The short answer: to read a clinical trial result, ignore the framing and check three things in order. Did the primary endpoint hit? What was the effect size in absolute terms? And where is the confidence interval? If a press release leads with a secondary endpoint or omits the confidence interval, that omission is usually the real story. You do not need a statistics degree to read a readout well. You need to know about a dozen things and apply them consistently.
Here is what those things are.
Why every trial press release is spun
A company reporting clinical data has lawyers, investor relations advisers, and a share price. It will present the result in the most favorable light the facts permit. The facts constrain what can be said, but not what gets emphasized, what gets led with, or what gets quietly left for the conference presentation four months later. So the skill is not reading the release. It is reading around it.
The concepts that decide whether a trial worked
- The primary endpoint. The one measure the trial was designed to test, declared in advance. If it missed, the trial failed, no matter how good the rest sounds. A release that leads with a secondary endpoint is telling you the primary missed.
- The p value. The probability the result happened by chance. Below 0.05 is called statistically significant. It tells you an effect probably exists. It says nothing about whether the effect matters.
- Effect size. How much better the drug is than the comparator. This is the number a physician and a payer actually care about. Statistical significance says an effect is real. Effect size says whether anyone should care.
- The confidence interval. The range within which the true effect probably lies. A wide interval means an unreliable answer. If a hazard ratio’s interval crosses 1.0, the result is not significant, whatever the point estimate looks like.
- Overall survival vs progression free survival. In oncology, overall survival is the endpoint nobody argues with. Progression free survival is weaker, and a strong progression free result with immature survival data is a warning sign.
The phrases that mean the opposite of what they say
Press releases are carefully worded. That precision is itself informative, because a company only reaches for careful construction when the plain version would be worse.
- “A trend toward significance” means it missed. There is no such thing.
- “Numerically greater” means better, but not significantly better, which is to say not better.
- “Well tolerated” means no safety numbers are being given. Go find the discontinuation rate.
- “In a prespecified subgroup” is fine if the overall trial hit. If it missed, this is a hypothesis dressed as a result.
- “Data remain immature” means the survival curve is not ready, which sometimes means it is not good.
- “Strategic review of the program” means the program is being killed and the announcement is being staged.
Six questions to ask of any readout
- Did the primary endpoint hit?
- What is the effect size, in absolute terms?
- Where is the confidence interval?
- Was the comparison fair, randomized and blinded, or single arm against an assumption?
- Who left the trial, and why? Asymmetric dropout is a safety signal the safety table hides.
- What are they not telling me? The missing number is usually the interesting one.
Frequently asked questions
What does statistically significant mean in a clinical trial?
It means the result is unlikely to have happened by chance, conventionally a p value below 0.05. It indicates an effect probably exists, but it does not tell you whether the effect is large enough to matter clinically.
What is the difference between a primary and secondary endpoint?
The primary endpoint is the single outcome the trial was designed and powered to test, declared in advance. Secondary endpoints are additional measures. If the primary endpoint misses, the trial is generally considered to have failed, even if secondary endpoints succeed.
What does “well tolerated” mean in a drug press release?
It is a phrase with no regulatory definition and no numbers behind it. Rather than rely on it, look for the specific safety data: the rate of discontinuation due to adverse events, serious adverse events by arm, and any deaths.
Why do companies report progression free survival instead of overall survival?
Progression free survival is faster and cheaper to measure, so trials often use it. It is a weaker basis for a claim, because a drug can delay progression on a scan without helping patients live longer, which is why regulators increasingly want a survival signal too.
Get the full decoder
We built a complete reference with twelve concepts explained in plain terms, a phrasebook of what companies say and what they mean, the six questions, and what each possible outcome means commercially.
