Edvard Munch, Self-Portrait with a Bottle of Wine, 1906
Patients who say they self-medicate may be describing pleasure, habit, or relief from withdrawal
STUDY: Marchi M and Galeazzi GM, Br J Psychiatry 2026
STUDY TYPE: Narrative review
FUNDING: Independent
Background
In 1985, psychiatrist Edward Khantzian proposed that people use drugs to relieve painful emotions. The idea caught on because it sounds compassionate and explains why mental illness and addiction often travel together.
Now, two Italian psychiatrists argue that the hypothesis is a misleading assumption, reviewing epidemiological, longitudinal, and moment-to-moment motive studies.
Summary
The self-medication story assumes four things, and each of these ideas does not hold up to the evidence:
- Patients use substances mainly to treat a symptom
- They choose the drug that fits the symptom, such as sedatives for sleep or alcohol for anxiety
- The drug helps
- The symptoms came first
Drug choice is shaped more by availability, cost, peers, and habit than by pharmacology. Pleasure and reward are strong influencers. In a meta-analysis of over 130,000 people, drinking for enjoyment predicted alcohol problems as strongly as drinking to cope.
Relief can mislead. Smokers feel cigarettes calm them, yet quitting improves depression and anxiety. Some of that “relief” may be the end of withdrawal.
The authors see three main risks:
- Paternalism: Treating patients as misguided self-prescribers and ignoring the rewarding aspects
- Diagnostic drift: Thinking that stimulant use suggests ADHD, alcohol use suggests social anxiety, or that cannabis use suggests a problem with anxiety or agitation
- Distraction: The “self-medication” idea diverts attention from other causes of substance use: trauma, shame, loneliness, and housing problems
The model can also push clinicians to prescribe substitute medications, such as SSRIs or gabapentinoids for alcohol use disorder, when better-proven addiction treatments exist.
Practice Implications
- When a patient says a drug helps, ask what they like about it, what it changes, and what it costs them. A calming response to cocaine does not diagnose ADHD.
- Substance use can point toward underlying diagnoses, but not in the way that intuitive ideas about would suggest. For example, studies tell us that:
- In bipolar disorder, cocaine use points more to bipolar I than II, while cannabis points more to bipolar II
- In the large BRIDGE study, the link between depression and alcohol use was entirely explained by mixed features
—Chris Aiken, MD
Director, Psych Partners
Editor in Chief, Carlat Psychiatry Report
