How to Read a Clinical Trial Without a Medical Degree
A short guide to spotting the difference between a headline-worthy result and one that should not change your weekend plans.
Overhead flat-lay of fresh whole foods on a deep navy surface.
Every week brings a press release announcing that a new study has changed everything. Most have not. Reading trials with a small amount of structure separates real signal from journal-promoted noise.
Start with the endpoint
A clinical trial measures something. That something — the primary endpoint — is the only result the trial was designed to detect. Everything else is exploratory.
If the headline result comes from a secondary endpoint, treat it as a hypothesis for the next trial, not a finding.
Relative vs absolute
A 50% relative risk reduction sounds dramatic. If the baseline risk is 2%, the absolute reduction is one percentage point. Both numbers are honest. Only one tells you whether to act.
Always ask: 50% of what?
Sample size and duration
A 200-person trial run for 12 weeks can detect blood-pressure changes. It cannot detect cardiovascular events. Match the question to the design.
Funding is not disqualifying
Industry funds most clinical research. That is not in itself a problem. What matters is whether the protocol was registered before enrollment, whether an independent committee monitored conduct, and whether the academic authors retained publication rights regardless of outcome.
- Check ClinicalTrials.gov for the registered protocol
- Compare the published primary endpoint to the registered primary endpoint
- A switched endpoint is a yellow flag
The press release is not the paper
Read the paper. If you cannot access the paper, read the abstract and the limitations section. Skip the press release entirely. It exists to be quoted, not to inform.
