Single question screening for problem drinking

For years I have used a single question that has been very effective. I usually use this question in the inpatient setting. When I see a patient who already has the evidence of alcoholism (laboratory results, trauma, family history, legal history, and so forth), and it seems obvious to me, I exchange introductions and brief historical information and then merely ask the patient: “How long have you been an alcoholic?” This question establishes that I know and that the patient knows; we are just trying to determine length. With careful use I have found this to be extremely accurate and helpful.

David Parrish, MD
Bayfront FP Residency, St. Petersburg, Florida

I would like to congratulate Williams and Vinson for developing a simple, easily remembered single screening question for the detection of problem drinking. No other screening tools are universally applicable.

The AUDIT is long and not likely to be used by busy physicians. Reducing the number of questions in the AUDIT to 5 yielded a sensitivity of 0.79, a specificity of 0.95,¹ a positive predictive value of 0.73, and a false-positive rate of less than 4%.² The AUDIT has since been reduced to 3 questions (AUDIT-C) and has been shown to be a valid primary care screening test for heavy drinking or active alcohol abuse or dependence.³ The AUDIT-C has been shown to be superior to the AUDIT for detecting heavy drinking (P=.03), but the AUDIT is better than the AUDIT-C for detecting active alcohol abuse or dependency (P <.001).

Because of the low participation rate in the study, it would be of interest if this were repeated and compared with either the AUDIT or the AUDIT-C in the hope of producing a simple, reliable, and validated alcohol screening test. The ultimate aim is to identify problem drinkers and refer for specialist alcohol counseling.

Dr Richard M. Lynch
York District Hospital, United Kingdom

REFERENCES

1. Mayfield D, McLeod G, Hall P. Am J Psychiatry 1974; 131:1121-23.
2. Piccinelli M, Tessari E, Bortolomasi M, et al. BMJ 1997; 314:420-24.
3. Bush K, Kivlahan DR, McDonell MB, Fihn SD, Bradley K. Arch Intern Med 1998; 158:1789-95.

Drs Williams and Vinson respond:

Dr Lynch suggests further study comparing the AUDIT and our single question. We agree. Future research should explore issues beyond sensitivity and specificity, such as the instruments’ acceptability to physicians and their effects on the alcohol-related discussions that follow positive (or negative) screening results.

We disagree with Dr Lynch that all problem drinkers need to be referred for counseling.^1,2 Some will accept referral and benefit from it, but brief interventions by primary care physicians are effective.

In the situation Dr Parrish describes, he already knows the patient has an alcohol problem. His question is not a screening question but a way of going directly to a discussion of the patient’s major problem. This approach will miss patients with hazardous drinking who have not yet experienced adverse consequences and those whose alcohol-related problems are not obvious to the physician. A physician who waits until the patient has obvious biomedical problems may miss good opportunities for earlier intervention that can prevent those problems.³

We encourage Dr Parrish and others to study his and other approaches. Our study examined only a small part of how physicians and patients can talk about alcohol more effectively and efficiently. Screening is an important part, but it is only the first step. Many questions remain, and Dr Parrish’s approach deserves further study. Regardless of how we screen, what do we do with the patient who screens positive? If it came from other physicians, would Dr Parrish’s question engage the patient in honest conversation, as it does coming from him? Would patients be intimidated by it or greet it with relief (as in “At last, a doctor who knows my problem!”)? And the most important question: What approaches more effectively help the problem drinker move toward change?

Randy Williams, PhD
Daniel C. Vinson, MD, MSPH
University of Missouri–Columbia

REFERENCES

1. Wallace P, Cutler S, Haines A. BMJ 1988; 297:663-68.
2. Fleming MF, Barry KL, Manwell LB, Johnson K, London R. JAMA 1997; 277:1039-44.
3. Fleming MF, Mundt MP, French MT, Manwell B, Stauffacher EA, Barry KL. Med Care 2000; 38:7-18.