ADHD medicines likely do not drive self-harm

Self-harm risk peaked before ADHD medicine started, not after.
*Multinational self-controlled case series; Level 3 (OCEBM).

Citation

Chan AYL, Beyene K, Park S, et al. BMJ Mental Health. 2026;29:1-9. doi:10.1136/bmjment-2026-302819

Background

Prior studies have disagreed on whether attention-deficit/hyperactivity disorder (ADHD) medicines raise self-harm risk. This large multinational study tested timing of self-harm before and after starting treatment.

Patients

People aged 6 years or older from Hong Kong, New Zealand, South Korea, Taiwan, and the United Kingdom who received ADHD medicine and had a first recorded self-harm event from 2001 to 2020. Exclusions included self-harm before the study period, missing age or sex, and cancer diagnosis.

Intervention

ADHD medicines, including stimulant and non-stimulant medicines.

Control

Unexposed time periods within the same person.

Outcome

First recorded self-harm, meaning intentional self-injury or self-poisoning, regardless of intent.

Follow-up Period

Individual observation between January 2001 and December 2020.

Results

Among 461,024 people prescribed ADHD medicine, 6,847 had incident self-harm and were included in the case-series analysis.

Risk period Adjusted rate ratio for self-harm Meaning
90 days before starting medicine 2.58 (95% CI 2.12 to 3.13) Highest risk
First 90 days of medicine use 1.78 (95% CI 1.23 to 2.57) Risk lower than before treatment
Later medicine use 1.27 (95% CI 1.01 to 1.60) Small remaining elevation

ADHD, attention-deficit/hyperactivity disorder; CI, confidence interval.

Analyses compared each person with themselves and adjusted for age, season, and other psychiatric medicines. The control was not another treatment.

Limitations

Administrative records may miss self-harm that never reached medical care. Prescriptions do not prove medicine was taken. Time-changing factors, such as worsening symptoms, family stress, or new depression, may still confound results. The study cannot prove protection or harm, only timing patterns.

Funding

Public research grants; several authors reported industry relationships.

Clinical Application

Do not delay indicated ADHD medication for self-harm fears; assess and monitor self-harm risk before starting and during early treatment.