Key facts
- Phase 2 role: find an early efficacy signal and select a dose, typically in tens to a few hundred patients.
- Phase 3 role: confirm efficacy and characterize safety with enough statistical power for regulatory approval, often in thousands of patients.
- Direction of change: effect sizes measured in phase 2 commonly get smaller in phase 3.
- Approval: most drugs need successful phase 3 confirmatory trials before a regulator approves them.
- Attrition: many drugs that look promising in phase 2 do not confirm their benefit in phase 3.
- Practical takeaway: an approved phase 3 medicine and an unapproved phase 2 candidate are not equivalent levels of evidence.
What a trial "phase" means
Clinical drug testing runs as a sequence of stages, and each stage answers a different question in a bigger study than the last. The phases are the map of how a candidate moves from "safe enough to give people" to "proven enough to approve" (PMID 32762387).
In broad strokes, phase 1 gives the drug to a small number of people — often healthy volunteers — mainly to check safety, tolerability and how the body handles it. Phase 2 then tests whether it appears to work in patients with the target condition, while continuing to watch safety. Phase 3 is the large confirmatory stage that regulators lean on to decide whether the benefits outweigh the risks (PMID 31536202). Understanding where a result sits on that ladder tells you how much weight it can bear.
What phase 2 tests
A phase 2 trial is where a drug first has to show it might actually do something useful. The core questions are "does this appear to work in the intended patients?" and "what dose looks best to carry forward?" To answer them, phase 2 typically enrols a modest number of participants — often tens to a few hundred — and may compare several doses against each other or against placebo over a relatively short window.
Because it is smaller and shorter, phase 2 is efficient but limited. It is well suited to detecting a signal and ruling out clearly ineffective candidates, but it is not powered to give a precise, reliable estimate of how large the benefit really is, nor to detect uncommon side effects. Its job is to justify the enormous cost of a phase 3 programme, not to settle the question. A strong phase 2 result is best read as "worth the bigger test," and this is the stage where many peptide candidates in weight loss and metabolic health currently sit.
What phase 3 tests
A phase 3 trial is the confirmatory stage. It takes the dose chosen in phase 2 and tests it in a much larger population — frequently thousands of participants — usually as a randomized, controlled, and often placebo-controlled study designed with enough statistical power to give a trustworthy answer. It also runs longer and on a broader, more representative group of patients, which is what makes its safety data far more informative than anything phase 2 can offer.
This is the evidence regulators such as the FDA and EMA rely on when they weigh benefit against risk. Approval normally requires one or more successful phase 3 trials, because only a study of this size and rigor can distinguish a real, durable clinical benefit from a promising early impression. The large cardiovascular and weight-loss outcome trials behind approved GLP-1 medicines — for example the SELECT trial of semaglutide — are phase 3 studies, which is exactly why their results are treated as settled rather than preliminary.
Why phase 2 results shrink in phase 3
One of the most consistent patterns in drug development is that the effect size looks smaller in phase 3 than it did in phase 2. This is not usually fraud or failure — it is statistics and study design working as expected. Several forces push in the same direction:
- Small samples are noisy. A phase 2 trial with a few hundred patients produces an estimate with wide uncertainty. By chance alone, some early results land high, and those are exactly the ones that get championed and advanced.
- Selection favors the flattering number. Because appealing phase 2 results are more likely to progress, the figures that reach the headlines are skewed toward the optimistic end — a form of regression to the mean once a bigger, fairer test is run.
- Phase 3 populations are broader. Larger trials enrol a wider mix of patients, including people who respond less well, so the average benefit moves toward a realistic middle. Our note on the responder rate explains why an average can hide very different individual outcomes.
- Endpoints get harder. Phase 3 often uses longer follow-up or more demanding clinical endpoints rather than the short-term surrogate measures a phase 2 study might rely on. The way a trial defines the treatment effect it is estimating — its estimand — can shift the number too.
The takeaway is not that phase 2 numbers are lies. It is that they are upper-bound impressions from a small sample, and the sober phase 3 figure is the one to plan around.
Approved vs unapproved: why it matters
This is where the distinction becomes practical. An approved medicine has cleared confirmatory phase 3 trials and a regulatory review of its complete efficacy and safety record. An unapproved phase 2 candidate has not — no matter how good its early numbers look.
Take retatrutide, an investigational triple-hormone-receptor agonist that has produced striking weight-loss figures in phase 2 studies. Those results are genuinely notable, and they are why a large phase 3 programme is underway. But until that programme reports and regulators complete their review, retatrutide's long-term benefit and safety are simply not established to the same standard as an approved drug like tirzepatide or semaglutide, both of which have cleared phase 3 outcome trials. A phase 2 headline and a phase 3 approval are different tiers of certainty, and treating them as interchangeable is one of the most common ways people overrate an experimental compound.
There is a separate point worth keeping distinct from the science: a drug still moving through trials is, by definition, not yet an approved product, which has its own consequences for how it can lawfully be obtained and marketed — see are peptides legal. The evidence question ("is it proven?") and the legal question ("is it approved?") tend to move together at this stage, but they are not the same question. For the broader map of which weight-loss compounds sit at which stage, see our pillar on peptides for weight loss.
Frequently asked questions
What is the difference between a phase 2 and a phase 3 clinical trial?
A phase 2 trial is a smaller study, typically in a few hundred patients, that asks whether a drug works at all and helps pick a dose to carry forward. A phase 3 trial is much larger, often thousands of patients, and is designed to confirm the effect and measure safety with enough statistical power to support regulatory approval. Phase 2 generates the hypothesis about efficacy; phase 3 is the confirmatory test of it.
Why do phase 2 results often shrink in phase 3?
Phase 2 trials are small, so their estimates are noisy and the ones that look most impressive are the most likely to advance, which flatters the starting number. Phase 3 then enrols a broader, more representative population, uses harder or longer endpoints, and is powered to detect a realistic effect rather than a best case. As a result, the effect size measured in phase 3 is often smaller than the eye-catching figure from phase 2, even when the drug genuinely works.
Does a positive phase 2 trial mean a drug works?
Not conclusively. A positive phase 2 result is an encouraging early signal that justified moving to a larger study, but many drugs that look promising in phase 2 fail to confirm their benefit, or show a smaller benefit, in phase 3. A single positive phase 2 trial should be read as a hypothesis worth testing, not as settled proof that the drug is effective and safe for general use.
Why is an unapproved phase 2 drug less certain than an approved one?
An approved drug has cleared confirmatory phase 3 trials and a regulatory review of its full efficacy and safety record; an unapproved phase 2 drug has not. For example, retatrutide has produced striking weight-loss numbers in phase 2, but until its phase 3 programme reports and regulators review it, its long-term benefit and safety are not established to the same standard as an approved medicine like tirzepatide or semaglutide. The phase 2 number is a promising signal, not a guarantee.
Sources
This glossary explainer draws on general references describing how clinical drug development and trial phases are structured:
- Drug Development 101: A Primer — International Journal of Toxicology, 2020; PubMed 32762387.
- Drug Trials — StatPearls; PubMed 31536202.