Screening for alcohol problems: what makes a test effective?

Alcohol Research & Health, Wntr, 2004 by Scott H. Stewart, Gerard J. Connors

Common Alcohol Screening Instruments in Medical Settings*

                                                         Time to
                Population to           Number of Items  Administer
Measure         Be Screened             (Subscales)      (Minutes)

Alcohol Use     Adults                  10 (3)            2
Disorders
Identification
Test (AUDIT)
CAGE            Adults and               4               <1
Questionnaire   adolescents > 16 years
Michigan        Adults and adolescents  25                8
Alcoholism
Screening
Test (MAST)
Self-           Adults                  35 (2)            5
Administered
Alcoholism
Screening Test
(SAAST)

* Briefer versions of some of these screening instruments (e.g., the
MAST and SAAST) also have been tested.
SOURCE: National Institute on Alcohol Abuse and Alcoholism (NIAAA).
Assessing Alcohol Problems: A Guide for Clinicians and Researchers, 2d
ed. NIH Pub. No. 03-3745, Washington, DC: U.S. Dept. of Health and
Human Services, Public Health Service, 2003.

               Disease Present?
Test Result  Yes             No

Positive     True positive   False positive
             (TP)            (FP)
Negative     False negative  True negative
             (FN)            (TN)

Characteristics of Screening Tests
Sensitivity* = TP / TP   FN
Specificity* = TN / FP   TN
Overall accuracy = (TP   TN) / (TP   FP   FN   TN)
Positive predictive value = TP / TP   FP
Negative predictive value = TN / TN   FN
Positive likelihood ratio = sensitivity / (1-specificity)
Negative likelihood ratio = (1-sensitivity) / specificity

Figure 1 Definitions of terms used to describe characteristics of
screening tests.**

*Sensitivity and specificity are mathematically expressed as numerical
values ranging between 0 and 1.
**This illustration assumes that a test can only yield positive or
negative results.

Test Scores on the AUDIT

A. Cutoff score of 4 or more for at-risk drinking

              Distribution
              Among Four        Sensitivity         Specificity
Test Results  Groups            (TP / TP   FN)      (TN / TN   FP)

9 positive    5 true positive   5 / (5   1) = 0.83  10 / (10   4) = 0.71
              (TP)
11 negative   4 false positive
              (FP)
              10 true negative
              (TN)
              1 false negative
              (FN)

B. Cutoff score of 8 or more for at-risk drinking

              Distribution
              Among Four        Sensitivity         Specificity
Test Results  Groups            (TP / TP   FN)      (TN / TN   FP)

4 positive    3 true positive   3 / (3   3) = 0.50  13 / (13   1) = 0.93
              (TP)
16 negative   1 false positive
              (FP)
              13 true negative
              (TN)
              3 false negative
              (FN)

C. Cutoff score of 10 or more for at-risk drinking

              Distribution
              Among Four        Sensitivity         Specificity
Test Results  Groups            (TP / TP   FN)      (TN / TN   FP)

1 positive    1 true positive   1 / (1   5) = 0.16  14 / (14   0) = 1.0
              (TP)
19 negative   0 false positive
              (FP)
              14 true negative
              (TN)
              5 false negative
              (FN)

Figure 2 Illustration of the effects of changes in the cutoff score of
a screening test (e.g., the AUDIT) on the test's sensitivity and
specificity. Among the 20 hypothetical people screened, 6 meet the gold
standard criteria for at-risk drinking (green), and 14 are nonrisk
drinkers (blue). Numbers below the hypothetical people indicate their
test scores on the AUDIT.