Opens in a new tab

CAGE Questionnaire:
Scoring and Interpretation

The CAGE questionnaire is a four-question yes or no screen for problematic alcohol use. CAGE questionnaire scoring is a simple 0 to 4 count, so this guide spends its time on what the count can and cannot say: the commonly studied two-positive screening rule, why it is not a measure of current drinking, SPSS and R code, and a model reporting paragraph.

ITEMS4
SUBSCALESNone
RESPONSE SCALEYes or no
SCORE RANGE0 to 4
DEVELOPEDEwing; validated by Mayfield et al., 1974

What the CAGE Questionnaire Measures

The four questions ask about experiences that may indicate a history of significant alcohol-related difficulties. They do not measure drinks per day. Ewing devised the questions, Mayfield and colleagues (1974) published the first validation among psychiatric patients, and Ewing (1984) described the instrument’s origin and clinical use.

It remains relevant in research because it is short, easy to administer and historically prominent, and CAGE questionnaire scoring takes seconds once the responses are coded. Its focus on established alcohol-related problems can make it less suited to detecting current risky consumption in some populations. If a study needs quantity and frequency, an instrument built for that purpose, such as the AUDIT, fits better than reading the CAGE total as a consumption score.

Structure of the CAGE

  • Items: four yes or no items, each scored 1 for an affirmative response, giving a sum from 0 to 4.
  • Subscales: none. The four items are a screening set and not four separately validated clinical subscales.
  • Direction: no reverse-scored items in the standard CAGE. Higher totals mean more endorsed concerns.
  • Time frame: affirmative answers can refer to experiences outside any fixed recent period, so document how the questions were presented.
  • Versions: adaptations such as the CAGE-AID, which extends the items to drug problems, should be distinguished from the standard CAGE and not pooled with it without version documentation.
  • Translations: use a validated version for the intended population and name the language and its source. If no validated translation exists, say so and describe how the version was adapted.

The instrument itself is available from the original publication (Mayfield et al., 1974) and the published clinical literature (Ewing, 1984).

How to Score the CAGE

  1. Confirm that the four responses are coded 1 for yes and 0 for no.
  2. Sum all four items to obtain cage_total, which ranges from 0 to 4.
  3. If a threshold such as two or more positives is used, identify the validation context and the reference outcome it was tested against.
  4. Apply a missing-item rule set by the protocol. No developer-endorsed prorating rule was verified, so the code below requires all four responses. Report missing items and never treat an unanswered question as a negative.
SPSS SYNTAX

* Illustrative syntax: standard CAGE, cage1 to cage4 coded 1 = yes, 0 = no.
COMPUTE cage_total = SUM.4(cage1 TO cage4).
COMPUTE cage_screen2 = (cage_total >= 2).
IF MISSING(cage_total) cage_screen2 = $SYSMIS.
EXECUTE.
R

# data: one row per respondent; items cage1 to cage4 coded 1 = yes, 0 = no.
items <- paste0("cage", 1:4)
stopifnot(all(vapply(data[items], function(x) all(is.na(x) | x %in% 0:1), logical(1))))
data$cage_total <- rowSums(data[items], na.rm = FALSE)
data$cage_screen2 <- ifelse(is.na(data$cage_total), NA, data$cage_total >= 2)

Interpreting CAGE Scores

Score Interpretation
0 No positive responses
1 One endorsed concern; interpretation depends on the setting
2 to 4 Meets the widely studied two-positive screening rule; diagnostic assessment may be appropriate

Source: Mayfield et al. (1974); Ewing (1984).

The original validation examined how different numbers of positive responses performed as thresholds. The score is best read as a history-oriented marker of alcohol-related problems. In populations where more severe alcohol problems are less common than in the original psychiatric sample, predictive values will differ.

The CAGE is not a DSM diagnosis and does not estimate current alcohol intake. Avoid calling a positive screen alcohol dependence without confirmatory assessment.

Psychometric Properties

Property Published finding Source
Original validation A study of 366 psychiatric admissions examined screening thresholds Mayfield et al. (1974)
Clinical review The brief identification role of the instrument was described Ewing (1984)
Diagnostic accuracy Performance varies by case definition and population, so no single sensitivity or specificity applies universally Mayfield et al. (1974)
Internal consistency With four binary items, report a suitable coefficient from your own sample if scale consistency matters Study-specific
Scope The CAGE addresses alcohol-related problems and not consumption frequency Ewing (1984)

These are source-specific findings for a given population and version. They cannot be assumed to hold as reliability or accuracy estimates in a new sample.

Using the CAGE in a Study

In prevalence research, define the denominator and the threshold before examining the data. Report the proportion above the threshold with a confidence interval, together with missing responses and any survey weighting. A yes or no screen suits logistic regression when the outcome definition matches the question asked.

When comparing groups, adjustment for age, clinical status and other covariates may be relevant, but it does not turn a cross-sectional association into a causal effect. If the CAGE is being validated against a clinical interview, report a two-by-two classification table with sensitivity and specificity, not percent agreement alone. Readers should not use the total as a proxy for drinking frequency.

Reporting the CAGE

Alcohol-related problems were screened with the four-item CAGE questionnaire (Mayfield et al., 1974; Ewing, 1984). Each affirmative response scored 1, and a total was calculated for complete records (range = 0 to 4). A score of [prespecified value] or more defined screen-positive status. The score was analyzed [method], and the study did not infer an alcohol use disorder diagnosis from this screen alone.

References

Ewing, J. A. (1984). Detecting alcoholism: The CAGE questionnaire. JAMA, 252(14), 1905-1907. https://doi.org/10.1001/jama.1984.03350140051025

Mayfield, D., McLeod, G., & Hall, P. (1974). The CAGE questionnaire: Validation of a new alcoholism screening instrument. American Journal of Psychiatry, 131(10), 1121-1123. https://doi.org/10.1176/ajp.131.10.1121

Frequently Asked Questions

Four yes or no items. Sum the affirmative responses to obtain a score from 0 to 4.

Two endorsed concerns meet a commonly used screening rule. It is not a confirmed diagnosis.

No. It describes alcohol-related concerns and not current consumption.

Give the score range, the exact positive-screen rule, the population, and the denominator and missingness.

NEXT STEP

Need CAGE Data Scored
or Checked?

OnlineSPSS is a statistical consultancy for researchers. Share your design and dataset, and a statistically trained Research Coordinator reviews it. The scope and one fixed project fee are confirmed before work begins.

No Commitment Required

Strictly Confidential

Fixed Project Fee

In-House PhD-Level Statisticians

Same-Day Response

Serving Researchers Since 2014