What did the placebo arm of SELECT report for dizziness?

Concretely: SELECT · dizziness.

I have read the primary source rather than the summary, which has left me with more questions.

I understand the headline. I do not understand the footnotes, and the footnotes look important.

How should I read this, and where are the traps?

The underlying point is that this is answerable from the published record, but only if you take the placebo arm seriously rather than reading the active arm alone.

Trial populations are selected. Exclusion criteria in this class routinely remove people with significant renal impairment, prior pancreatitis and unstable psychiatric illness, which is exactly the population the results are then quoted for.

Open-label extensions are not the same evidence as the randomised phase. Once everyone knows what they are taking, the reported outcomes acquire a bias that no analysis fully removes.

Registry entries at ClinicalTrials.gov carry the pre-specified primary endpoint with a timestamp, which is the cheapest available check on whether an endpoint was changed after the data were seen.

I am not a clinician and this is not medical advice; it is a reading of a published protocol.

When two sources disagree, the answer is almost always in the methods section of the one you have not read.

The short version: the effect is real, the magnitude depends on the population, and the population is usually the part that gets dropped when a result is quoted second-hand.

Duration decides what can be seen. A 68-week trial can measure weight and glycaemia; it cannot measure anything whose event rate is one per cent per year without enrolling tens of thousands.

Put another way, non-inferiority and superiority designs are not interchangeable. A non-inferiority result says the new agent is not meaningfully worse against a pre-specified margin — it does not say it is as good, and it certainly does not say it is better.

If a claim cannot be traced to a named trial with a named endpoint, treat it as a claim rather than as evidence.

Look at the discontinuation rate alongside the efficacy figure. A large effect in the two thirds who stayed is a different result from a large effect in everyone.

Intention-to-treat and per-protocol analyses answer different questions. ITT asks what happens if you offer the treatment; per-protocol asks what happens if it is taken as directed. The gap between the two is a measure of how tolerable the protocol was.

Meta-analyses in this area are dominated by whichever trial contributed the most participants, so read the forest plot rather than the summary estimate.

The caveat is that trial evidence is about licensed product administered under supervision. None of it transfers automatically to research-grade material of unverified content.

Read the protocol and the statistical analysis plan if the result matters to you. Both are usually published alongside.

edited 17 Jul 2026 by h_villanueva — added a caveat about sampling

Before comparing two trials, check whether they share an endpoint definition. Frequently they do not, and the numbers then are not comparable in any sense.

Confidence intervals matter more than point estimates when two trials disagree. Two studies reporting fifteen and twenty per cent whose intervals overlap heavily have not disagreed about anything.

One qualification: absence of a signal in a trial of this size is not evidence of absence for a rare event. It is evidence that the event is rarer than the trial could detect.

Quote the interval alongside the estimate and half the disagreements on this site would not start.