Chang · Journal of child psychology and psychiatry, and allied disciplines 2014 · nationwide longitudinal cohort study · n=38,753

Stimulant ADHD medication and risk for substance abuse.

Cited 220 times in the scientific literature.

Level 3 - non-randomized controlled study

Nationwide non-randomized longitudinal cohort study.

PubMed 25158998 · doi:10.1111/jcpp.12164 · record verified 2026-08-26

What was done

Using Swedish national registers, researchers analyzed individuals born between 1960 and 1998 diagnosed with ADHD (26,249 men and 12,504 women; total n = 38,753). They investigated the longitudinal association between stimulant ADHD medication prescribed in 2006 and substance abuse events occurring in 2009. Substance abuse was defined through registry records of substance-related hospital admissions, criminal convictions, or deaths. Statistical models controlled for concurrent medication use in 2009 and other covariates.

What was found

Prescription of ADHD medication was not linked to elevated substance abuse. Participants prescribed ADHD medication in 2006 had a 31% lower rate of substance abuse in 2009 compared to unmedicated peers after adjusting for 2009 medication and covariates (hazard ratio: 0.69; 95% confidence interval: 0.57–0.84). Longer treatment duration was associated with lower substance abuse rates. Risk reductions were also noted for children and in analyses of concomitant short-term abuse, though specific numerical estimates for these subsets were not reported in the abstract.

Why it matters

These nationwide population data address long-standing safety concerns by demonstrating that stimulant therapy for ADHD is associated with a reduced, rather than increased, risk of subsequent substance abuse.

Limits

The study is observational and susceptible to residual confounding. The outcome definition captured only severe events (substance-related crimes, hospital visits, or mortality), missing milder or untreated misuse. Specific numerical data for duration effects and child subgroups were not provided in the abstract.

Cited by