How this instrument works
This instrument asks ten questions in three groups: three cover how much and how often someone drinks, three ask about dependence symptoms like being unable to stop once starting or needing a drink to get going in the morning, and four ask about harm already caused — guilt, blackouts, injuries to yourself or someone else, and whether a relative, friend, or doctor has voiced concern. Eight of the ten items score from 0 to 4; the last two, about injury and others' concern, score only 0, 2, or 4, since their answer choices are just no, yes-but-not-recently, and yes-recently. Add all ten and the sum runs from 0 to 40.
Saunders, Aasland, Babor, de la Fuente, and Grant published the questionnaire in 1993 through a World Health Organization collaborative project spanning six countries, aiming for a screen that worked across different drinking cultures rather than one built around a single country's habits. The WHO manual that followed sets out four risk zones for the total: 0-7 is Zone I, low risk; 8-15 is Zone II, hazardous drinking where brief advice is recommended; 16-19 is Zone III, harmful drinking calling for brief counseling and continued monitoring; and 20-40 is Zone IV, a range suggesting possible dependence and warranting referral to a specialist for further assessment. Those exact bands, not the looser 8-14/15-plus split sometimes quoted elsewhere, are what the original manual specifies.
This site also carries a shorter version built from just the first three consumption items here — how often someone drinks, how many drinks on a typical day, and how often six or more happen at once. That shorter form is faster and works well as an initial pass, but it can only see consumption; it has no way to catch the seven markers this longer version adds, the morning drink, the failed attempt to cut down, the guilt after a heavy night, the blackout, the injury, the worried friend. A high reading on the short form says drinking is heavy; a high reading here says something more specific about whether that drinking has already produced dependence or harm.
- Answer each of the ten items using the frequency or severity options shown for that question.
- Items 9 and 10 use a different scale — 0 for no, 2 for yes but not in the past year, 4 for yes in the past year.
- The total sums automatically as you answer; leaving an item at its default of 0 counts it as the lowest response.
- Read the total (0-40) against the WHO's four zones to see which range it falls in.
Worked example — a total of 13, WHO Zone II
A mix of moderate answers — 3, 2, 2, 2, 1, 1, 1, 1, 0, and 0 across the ten items — adds straight across: 3+2+2+2+1+1+1+1+0+0 = 13. That lands in Zone II (8-15), hazardous drinking, the WHO manual's threshold for offering brief advice rather than watching and waiting.
The two extremes bracket it cleanly. Answering 0 on every item — no concerning answer anywhere — sums to 0, Zone I, low risk. Answering the maximum on all ten instead — 4 on every item, including the top response on items 9 and 10 — sums to 40, the highest reading the test allows, deep into Zone IV (20-40), where the manual recommends referral to a specialist for a fuller diagnostic look.
Questions
What are the WHO's four AUDIT risk zones?
Zone I is 0-7, low risk, often needing nothing beyond simple alcohol education. Zone II is 8-15, hazardous drinking, where the manual recommends brief advice. Zone III is 16-19, harmful drinking, calling for brief counseling plus continued monitoring. Zone IV is 20-40, where dependence becomes likely enough that referral to a specialist is recommended. These four bands come directly from the WHO manual, not the 8-14/15-plus split that sometimes circulates instead.
How is the full AUDIT different from AUDIT-C?
AUDIT-C, also on this site, is just the first three questions here — frequency, typical quantity, and how often six-plus drinks happen at once — scored 0-12. This longer version keeps those three and adds seven more covering dependence symptoms, like being unable to stop once starting or needing a morning drink, and alcohol-related harm, like guilt, blackouts, injuries, and others' concern. The short form screens consumption fast; this one can also catch dependence and harm a consumption-only reading has no way to see.
Why do items 9 and 10 only allow scores of 0, 2, or 4?
Because their answer choices aren't a frequency scale like the other eight items — they're a three-way history question: no, yes but not in the past year, or yes in the past year. WHO scoring assigns 0, 2, and 4 to those three answers respectively, skipping the odd numbers entirely, since no meaningful answer falls between them.
Is a high AUDIT score a diagnosis of alcohol use disorder?
No. It's a screening total, not a clinical diagnosis. A reading in Zone III or IV is the manual's cue for a fuller evaluation by a trained clinician, who weighs the pattern of answers alongside history and context before reaching any conclusion. The test is designed to flag who needs a closer look, not to replace that look.
Who developed the AUDIT and why?
Saunders, Aasland, Babor, de la Fuente, and Grant published it in 1993 as the product of a World Health Organization collaborative project run across six countries, built specifically so one screening tool could work across different drinking cultures rather than being tuned to a single country's habits.
Can the AUDIT be self-administered?
Yes — it was designed to work either as a self-report form or read aloud by a clinician, and both formats see real-world use. What matters for scoring is answering all ten items honestly against the actual response choices; a partially completed form understates the true total and can push a genuinely hazardous pattern into a falsely reassuring zone.
References
- Saunders JB et al., Addiction, 1993 (PubMed)
- WHO AUDIT Manual, 2nd ed. (Babor et al., PAHO-hosted PDF)
- NIAAA — Screen and Assess: Quick, Effective Methods
Read this first: This instrument computes a screening figure from population formulas — it is not a diagnosis, and it cannot see the whole picture a clinician can. Use it to inform a conversation, not to replace one.