No Causal Link Between Stimulants and Psychotic Experiences in Kids With ADHD

—
Findings offer reassurance for clinicians and families, researcher said

by
Jennifer Henderson, Enterprise & Investigative Writer, MedPage Today
May 12, 2025 • 3 min read

There was no evidence to support a causal relationship between a common treatment for attention deficit-hyperactivity disorder (ADHD) and psychotic experiences in children, according to a large observational study.

In unweighted analyses, stimulant prescription for children, ages 9-14 years, was tied to subsequent psychotic experiences with an odds ratio (OR) 1.46 (95% CI 1.15-1.84), reported Ian Kelleher, MD, PhD, of the University of Edinburgh in Scotland, and colleagues.

But the reverse was also true in that baseline psychotic experiences predicted subsequent stimulant treatment (OR 1.93, 95% CI 1.57-2.37), they stated in Pediatrics.

Ultimately, when applying doubly robust estimation, there was no evidence of a causal effect of stimulant prescription on the subsequent occurrence of psychotic experiences (OR 1.09, 95% CI 0.71-1.56), they said.

“Although treatment with stimulants has been found to be effective at reducing ADHD symptoms in children and adolescents, concerns have been raised that stimulants could increase risk for psychotic experiences (ie, subclinical hallucinations and delusions),” Kelleher and colleagues wrote.

Indeed, in 2006, the FDA issued an alert that such experiences “can occur in some patients with no identifiable risk factors, at usual doses” of stimulants to treat ADHD, and the agency called for more prominent warnings and additional monitoring of this risk.

“Some research has suggested that stimulant medication treatment can increase the risk of psychotic experiences in kids with ADHD,” Kelleher further told MedPage Today in an email. “Understandably, that can be quite worrying to young people, parents, and doctors. But research to date has not done a good job at figuring out whether it’s actually stimulants that cause these experiences or it’s something else altogether. For example, stimulants are more likely to be prescribed for kids with more severe ADHD. And it could be that ADHD severity, rather than stimulants, increases your risk of psychotic experiences.”

“These findings provide reassurance to clinicians, young people, and families that, at standard doses, stimulant medications are unlikely to be a cause of psychotic experiences,” he added.

Overall, there were 8,391 kids included in the analytical sample. Just shy of 19% reported one or more psychotic experience of at least moderate distress at baseline, and the most common psychotic experiences were paranoid ideas, auditory hallucinations, and perceiving a supernatural presence.

Kids included in the sample had no prior prescription of stimulants (methylphenidate, dexmethylphenidate, amphetamine, dextroamphetamine, or lisdexamfetamine) in the past year. They were required to have attended a follow-up visit at 1 year, during which psychotic experiences were measured using the Prodromal Questionnaire-Brief (PQ-B). The outcome was psychotic experiences of at least moderate distress.

At the 1-year follow-up visit, 15.1% of participants reported this outcome, and 5.5% of the sample had received a first stimulant prescription by the follow-up visit (as reported by a caregiver).

Kids who were prescribed stimulants were more likely to be male, and had higher scores on psychopathology measures at baseline, Kelleher and colleagues noted. And psychopathology, including ADHD symptoms, also predicted later psychotic experiences, as did age, parental income, and race.

The authors explained that they used a “target trial emulation framework whereby observational analyses are designed to explicitly emulate a hypothetical,” randomized controlled trial, specifically the Adolescent Brain Cognitive Development Study.

Treatment propensities were derived using covariates indexing demographic factors and mental illness severity. The average causal effect of first stimulant prescription on psychotic experiences by the 1-year follow-up visit was derived using inverse probability of treatment weighting followed by standardization (doubly robust estimation).

Kelleher’s group noted that the sample size of individuals who were prescribed stimulants may have limited the precision of causal effect estimates. They also noted that they did not have data on cumulative doses of stimulants, and that they did not investigate whether kids prescribed stimulant medications were also prescribed nonstimulants.

Furthermore, data on adherence to stimulant medications were not available. Despite controlling for confounding factors, there may have been residual confounding from other unmeasured factors and/or measurement error.