Reading a trial result

A trial 'works' only if it meets its pre-specified primary endpoint with statistical significance. The details matter as much as the headline.

Key terms

• Primary endpoint — the main pre-defined measure of success; missing it usually sinks the stock regardless of anything else.
• Secondary endpoints — supporting measures; nice to have, not sufficient on their own.
• Statistical significance (p-value) — p < 0.05 is the common bar; it means the result is unlikely to be chance.
• Effect size — how big the benefit is, not just whether it's significant; a 'significant' but tiny benefit can still disappoint.
• Safety / tolerability — adverse events can sink an otherwise positive trial.

Common traps

• Hitting a secondary while missing the primary — often spun positively, but usually a fail.
• 'Trends toward significance' — i.e. it missed.
• Subgroup wins — post-hoc subgroups are hypothesis-generating, not confirmatory.