Alcoholism: the keys to the CAGE

Am J Med. 1993 May;94(5):520-523. doi: 10.1016/0002-9343(93)90088-7.


Purpose: To compare the efficiency of screening for alcoholism using two different introductions to the CAGE questions.

Patients: Forty-three alcoholics on inpatient medical and surgical services were identified through random distribution of the Michigan Alcoholism Screening Test.

Methods: Participants were randomized to two groups receiving different introductions to the CAGE questions. Group I was introduced by a simple open-ended question. Group II patients were asked first to quantitate the volume and frequency of their drinking. The outcome measures were the scores on the four CAGE questions recorded by an independent observer.

Results: Eighteen of 21 (86%) patients in Group I admitted to prior attempts to cut back on their drinking as compared with 8 of 22 (36%) in Group II (p = 0.002). Likewise 10 of 21 (49%) patients in the first group admitted to annoyance compared with just 3 of 22 (14%) in Group II. The average scores per patient in the two groups were significantly different: Group I = 2.52 versus Group II = 1.23 (p = 0.0002). Using a screening CAGE score of two or more to identify an alcoholic patient, the open-ended introduction identified 95% of the alcoholics compared with only 32% when the closed-ended introduction was used (p < 0.0001).

Conclusion: Our study demonstrates that the sensitivity of the CAGE questionnaire in screening for alcoholism is dramatically enhanced by an open-ended introduction. In contrast, the sensitivity of these questions is greatly reduced when preceded by inquiries that seek to define the quantity and frequency of drinking.

Publication types

  • Clinical Trial
  • Randomized Controlled Trial

MeSH terms

  • Alcoholism / diagnosis*
  • Alcoholism / psychology
  • Female
  • Humans
  • Male
  • Mass Screening / methods
  • Middle Aged
  • Observer Variation
  • Predictive Value of Tests
  • Random Allocation
  • Sensitivity and Specificity
  • Surveys and Questionnaires