
Adderall is commonly prescribed medication for ADHD. (© Alex DiStasi - stock.adobe.com)
Study Exposes Critical Gap in Long-Term Research
In A Nutshell
- Methylphenidate (Ritalin) leads for all ages. The medication showed consistent benefits in both children and adults across multiple studies. However, tolerability remains a concern, particularly in adults.
- What works for kids doesn’t always work for adults. Children responded well to five different medications (methylphenidate, amphetamines, atomoxetine, alpha-2 agonists, and viloxazine), while adults showed strong evidence for only methylphenidate, atomoxetine, and cognitive behavioral therapy.
- Long-term data is nearly nonexistent. No intervention showed reliable evidence of sustained benefits beyond 26 weeks, despite millions of patients taking ADHD medications for years. Treatment decisions currently rely almost entirely on short-term studies lasting around 12 weeks.
- Non-drug options show promise but lack solid evidence. Acupuncture, cognitive behavioral therapy in children, and mindfulness in adults demonstrated large effect sizes, but methodological problems in the underlying research prevent confident recommendations.
One of the largest reviews of ADHD treatments to date has analyzed dozens of medications and therapies, revealing that what works best for children often differs sharply from what helps adults.
Researchers synthesized data from 47 systematic reviews and re-analyzed 221 different combinations of treatments and outcomes from past trials, creating a roadmap that clinicians and patients can use to navigate the array of ADHD treatment options. The team then translated their findings into a free online platform designed to help the roughly 5% of children with ADHD, many of whom continue to have symptoms as adults, make informed decisions.
For children and adolescents, methylphenidate emerged as the most consistently effective medication. The drug showed benefits regardless of who evaluated the child, whether clinicians, parents, or teachers, with effect sizes exceeding 0.75 and moderate to high quality evidence backing those findings. To put that in perspective, effects above 0.5 are considered medium and those above 0.8 are large, meaning methylphenidate’s performance fell solidly in the clinically meaningful range.
Four other medications also demonstrated meaningful short-term symptom reductions in younger patients: amphetamines, atomoxetine, alpha-2 agonists, and viloxazine. All showed medium to large effect sizes in reducing ADHD symptoms in the short term with moderate to high certainty evidence, though the certainty varied by who rated the symptoms.
Adult ADHD Treatments Show Different Patterns
In adults, only three interventions achieved moderate to high certainty evidence of efficacy. Methylphenidate (Ritalin) and atomoxetine (Strattera) both showed medium effect sizes for symptom reduction. Cognitive behavioral therapy, a non-drug approach, also demonstrated moderate certainty evidence of benefit when clinicians rated symptoms.
The research, published in BMJ, identified amphetamines as effective in adults, but only when the analysis was restricted to high-quality trials. This distinction matters because it means that weaker studies may have inflated or deflated the drug’s apparent benefits.
A notable pattern emerged across all age groups: the medications that worked best often proved hardest to tolerate. In adults, methylphenidate, amphetamines, and atomoxetine all had high certainty evidence that more patients stopped taking them due to side effects compared with placebo. For children and adolescents, amphetamines showed significantly worse tolerability than placebo with moderate certainty evidence.
Long-Term Evidence Remains Scarce
The research team, led by Corentin J. Gosling from Université Paris Nanterre and Samuele Cortese from the University of Southampton, uncovered a troubling gap in the research landscape. At medium and long-term follow-up points, typically around 26 and 52 weeks, no intervention showed high or moderate certainty evidence of sustained benefit, regardless of whether patients were children or adults.
This absence of reliable long-term data creates a problem for the millions of patients who take ADHD medications for years. Current treatment decisions rest almost entirely on short-term studies, usually lasting around 12 weeks.
Some non-drug interventions showed large effect sizes but couldn’t be recommended with confidence due to methodological problems in the underlying research. Acupuncture and cognitive behavioral therapy in children and adolescents, along with mindfulness in adults, all demonstrated apparently substantial benefits, but the evidence quality was too low to draw firm conclusions.
Medications Affect More Than Just Symptoms
The review looked at more than just core ADHD symptoms. In children and adolescents, amphetamines showed medium improvements in academic performance with moderate certainty evidence. Atomoxetine demonstrated small to medium improvements in quality of life, also with moderate certainty evidence.
For adults, atomoxetine showed small improvements in emotional dysregulation with high certainty evidence. Methylphenidate produced small improvements in executive functions with moderate certainty evidence. Both findings matter because ADHD affects more than just attention and hyperactivity.
One safety concern that many parents worry about found reassuring news in the data: methylphenidate was not significantly different from placebo for suicidal ideation or behavior in children and adolescents, based on moderate certainty evidence.
New Online Platform Makes Data Accessible
Researchers built an interactive website called EBI-ADHD where patients, families, and clinicians can explore the data themselves. Users can filter by age group, see effect sizes with confidence intervals, and read plain-language explanations of what the numbers mean.
The platform provides a new avenue for how medical evidence gets communicated. Rather than forcing patients to decode journal articles or rely solely on their doctor’s interpretation, the tool puts comparative data directly in their hands.
Study authors explain that they focused on placebo-controlled evidence rather than head-to-head drug comparisons because network meta-analyses that tried to directly compare medications often showed inconsistencies between direct and indirect evidence, undermining their reliability.
The team plans to update the platform annually through 2028, incorporating new evidence as it emerges. This “living evidence synthesis” approach means the recommendations won’t fossilize but will evolve as science advances.
Several caveats temper the findings. The results apply at the group level, which means individual patients may respond very differently than the average person in these studies. The researchers couldn’t break down results by sex or by age at diagnosis, both of which likely influence treatment response.
The review also found limited evidence on combination treatments, even though many patients take both medication and receive therapy. That gap reflects the design of most clinical trials, which typically test single interventions rather than the multimodal approaches common in real-world practice.
For parents weighing whether to start their child on ADHD medication, or adults considering treatment for the first time, effective options exist. Methylphenidate stands out for both age groups, but the medications come with trade-offs around tolerability, and the long-term effects remain inadequately studied. The creation of a continuously updated, publicly accessible database marks a step toward transparency in a field where treatment decisions have often relied on incomplete information.
Disclaimer: This article summarizes findings from a peer-reviewed umbrella review published in BMJ. The information presented is for educational purposes only and should not be used as a substitute for professional medical advice, diagnosis, or treatment. ADHD treatment decisions should be made in consultation with a qualified healthcare provider who can assess individual circumstances, medical history, and personal response to treatment. The effectiveness and tolerability of medications vary significantly between individuals, and what works at the population level may not reflect individual outcomes.
Paper Notes
Study Limitations
The study’s findings apply only at group level, potentially masking important individual differences in treatment response or tolerability. Researchers were unable to stratify results by participant sex or age at ADHD diagnosis, which would have been informative for clinical practice. Methodological inconsistencies between included meta-analyses may have persisted despite efforts to identify potential errors. The review identified limited meta-analytical results for specific multimodal interventions compared with neutral control, though combination therapy is common in some countries’ clinical practice. The review relied on existing eligible meta-analyses and could not capture individual randomized controlled trials on particular interventions or outcomes not yet included in systematic reviews and meta-analyses.
Funding and Disclosures
The work was carried out within the framework of ANR-23-CE28-0019-01 EBIA-CT awarded to CJG and supported by the French National Research Agency and France 2030 programme under reference ANR-23-IAIIU-0010 awarded to RD. SCo, NIHR Research Professor (NIHR303122), was funded by the NIHR for this research project. Multiple authors declared receiving honorariums, consultancy fees, research funds, or other financial relationships with pharmaceutical companies including Angelini, Lundbeck, Janssen, Otsuka, Takeda, and others, all outside the submitted work.
Publication Details
The study was authored by Corentin J. Gosling and colleagues from institutions including Université Paris Nanterre, Karolinska Institutet, University of Southampton, King’s College London, University of Oxford, and multiple other European and North American universities. Published in BMJ 2025;391:e085875 with DOI: 10.1136/bmj-2025-085875. The protocol was pre-registered on the Open Science Framework at https://osf.io/ugqy6/. Researchers screened 4632 references and assessed 414 full text articles, with 115 meta-analytical reports deemed eligible for inclusion, comprising 221 unique combinations of participants, interventions, comparators, and outcomes.







