For the millions of people managing ADHD—and the clinicians prescribing for them—knowing not just which medication works, but at what dose, is the difference between therapeutic benefit and inadequate treatment. Growing concerns about subtherapeutic prescribing have gone largely unaddressed by clinical guidelines, which rarely specify dose-effect relationships with precision. This landmark analysis changes that calculus substantially.
Published in The Lancet Psychiatry, this systematic review and dose-effect network meta-analysis synthesized data from double-blind randomized controlled trials drawn from the MED-ADHD database, encompassing both published and unpublished aggregated data, with no language restrictions. The analysis included participants aged five and older meeting standardized ADHD diagnostic criteria and evaluated oral monotherapy with stimulants and non-stimulants. Using hierarchical Bayesian modeling with restricted cubic splines, researchers constructed dose-effect curves for both efficacy—measured through validated clinical rating scales—and tolerability, defined as discontinuation due to adverse events. Critically, analyses were stratified by age group, separating children and adolescents under 18 from adults, acknowledging that pharmacodynamic responses differ meaningfully across development.
This work sits at the top of the evidence hierarchy for psychiatric pharmacology. Network meta-analyses that model dose-response relationships—rather than simply comparing medications head-to-head—are rare and methodologically demanding. The Bayesian spline approach is well-suited to capturing the nonlinear dose-efficacy curves typical of CNS medications, where benefits plateau and tolerability worsens at higher doses. A key limitation is reliance on aggregated rather than individual patient data, which constrains the ability to examine moderators like body weight, comorbidities, or symptom subtype. The exclusion of withdrawal-phase designs and treatment-resistant populations also limits generalizability to real-world clinical complexity. Nonetheless, this analysis represents an incremental but genuinely meaningful advance—offering the most comprehensive dose-optimization map for ADHD pharmacotherapy yet assembled, with direct implications for guideline revision.