The ASSIST Screening Test for Drug Use: A Clinician Guide

The Alcohol, Smoking and Substance Involvement Screening Test (ASSIST) is a World Health Organization instrument designed to screen for problematic use across multiple substances in a single administration. The version built into LetPsyc is a drug-only adaptation: it screens for illicit and non-medical prescription drug use across ten substance classes and for injecting drug use. It does not screen for tobacco or alcohol - clinicians who need those should use AUDIT or the full WHO ASSIST v3.0 alongside it. This guide covers what the tool measures, exactly how per-substance scoring works, its risk tiers, and where it fits against a narrower instrument like the DAST-10.

What the ASSIST measures

The ASSIST does not produce a single overall score the way a depression or anxiety screener does. Instead it produces a separate involvement score for each substance class a patient reports ever using, covering domains such as frequency of use, urge or desire to use, health/social/legal/financial problems attributed to use, failure to meet expectations because of use, concern expressed by others, and unsuccessful attempts to cut down or stop. Because each drug class is scored independently, the ASSIST can distinguish, for example, moderate-risk cannabis use from high-risk opioid use in the same patient, rather than collapsing everything into one number.

The ten substance classes and skip logic

Patients are first asked, for each of the ten substance classes, whether they have ever used it in their lifetime. This lifetime-use question acts as a skip gate: only substances the patient endorses at this first step trigger the full set of follow-up questions for that substance. A patient who denies ever using a given class skips straight to the next substance, which keeps administration efficient - most patients end up answering the detailed items for only one or two substance classes rather than all ten.

How to score the ASSIST

For every substance class the patient endorses at the lifetime-use gate, the follow-up questions are summed to produce that substance's own involvement score. This is the single most important scoring rule to get right: the ASSIST does not have one grand total across all drugs. A patient who reports problematic use of two substances receives two separate scores, each interpreted against its own risk tier, not added together into a combined figure. Clinicians and administrators new to the ASSIST sometimes assume a single summary number exists; it does not, and averaging or summing across substances produces a meaningless figure.

Risk tiers and the cannabis exception

Each substance's score sorts into a risk tier - broadly a lower-risk band, a moderate-risk band, and a high-risk band - that indicates whether no intervention, a brief intervention, or a more intensive assessment and referral is appropriate for that specific substance. One detail matters for accurate scoring: cannabis uses a lowest-risk band that is one point higher than every other substance class (0-4 for cannabis versus 0-3 for the other nine classes). This reflects cannabis's distinct risk profile relative to substances like opioids or amphetamines, and it means a raw score that would tip another substance into the moderate-risk tier can still sit within cannabis's lower-risk band. Getting this exception wrong is the most common per-substance scoring error.

Injecting drug use is a separate risk pathway

The ASSIST includes a distinct question asking whether the patient has ever injected a drug. This item is not added into any substance's involvement score. Instead, a positive response flags a separate injection-related risk pathway - reflecting the additional harms associated with injecting (overdose, blood-borne virus transmission, vein damage) that are not captured by a substance-specific involvement total. A clinician reviewing an ASSIST report should treat a positive injection flag as its own clinical signal, independent of whichever substances scored in the moderate or high-risk tiers.

ASSIST vs DAST-10

Clinicians often weigh the ASSIST against the DAST-10. The DAST-10 produces one quick global score (0-10) across all drug use combined, which is fast to administer and easy to trend over time, but it cannot tell you which substance is driving the concern. The ASSIST takes longer because of its skip logic across ten classes, but it returns substance-specific detail - useful when a treatment plan needs to target a particular drug class, or when a patient uses more than one substance with very different risk levels. Many services use the DAST-10 as a fast initial flag and reserve the ASSIST for patients who need substance-by-substance detail.

Who administers it, and how long does it take

The ASSIST is typically completed with support from a clinician or trained staff member, either as a brief structured interview or self-administered on a tablet or device with a clinician available to clarify items. Administration time varies with how many substances a patient endorses at the lifetime-use gate: a patient who denies most substances finishes in a few minutes, while a patient with polysubstance use answering the full follow-up set for several classes takes longer.

Limitations

The most important limitation to communicate to any team adopting this tool is the one built into its name in this context: the LetPsyc ASSIST implementation does not screen tobacco or alcohol. A clinic relying on it alone will systematically miss alcohol-related risk, which needs to be screened separately. Beyond that, like any self-report instrument, the ASSIST depends on honest disclosure, and its per-substance risk tiers indicate the likely need for intervention rather than a diagnosis of a substance use disorder - a moderate or high-risk score should prompt further clinical assessment, not stand in for one.

Using the ASSIST digitally

Digital administration is well suited to the ASSIST's skip logic: the platform can automatically route a patient past substances they deny at the lifetime-use gate, present follow-up items only for endorsed substances, and calculate each substance's involvement score and risk tier instantly - including correctly applying the cannabis exception without manual lookup. The injection-use flag can be surfaced prominently in the clinician report so it is never missed among multiple substance scores. This reduces the scoring burden that makes the ASSIST time-consuming on paper, while supporting measurement-based care across repeat administrations.

Key takeaways

• LetPsyc's ASSIST is a drug-only adaptation - it does not screen tobacco or alcohol; use AUDIT or the full WHO ASSIST v3.0 for those.
• Scoring is per-substance: each endorsed drug class gets its own involvement score, not one combined total.
• Cannabis has a one-point-higher lowest-risk band (0-4) than every other substance class (0-3).
• Injecting drug use is flagged through a separate risk pathway, not summed into any substance score.
• A moderate or high-risk score indicates the likely need for intervention, not a diagnosis.