What medication holidays actually are
A "medication holiday" (also called a "drug holiday" or "structured treatment interruption") means intentionally stopping ADHD medication for a period, typically weekends, school breaks, or summer months. The practice is most common with stimulant medications like methylphenidate and amphetamine-based drugs.
Parents, and sometimes clinicians, suggest medication holidays for several reasons: to reduce side effects, to assess whether the medication is still needed, or to allow "catch-up" growth in children. But the evidence on whether this is a good idea is more nuanced than the advice suggests.
What the research shows
The 2021 World Federation of ADHD Consensus Statement (Faraone et al.) reviewed evidence on medication interruptions and found mixed results. Structured breaks can be reasonable in specific circumstances, but blanket recommendations for drug holidays aren't supported.
Key findings from the research:
- Symptom return is rapid. Stimulant medications don't build up in the body. When you stop, the effect stops. ADHD symptoms return fully, often within hours of the missed dose.
- Growth effects are real but modest. MTA follow-up data showed a 2 cm reduction at 3 years, with long-term consistent use associated with about 2.5 cm shorter adult height on average (Swanson et al., 2017). Discuss growth monitoring with your prescriber.
- Restarting can feel rough. Some people report that side effects like reduced appetite or trouble sleeping feel stronger for a day or two when they restart after a break. Tell your prescriber if this happens.
- Tolerance is not well-established for therapeutic doses. Long-term studies suggest most people keep benefiting at stable prescribed doses, and when medication seems to stop working, sleep, stress, or changing demands are often the cause. Breaks to "reset tolerance" may not be addressing a real phenomenon.
When a medication break makes sense
- To reassess need. Periodically checking whether symptoms persist without medication is reasonable, especially in children whose brains are developing. This should be done with your prescriber's guidance, not unilaterally.
- To manage specific side effects. Appetite suppression or sleep problems may improve during breaks, giving the body a chance to recover.
- During genuinely low-demand periods. If a weekend involves no driving, no responsibilities, and no interpersonal demands, a break might be fine. But this describes very few real weekends.
When breaks are risky
- Driving. ADHD increases crash risk. In a study of more than 2.3 million people with ADHD, Chang et al. (2017) found that men and women were 38% and 42% less likely to have a crash during months they were taking medication than during months they were not. If you drive on weekends, factor that into any break.
- Social and family time. ADHD symptoms don't just affect productivity. Emotional dysregulation, impulsivity, and inattention affect relationships too. Unmedicated weekends can strain family dynamics.
- Self-esteem in children. If a child notices they're "worse" on weekends, this can reinforce the idea that they need medication to be acceptable, which creates its own psychological harm.
How to decide
Talk to your prescriber. The decision should factor in symptom severity, side effect burden, lifestyle demands, and individual response. There's no universal right answer, but the decision should be informed by evidence rather than general anxiety about medication.
If tracking how you feel on and off medication helps you have a more productive conversation with your doctor, try UpOrbit. It's free, private, and can help you notice patterns over time.
References
- Faraone et al. (2021). World Federation of ADHD Consensus Statement. Neuroscience & Biobehavioral Reviews, 128, 789-818.
- Chang et al. (2017). Association between medication use for ADHD and risk of motor vehicle crashes. JAMA Psychiatry, 74(6), 597-603.
- Swanson et al. (2017). Young adult outcomes in the follow-up of the Multimodal Treatment Study of ADHD. J Child Psychol Psychiatry, 58(6), 663-678.