Assessment, Screening, and Diagnosis of Substance Use and Co-occurring Disorders
Core knowledge
DSM-5-TR defines substance use disorder across 11 criteria in four clusters: impaired control (larger amounts or longer use, unsuccessful cutting down, time spent, craving), social impairment (role failure, interpersonal problems, activities given up), risky use (hazardous situations, continued use despite physical or psychological harm), and pharmacological (tolerance, withdrawal). Two or more criteria within 12 months establish the diagnosis; severity is graded mild (2-3), moderate (4-5), or severe (6 or more). Screening is not diagnosis: AUDIT and AUDIT-C for alcohol, DAST-10 for other drugs, the NIDA Quick Screen, TAPS, and SBIRT-framed brief screens are validated and time-efficient; CAGE is short but insensitive to low-level use and performs poorly in women. Withdrawal severity instruments - CIWA-Ar for alcohol, COWS for opioids, CINA for cannabis - guide treatment intensity, not diagnosis. Co-occurring disorders require an integrated assessment: establish chronology (did psychiatric symptoms predate heavy use or appear only in intoxication or withdrawal?), use PHQ-9, GAD-7, and a structured suicide risk assessment such as the C-SSRS at intake.
Traps
- DSM-5-TR does not count tolerance and withdrawal when a substance is taken as prescribed under appropriate medical supervision; do not label a pain patient with severe SUD on that basis alone.
- A positive urine drug screen reflects exposure, not disorder.
- Immunoassays produce false positives (poppy seeds, some cold preparations); confirm with GC-MS or LC-MS before consequential conclusions.
- Conflating intoxication, withdrawal, and substance-induced psychiatric syndromes hides independent diagnoses.
Rule of thumb
Count criteria, then grade severity - never diagnose from one data point. Worked example: a patient meets four criteria (craving, larger amounts, continued use despite depression, giving up hobbies) with a CIWA-Ar of 12. The diagnosis is moderate SUD, and the score signals protocol-driven withdrawal monitoring rather than a change in severity.