Nadia Hollins’s son was referred for behaviour three times before anybody in the room asked what his nights were like.

The diary runs for two weeks before anything is changed. Retrospective estimates of sleep are unreliable in both directions, and the diary is the correction.

Sleep hygiene alone has a poor evidence base as a standalone treatment, which is awkward, because it is the most widely distributed sleep advice in existence.

Sleepy children don’t look sleepy. They look difficult. That single fact costs us years.

Illustrative composite

Adherence to therapy is reviewed at two weeks, six weeks and three months, because almost all abandonment happens inside the first month.

“He’s the same child,” she says. “He’s just had a full night’s sleep for eleven months.”