Design: Systematic review and meta-analysis of RCTs
Population: 46 RCTs in children and young adults up to 25 with type 1 diabetes
Primary outcome: HbA1c
Effect: Overall g = 0.12; family-focused g = 0.29; youth-focused g = 0.05

This meta-analysis updated a decade-old review of randomised trials of behavioural interventions to improve type 1 diabetes self-management in people up to age 25. Of 62 trials, 46 had enough data to pool.

Overall the effect on HbA1c was small (Hedges' g 0.12). Pilot trials showed larger effects (g 0.36) than full trials (g 0.07), a sign that early enthusiasm shrinks at scale. Multicomponent interventions did better (g 0.19) than single-focus ones, and family-focused approaches did best (g 0.29), against 0.05 for youth-focused and 0.04 for parent-focused work.

Diabetes technology has changed a great deal in ten years, but family behaviour still matters. The benefit is modest; the message is where to put the effort — with the family together, not with the adolescent or parent alone.

Involve the whole family in diabetes self-management support, not just the child or one parent.
Combine practical skills (dosing, monitoring) with problem-solving and communication in the family.
Expect modest HbA1c gains from behavioural support; it complements, not replaces, technology.
Watch adolescence closely — the handover of responsibility is where control often slips.

Why it matters

It points limited psychological support time towards the family unit, where it helps most.

The statistics, in plain English

Hedges' g is a standardised effect size: 0.2 is usually called small and 0.5 moderate. An overall g of 0.12 is small; family-focused interventions at 0.29 are small-to-moderate. The gap between pilot and full trials suggests smaller studies overestimate benefit.