Recorded self-harm was most common in the 90 days before people received a first prescription for ADHD medication, according to an analysis of health records from five health systems.
The study included 461,024 people aged 6 and older who had at least one ADHD medication prescription. Among them, 6,847 had both a prescription and a recorded self-harm event.
Compared with each person’s own unmedicated periods, the recorded self-harm rate was 2.58 times higher in the 90 days before treatment began. The rate was also higher during the first 90 days on medication, at 1.78 times the unmedicated rate, and during later treatment, at 1.27 times the unmedicated rate.
What the study compared
Researchers used a self-controlled case-series design. Instead of comparing different groups of people, they compared several periods within the same person’s life: unmedicated time, the 90 days before the first prescription, the first 90 days of treatment and later medication use.
The analysis adjusted for age, season and use of other psychiatric medicines. The pattern was seen across Hong Kong, New Zealand, South Korea, Taiwan and the United Kingdom, although the size of the highest pre-treatment rate ratio varied from 2.07 in the UK to 4.10 in Hong Kong.
A follow-up analysis at three sites found that the rate began to rise at least 100 days before the first prescription and fell afterward. The researchers concluded that the results do not support ADHD medication as the main driver of self-harm in this group.
Why timing matters
The period before treatment may include worsening ADHD symptoms or other changes that prompted a person or family to seek care. The available records did not fully capture factors such as new depression or anxiety, stressful events or changes in symptoms around treatment initiation.
Starting treatment can also involve clinical assessment, counseling, closer follow-up and care for other conditions. These changes in care may account for part of the decline after the pre-treatment peak. More frequent contact with healthcare services could also affect how often self-harm was recorded.
The overall comparison of all medicated time with all unmedicated time produced a rate ratio of 1.20, with a 95% confidence interval of 0.96 to 1.49. This result was not statistically significant. However, the rate during later treatment remained 27% above unmedicated time, and later-treatment rate ratios ranged from 0.87 in Taiwan to 2.22 in South Korea.
What was—and was not—measured
The outcome was the first self-harm event recorded in medical records. It included intentional self-poisoning or self-injury regardless of intent. Suicide deaths were not included.
Stimulant and nonstimulant medications showed the same general timing pattern. For stimulants, rate ratios were 2.74 before treatment, 1.92 during the first 90 days and 1.23 later. For nonstimulants, the corresponding figures were 2.87, 1.66 and 1.13.
Because the study used prescriptions rather than confirmed pill-taking, people who did not take the medication or took breaks could still be counted as exposed. The records also missed events that did not reach medical care or were not coded, while including some minor events.
The findings describe patterns in a large population; they cannot provide an individual’s absolute risk after starting ADHD medication. They also do not show that medication lowered self-harm on its own. The researchers called for close follow-up around treatment initiation, including attention to changes in mood, behavior and self-harm thoughts. Decisions about starting, stopping or changing medication belong with a prescribing clinician who knows the person’s history.
AI tools were used to assist with the preparation of this article under our editorial standards.
Leave a Reply