- Design
- Pragmatic single-blind sequential multiple assignment randomised trial, 24 weeks
- Population
- 316 youths aged 8–17 with DSM-5 anxiety disorders
- Primary outcome
- Youth-reported SCARED score
- Effect
- Start CBT vs fluoxetine: 1.45 (−2.25 to 5.16); add vs continue in non-remitters: −2.74 (−6.53 to 1.05)
This pragmatic sequential multiple assignment randomised trial, published in June, recruited 316 children and adolescents aged 8–17 with a DSM-5 anxiety disorder from primary care and mental health clinics. Most had severe anxiety, co-occurring diagnoses and social disadvantage. They were randomised to fluoxetine or exposure-based cognitive behavioural therapy (CBT) for 12 weeks; those not in remission were randomised again to continue or to add the other treatment.
Self-reported anxiety scores fell by about 32% over 24 weeks. The starting treatment made no significant difference (difference 1.45 points, 95% CI −2.25 to 5.16, numerically favouring CBT). Among children not in remission at 12 weeks, adding the second treatment did not beat continuing the first (−2.74, −6.53 to 1.05). CBT followed by combination separated from other sequences on some secondary measures, and there were differences by ethnicity that the trial was not designed to test.
In practice, the choice between medicine and therapy can follow availability and family preference. Where waiting lists for exposure-based CBT are long, starting fluoxetine is a reasonable option rather than a compromise.
- Offer families the choice of fluoxetine or exposure-based CBT as equally reasonable starting points
- Do not delay treatment for a CBT waiting list if the family prefers to start medicine
- Review response formally at 12 weeks with a rating scale
- Monitor for activation and suicidal thoughts in the early weeks of an SSRI
Why it matters
Waiting for therapy is not clinically better than starting medicine for most anxious children.
Don't overread it
The differences by ethnicity came from subgroup analyses and do not justify choosing treatment by ethnicity.
The statistics, in plain English
Both confidence intervals cross zero, so neither the starting treatment nor adding a second one showed a clear advantage. The ethnicity findings are subgroup results and should be treated as hypotheses, not guidance.
Read the rest in the app
You have read your two free briefings this month. The app carries all 27 specialties, every morning, free — and this finding is waiting in it.
Scan to keep reading on your phone. No account needed to start.
Tomorrow morning, before your first patient
One edition a day for psychiatry, written by the desk, every claim tied to its paper. Six minutes.
Get the app — free