There is a lot of research completed or underway regarding various substance use screening and assessment tools. Some of that work is collected below.
Evidence of Effectiveness of Screening Tests
The following two tables provide some quantitative data regarding the effectiveness of four screening tools, including multiple derivatives of several tools: the AUDIT, CAGE, ASSIST, and DAST. This information may be useful when comparing screening tools to other tests, such as examination of biomarkers.
Different studies, with varied patient populations (and drugs for the ASSIST) found different sensitivity and specificity for the above noted screening instruments. The ranges given reflect the varied results.
Other “Screening Tests”: Biomarkers & Patient Reports
Screening in a medical setting involves at least two components: biomarkers and patient reports. Biomarkers are objective evidence (not necessarily proof) that an individual may abuse drugs, such as a urine drug screen or transaminases. Patient reports are based on questionnaires designed to get a “big picture” of the individual’s substance use and to identify potential red flags.
Screening for Unhealthy Alcohol Use: A Questionnaire Is Best
Given that there are multiple potential means of screening, a study was conducted to determine the best way to screen for unhealthy alcohol use.
- To compare blood tests with self-report questionnaires for alcohol screening.
- AUDIT administered to 1794 men in Wales.
- Blood tests on 112 men who screened positive on the AUDIT (score of >=8) and 82 who screened negative.
- For unhealthy alcohol use, sensitivity and specificity were highest for the AUDIT.
- For alcohol dependence, sensitivity and specificity were also highest for the AUDIT (84% and 83%, respectively).
- The cost of identifying a patient with unhealthy alcohol use was lowest for the AUDIT.
|Mean corpuscular volume (MCV)||32%|
|Aspartate aminotransferase (AST)||20%|
Do Biomarkers Improve the Accuracy of Alcohol Screening in Acutely Injured Adults?
Although the AUDIT displayed higher sensitivity than biomarkers in the previous study, this doesn’t mean that biomarkers don’t have utility. Another study investigated whether biomarkers improve the accuracy of alcohol screening among a specific population (acutely injured adults). The results are displayed below and are separated by gender.
Screening for Alcohol Problems in Primary Care – A Systematic Review
Before moving on to actual approaches to the screening process, we wanted to share some data related to the effectiveness of several different screening tools and their application to various special populations in the primary care setting. These data are from a systematic review of 38 studies of screening for alcohol misuse by adults in primary care settings.
- The Alcohol Use Disorders Identification Test (AUDIT) was most effective in identifying subjects with at-risk, hazardous, or harmful drinking (sensitivity, 57%-97%, specificity, 78%-96%) with a cut off score of >/= 8.
- CAGE questions proved superior for detecting current alcohol abuse and dependence (sensitivity, 77%-94%; specificity, 79%-97%) with cut off score of >/= 2.
- MAST (Michigan Alcoholism Screening Test) and SMAST (Short Michigan Alcoholism Screening Test) reported mixed results (sensitivity 68%, specificity 92% for at risk drinking and sensitivity 38%-82%, specificity 88%-97% for lifetime alcohol abuse/dependence with cut off score of >= 2).
- TWEAK Questionnaire with reported sensitivity of 75% and specificity of 90% in general primary care population (Cherpitel, 1998).
- TWEAK Questionnaire appears to be the optimal screening questionnaire for identifying women with heavy drinking or alcohol abuse and dependence in racially mixed populations with sensitivity of 67%-80%, specificity of 90%-94% with cut off score of >/= 3 (Bradley, 1998).
- CRAFFT instrument was specifically designed to screen for alcohol and drug problems in adolescents.
- In a cross-sectional study of 358 young persons with cut off score of 2, the CRAFFT had sensitivity of 94%, specificity of 33% (Cook, 2005).