Nicodemo · Journal of population economics 2026 · quasi-experimental cohort study (regression discontinuity design) · n=?

Why are the youngest in class more often prescribed pharmacological treatment for ADHD?

Cited 0 times in the scientific literature.

Level 2 - randomized trial

Quasi-experimental regression discontinuity design using administrative data (graded by design analogy, not clinical CEBM)

PubMed 42495092 · doi:10.1007/s00148-026-01190-y · record verified 2026-08-29

What was done

The authors analyzed administrative data from England tracking children from ages 5 to 15 to assess how school starting age affects ADHD pharmacological treatment rates. Exploiting the September 1 school-entry cutoff in a regression discontinuity design, they compared children born just before and after the threshold. A decomposition strategy was used to evaluate the separate impacts of relative age, absolute age, length of school exposure, and age at school entry.

What was found

Early school starters were 40% to 50% more likely to receive ADHD medication by age 15. This long-term gap was entirely driven by first-time prescriptions occurring between ages 5 and 8. Relative age within the peer group was identified as the primary long-term driver of the prescribing gap, whereas absolute age and length of school exposure exerted only short-lived effects.

Why it matters

This study provides evidence that peer-comparison bias and relative developmental immaturity within a classroom drive elevated ADHD prescription rates for the youngest pupils, rather than absolute biological age or cumulative school exposure. It points to potential overdiagnosis of younger children and supports updating diagnostic guidelines and classroom age grouping practices.

Limits

The abstract does not state the sample size (n) or exact numerical confidence intervals. Findings from administrative prescription records in England may not generalize to educational or clinical systems with different diagnostic practices. Direct symptom assessments and non-pharmacological interventions were not reported in the abstract.

Cited by