The paediatric anxiety trial found no difference between fluoxetine, CBT, their combination, or any sequencing of them. It is easy to read that as a disappointing result. It is closer to the opposite.
A null from an underpowered trial says we could not tell. A null from a well-powered pragmatic trial in a representative population says these options are similar — and that changes who should be making the choice. When several treatments produce the same outcome, the clinician's technical judgement stops being the deciding input, and what remains are the things the patient and family know better: what they will actually attend, what they can afford, whether a young person will take a daily tablet, whether a parent can bring them weekly.
The habit is to notice when a comparative trial has genuinely equalised the options, and then to stop making the decision alone. The mistake is to keep expressing a preference the evidence no longer supports — which quietly converts a legitimate patient choice into a clinician's default.
- Distinguish a well-powered null from an underpowered inconclusive one
- Equivalence moves the decision toward preference, adherence and access
- Ask what the family can actually sustain for the treatment duration
- A clinician default expressed as advice removes a real choice
- Check the trial population resembles the patient before applying this
The statistics, in plain English
Absence of evidence and evidence of absence are distinguished by the confidence interval, not the p value. Here the 24-week difference of 1.45 with an interval from -2.25 to 5.16 excludes any large advantage in either direction, which is what licenses the equivalence reading. A wide interval containing zero would have licensed nothing.
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