A pattern I’ve been tracking across my patient panel lately: adults with ADHD who arrive carrying years of suboptimal stimulant doses — 18 mg of methylphenidate, unchanged since their pediatrician first prescribed it a decade ago — alongside a parallel cohort who’ve been pushed well past labeled ceilings by well-meaning clinicians chasing symptom control that never fully arrives. Both groups are frustrated. Both groups are, in different ways, the product of a field that hasn’t had rigorous dose-response data to anchor clinical decisions. I see this pattern in my clinic every week, and it tells me the field has been operating with inadequate maps.
That map just got substantially better.
What the Network Meta-Analysis Actually Shows
A systematic review and dose-effect network meta-analysis published in Lancet Psychiatry has now done something the field has needed for years: mapped the dose-response relationship across ADHD pharmacotherapy with enough granularity to challenge two opposing clinical mistakes simultaneously. The authors identify therapeutic inertia — accepting a suboptimal response without further dose titration — and uncritical dose escalation beyond licensed limits as twin hazards, each carrying its own risk profile. That framing matters. Most clinical discussions about ADHD dosing collapse into one of these failure modes without acknowledging the other.
The neuropharmacology here is worth holding onto. Stimulants operate on dopaminergic and noradrenergic systems with steep dose-response curves that are highly individual — the same 36 mg of Concerta that produces executive function normalization in one patient produces appetite suppression and cardiovascular strain in another with no additional cognitive benefit. FDA-approved ceiling doses for extended-release methylphenidate sit at 72 mg daily, and for amphetamine formulations the ranges vary by product and age — but those ceilings were established on efficacy and tolerability data, not on optimized dose-response modeling of the kind this meta-analysis attempts. A prior Lancet Psychiatry network meta-analysis across 113 randomized controlled trials in more than 14,800 adults confirmed stimulants and atomoxetine as the only treatments with proven efficacy — but it didn’t resolve the dosing question. This new analysis moves us closer to answering the question that practicing clinicians actually ask every appointment: how much, in whom, and when to stop titrating.
The cardiovascular signal is the part that deserves clinical attention without being dramatized. Long-term stimulant exposure has been associated with increased risk of hypertension and arterial disease in dose-dependent fashion, which means the uncritical escalation arm of this story carries real population-level consequences. That risk doesn’t make high-dose therapy wrong — it makes precise dose-response data essential, which is exactly what this analysis is trying to provide.
There is a specific driver of that therapeutic inertia I see more clearly than most, because much of my work is in addiction psychiatry: the fear of abuse. In my experience, an unsubstantiated worry about diversion or dependence is one of the most reliable reasons a clinician holds a dose below what a patient actually needs. The caution feels responsible, so it rarely gets named or questioned — and the patient absorbs the cost quietly, carrying a dose that keeps them technically treated and functionally undertreated, sometimes for years. The irony is that genuine misuse risk is precisely what I am trained to assess directly; it does not have to be managed through chronic underdosing. When abuse fear substitutes for that assessment, it stops being prudence and becomes its own failure mode — exactly the kind this meta-analysis can help correct by giving clinicians something firmer than apprehension to titrate against.
What Trials Keep Getting Wrong
Here is where the clinical pattern and the research landscape intersect in a way that should concern trial designers. A retrospective observational study across 1,229,972 individuals initiating ADHD medication between 2010 and 2020 documented high discontinuation rates within the first year of treatment — a signal that’s consistent with what I see in practice, where patients cycle off medications not because the drug class failed them, but because the dose was never optimized. Therapeutic inertia is a recruitment and retention problem in disguise. When a patient abandons treatment because an inadequate dose produced side effects without adequate benefit, they enter trials — if they enter them at all — with negative treatment histories that confound response modeling and inflate placebo arms.
The design implication is structural. Most ADHD pharmacotherapy trials still use fixed-dose or narrow flexible-dose designs, which means they enroll patients at doses that may fall systematically in the suboptimal range for a meaningful fraction of the population. The MTA Cooperative Group’s landmark trial, which enrolled 579 children aged 7 to 9.9 years, demonstrated that intensive medication management — with systematic titration and follow-up — outperformed routine community care, and that finding has held for decades. But the field’s trial architecture hasn’t fully absorbed the lesson: rigid dosing schedules don’t reflect how clinicians actually optimize these medications, and they produce efficacy signals that underestimate what the drug can do at an individually titrated dose. If this new meta-analysis is going to inform clinical guidelines, it should also push sponsors toward adaptive dose-finding designs that allow titration to individual response within a protocol-defined range, rather than treating dose as a fixed variable.
Shared Decisions Require Better Data
The Lancet Psychiatry authors explicitly frame their findings as support for shared decision-making around ADHD medication dosage. That framing is clinically correct and practically underused. Research by Zeitler and colleagues, published in Patient Education and Counseling, demonstrated that structured shared decision-making interventions significantly increased parental knowledge and involvement in ADHD treatment decisions — but shared decision-making requires that the clinician bring real dose-response data to the conversation, not just heuristics accumulated from individual prescribing experience. That’s precisely what this meta-analysis begins to supply. In my own practice, particularly with Spanish-speaking adult patients who’ve often had ADHD undertreated or misdiagnosed for years, the conversation about dosing is never just pharmacological — it’s about trust, about side-effect tolerance shaped by work demands and family structure, about what “good enough” means to a patient who has never experienced adequate treatment. That conversation requires data granular enough to say: here is where the benefit curve flattens, and here is where the risk curve starts climbing.
I’ll be watching for whether this analysis shapes the dosing language in the upcoming DSM and international ADHD guideline revisions — and more specifically, whether any sponsor uses it to justify a protocol design that finally treats dose as the dynamic, individually calibrated variable it actually is in every clinic running these patients.
References
- Lancet Psychiatry — “Pharmacological interventions for ADHD: a systematic review and dose–effect network meta-analysis”
- Oxford Health NHS — “Oxford and Southampton study identifies the most effective treatments for ADHD in adults” (Lancet Psychiatry network meta-analysis, 113 RCTs, 14,800+ adults)
- PubMed — “ADHD medication discontinuation and persistence across the lifespan: a retrospective observational study using population-based databases” (n=1,229,972)
- Johns Hopkins — “Findings from the NIMH Multimodal Treatment Study of ADHD (MTA)” — 579 children, ages 7–9.9
- ResearchGate — Zeitler et al., “Shared decision-making to improve attention-deficit hyperactivity disorder care,” Patient Education and Counseling, 2013
Dr. Leonardo Vando, MD is Principal Investigator of Psychiatry at Vando Medical Services.
