The phases of clinical trials are the stages through which a new medicine or intervention is tested in people. Phase 1 looks mainly at safety and dose, Phase 2 at early evidence of benefit, Phase 3 at confirmed benefit versus a comparator, and Phase 4 at what happens after approval.
Free sandbox · No credit card · 21 CFR Part 11 aligned
Clinical trial phases
noun · clinical research
The staged progression of human studies for a new treatment: safety and dose (Phase 1), preliminary efficacy (Phase 2), confirmatory efficacy and safety (Phase 3) and post-marketing follow-up (Phase 4).
Related terms
In short
Explained
Before any human study, a candidate treatment goes through laboratory and animal (preclinical) work. Clinical development then proceeds in phases, each answering a narrower question than the last and each justified by the one before. In the United States the regulation for drugs under an investigational new drug application (21 CFR 312.21) describes three phases for a previously untested drug, notes that they are generally conducted sequentially but may overlap, and defines Phase 1 as the first introduction of the drug into humans. Check the current text for exact wording. Phase 4 is the common name for studies after approval.
The participant numbers quoted for each phase are rules of thumb, not regulatory thresholds. A first-in-human study might enrol a few dozen people; a Phase 2 study tens to a few hundred patients; a Phase 3 study several hundred to several thousand, depending on the condition, the effect size being detected and the safety database regulators expect. A sample size calculation based on the primary endpoint, not the phase label, decides the real number.
Phase 1 studies give the treatment to people for the first time. Many use healthy volunteers; for toxic treatments such as cytotoxic cancer drugs, patients are enrolled instead. Typical designs are single ascending dose (SAD) and multiple ascending dose (MAD) cohorts, where each cohort receives a higher dose only after a safety review of the one before. Pharmacokinetics (what the body does to the drug) and early pharmacodynamics are measured with dense sampling, so the schedule of assessments is intense but the participant count is small.
Data needs are specific: dose-level and cohort tracking, safety review decisions recorded against each escalation step, and precise timing of doses and samples. See EDC for Phase 1 clinical trials and EDC for first-in-human studies.
Phase 2 enrols patients who have the target condition. Phase 2a often tests proof of concept (is there any signal?) and Phase 2b refines dose and regimen (which dose is worth taking forward?). Designs are frequently randomized and may be blinded, with a placebo or active comparator, though single-arm designs are common in oncology. The primary question is whether the effect on a chosen endpoint is large enough to justify a pivotal study.
Here data quality starts to carry regulatory weight. See EDC for Phase 2 clinical trials and proof-of-concept studies.
Phase 3 trials are the large, usually randomized and controlled studies designed to confirm efficacy and characterise safety in a population that resembles real use. Results from these studies are typically the core of a marketing application. They run across many sites and often several countries, so site management, query resolution, monitoring and consistent data capture matter as much as the science. Source data verification and risk-based monitoring plans are normally in place, and the dataset is frozen at database lock before unblinding and analysis. See EDC for Phase 3 clinical trials.
Phase 4 studies, also called post-marketing studies, look at safety and effectiveness in wider use, in subgroups not well represented earlier, or over a longer period. Some are required by a regulator as a condition of approval; others are sponsor-initiated. Registries and pragmatic designs are common here, see pragmatic clinical trials and EDC for Phase 4 post-marketing studies.
Years
Lab and animal work
Safety, dose
Signal, dose-finding
Confirmatory
Regulatory review
Side by side
Typical characteristics, not rules. Always follow the protocol and the regulator for your product.
| Phase 1 | Phase 2 | Phase 3 | Phase 4 | |
|---|---|---|---|---|
| Main question | Safe? How is it handled? | Any signal? Which dose? | Better than comparator? | Long-term and wider use |
| Population | Healthy volunteers or patients | Patients with the condition | Broad patient population | Real-world patients |
| Typical size | Tens | Tens to a few hundred | Hundreds to thousands | Varies, can be very large |
| Design | Dose escalation, often open label | Randomized or single arm | Randomized, controlled, often blinded | Observational, registry or pragmatic |
| Data emphasis | Dose cohorts, PK timing, safety review | Efficacy endpoints, dose groups | Multi-site quality, queries, monitoring, SDTM | Long-term follow-up, safety signals |
Hybrid labels such as Phase 1/2 and Phase 2/3 combine the questions of adjacent phases in one protocol.
Try the free Capture sandbox: build the forms, randomization and visit schedule for your phase, enter sample data and test the checks before you enrol anyone.
In practice
It is tempting to choose software by phase: a light tool for Phase 1, a heavy enterprise suite for Phase 3. In practice what matters is the protocol: how many sites, whether there is randomization or blinding, which endpoints are measured and how often, and what the regulator expects at submission. A small, tightly run Phase 3 can need less than a sprawling Phase 2, and a Phase 1 can be complex because of dose-escalation logic and time-critical sampling.
The same core controls apply at every phase: validated, role-based access, a field-level audit trail, edit checks that raise queries at entry, and exports that map to submission standards such as CDISC. The difference between phases is mostly volume and the number of parties involved, including a CRO if the sponsor delegates operational work.
In Capture
Capture is a self-serve platform for Phase 1, 2 and 3 studies. You build eCRFs with skip logic and edit checks, set up the visit schedule, add randomization where the design needs it, collect ePRO and eConsent in the same system and keep one field-level audit trail. Test everything in the free sandbox, and pay only when you go live with real participants.
| Assessment | Screening | Day 1 | Week 4 | Week 12 | Follow-up |
|---|---|---|---|---|---|
| Informed consent | |||||
| Vital signs | |||||
| Lab panel | |||||
| Symptom diary | D | D | D |
x = done at visit, D = ePRO diary
Phase 1 tests safety, tolerability and dose, often in a small group. Phase 2 looks for early evidence of benefit and the right dose in patients. Phase 3 confirms benefit and safety against a comparator in a larger population. Phase 4 studies the treatment after approval.
No. The phase system is mainly used for drugs and biologics. Many studies are labelled Phase 1/2 or Phase 2/3, some products skip or combine stages, and device, behavioural and nutrition studies follow different pathways.
It varies. As a rough guide Phase 1 is tens of participants, Phase 2 tens to a few hundred, and Phase 3 hundreds to thousands. These are conventions, not regulatory thresholds; the sample size comes from the statistical plan for the primary endpoint.
Phase 2 looks for a signal of benefit and chooses the dose, usually in a smaller group. Phase 3 is the large confirmatory study that tests the chosen dose against a comparator and builds the safety database.
Phase 0 usually means a very small exploratory study with sub-therapeutic doses. Phase 1/2 and Phase 2/3 are combined designs that answer questions from two adjacent phases in one protocol.
Yes. Capture is suitable for Phase 1, 2 and 3 studies. You can build the study, test it in the free sandbox with every feature and pay only when you go live with real participants.
Keep exploring
EDC for Phase 1 trials
Dose escalation and PK timing.
EDC for Phase 2 trials
Signal and dose-finding.
EDC for Phase 3 trials
Confirmatory multi-site studies.
EDC for Phase 4 studies
Post-marketing follow-up.
Clinical trial endpoints
What each phase measures.
Sample size calculation
How many participants.
Blinding
Single, double and triple blind.
Free sandbox with every feature. No credit card, no time limit.