Home › Explainers › Trial phases
What the phases of a clinical trial mean
"Phase 3" sounds like a grade, and many people read it as one. It is better understood as a position in a sequence of questions. Here is what each phase is trying to find out and how to use that when you read about a study.

Phases are questions in order
New medical approaches are tested step by step because each step answers something the previous one could not. First comes safety in a small group. Then comes a first look at whether the approach does anything. Then a large comparison against what is already available. Then watching what happens once it reaches everyday use.
The label tells you which question the study is asking. It does not tell you the quality of the study, and it does not say how likely the approach is to work.
The four phases side by side
| Phase | Who takes part | What the researchers want to learn |
|---|---|---|
| Phase 1 | A small group, often healthy volunteers or people with few other options | How the body handles it, which dose is tolerated and what side effects appear first |
| Phase 2 | Several dozen to a few hundred people who have the condition | Early signs of benefit and a closer look at side effects |
| Phase 3 | Hundreds or thousands, often at many sites | Whether it works as well as, or better than, the current standard, and how safe it is across a wide group |
| Phase 4 | Large groups after the product is approved | Rare problems, long-term results and use in real-world care |
What the phases leave out
Some research does not follow phases at all. Studies of surgical methods, scans, diets, exercise programs and counseling are common examples. Studies that watch people without assigning anything are not phased either. When a listing says "not applicable" for phase, that is normal.
Also, early does not mean dangerous and late does not mean safe. A Phase 1 study of a gentle supplement may carry little risk, while a Phase 3 study of a strong drug may carry several. Read the risk section of the consent form whatever the number says.
Words that sit next to the phase
- Randomized
- People are assigned to groups by chance.
- Double-blind
- Neither participants nor the staff assessing them know who receives which treatment.
- Open-label
- Everyone knows what treatment is being given.
- Dose-escalation
- Small groups receive increasing doses so researchers can find a safe range.
How long each phase tends to take
Timelines vary a great deal, but a rough picture helps. Early safety studies often run for months. Middle-stage studies may take a year or two. Large comparison studies can take several years, partly because they need many people and partly because some outcomes take time to appear. A medicine can spend a decade or more moving through all of them.
For you, the length of the study you are offered matters more than the length of the whole program. Ask how long you would be involved, including any follow-up calls after the last visit.
Worked example
Imagine you read: "A Phase 2, randomized, open-label study of a new cream for chronic hand eczema, 150 adults, 16 weeks." Translation: the cream has passed its first safety check, the researchers now want an early read on whether it helps, people are placed in groups by chance, nobody is kept in the dark about which cream they get, and the study runs just under four months. The big open question is "does it seem to help?", not "is it proven?".
A short list to use when you meet a phase label
- Say the phase out loud as a question, using the table.
- Check whether the study uses a comparison group.
- Look for the number of people and how long the study lasts.
- Open the consent form to the risks section.
- Write down anything that still puzzles you.
Questions worth asking the study team
- What happened in the earlier phases, and what did the results show?
- Why is this study needed now?
- Will I be told how the study turned out?
Questions people often raise
- "Is Phase 3 safer than Phase 1?"
- Not in a simple way. Later phases have more safety data behind them, but the study itself can still involve a strong treatment with real side effects. Safety is judged by the specific risks listed, not by the phase number.
- "Does a study stop if the early results look bad?"
- It can. Safety reviewers watch the information as it comes in, and studies can be paused, changed or ended early when something worrying appears or when the answer is already clear.
- "What if no phase is listed?"
- Many studies of behavior, devices, diet and surgery do not use phases. The listing may say not applicable. Ask the team what stage the idea is at, in their own words.
The phase label is a good first question. The consent form and the study team supply the rest.
Where to go next
Next, see how informed consent works, or learn about the placebo.
Keep an explainer on your phone
Every explainer is open to read here. If you like, the get-started page has two separate optional boxes: one to be texted the explainer you pick, one for notices of new explainers. Up to 8 messages per month; reply STOP to opt out.