consumption. BMJ 297:663–668.

Watkins K, Pincus HA, Tanielian TL, Lloyd J (2003) Using the chronic

care model to improve treatment of alcohol use disorders in primary

care settings. J Stud Alcohol 64:209–218.

Whitlock EP, Polen MR, Green CA, Orleans T, Klein J (2004) Behav-

ioral counseling interventions in primary care to reduce risky/harmful alcohol use by adults: a summary of the evidence for the U.S. preven-

tive services task force. Ann Intern Med 140:557–568.

Williams RH, Vinson DC (2001) Validation of a single question screen for problem drinking. J Fam Pract 50:307–312.

Yarnall KSH, Pollak KI, Ostbye T, Krause KM, Michener JL (2003)

Primary care: is there enough time for prevention? Am J Public Health

93:635–641.

1398 VINSON ET AL.