Melatonin versus placebo in children with autism spectrum conditions and severe sleep problems not amenable to behaviour management strategies: a randomised controlled crossover trial.

Wright et al. (2011) ran a crossover trial. Each autistic child took melatonin for three months and placebo for three months. The order flipped halfway through. No one knew which pill was which.

Kids had to have severe sleep problems that behavior plans had already failed to fix. Dose started low and could climb to 10 mg at bedtime.

Melatonin cut the time to fall asleep by about 47 minutes. Total nightly sleep stretched 52 minutes longer. Night wakings did not change.

Parents reported the gains, and doctors saw no serious side effects. The benefit washed out when kids switched back to placebo.

Tyrer et al. (2006) tried the same idea first in an open-label study. All 25 kids slept better, but there was no blind. Barry’s later RCT proves the effect is real, not placebo.

McLay et al. (2019) looks like a contradiction. They fixed severe sleep problems in autistic kids using only parent-run FBA plans. The difference: their families had not yet tried full behavioral sleep packages. Barry’s trial picked the toughest cases—behavior plans had already failed—so medication was the next step.

Van Hanegem et al. (2014) followed up with low-dose melatonin and found that kids who respond have normal melatonin metabolism. You do not need pricey lab work before prescribing.

If you have an autistic client who still lies awake for hours after Mom has done bedtime fading, faded co-sleeping, and ruled out medical issues, this study gives you green light to talk with the pediatrician about a three-month melatonin trial. Track sleep with a simple diary; you may gain nearly an extra hour of sleep for both child and parent with little risk.