Vaccine safety testing is one of the most heavily documented processes in modern medicine, and yet it is widely misunderstood — partly because the documentation is technical and partly because the process is genuinely long. This article walks through what actually happens between a candidate vaccine in a lab and an approved product on a clinic shelf.
Nothing here is medical advice. The goal is to describe, in plain language, how the FDA, CDC, and independent advisory bodies evaluate vaccines, and where readers can find the underlying data for any specific product.
Preclinical research
Before a vaccine candidate is given to any person, researchers spend years studying it in cells and animals. The goal of preclinical work is to characterize how the candidate triggers an immune response, identify the dose range that produces a response without unacceptable toxicity, and rule out obvious safety problems. Only a small fraction of candidates that look promising in preclinical work make it past this stage.
Phase 1: small human trials
Phase 1 trials enroll a small group of healthy adult volunteers — usually 20 to 100 — and focus primarily on safety, dose-finding, and an early look at the immune response. Adverse events are tracked closely. If the candidate causes serious or unexpected problems, development stops or restarts at a different dose or formulation.
Phase 2: hundreds of participants
Phase 2 expands the trial to several hundred participants, often including subgroups the vaccine is intended to protect (different ages, conditions). The trial continues to track safety and immune response and refines the dosing schedule. Comparisons with placebo or an existing vaccine begin in earnest.
Phase 3: thousands to tens of thousands
Phase 3 is the large, randomized, controlled trial most people picture when they hear "clinical trial." It enrolls thousands and sometimes tens of thousands of participants, randomly assigned to the vaccine or a placebo (or comparator), and is the primary evidence the FDA uses to evaluate effectiveness and detect less-common adverse events. Pre-specified statistical thresholds determine whether the result is considered meaningful.
Phase 3 trials also include diverse populations to detect differences by age, sex, race, and underlying conditions. Trial results are submitted to the FDA as a Biologics License Application (BLA), accompanied by detailed manufacturing information.
Regulatory review and the advisory committee
The FDA's Center for Biologics Evaluation and Research (CBER) reviews every page of the application, including raw trial data. For most vaccines, the FDA also convenes the Vaccines and Related Biological Products Advisory Committee (VRBPAC) — an independent panel of outside scientists and physicians — to publicly debate the evidence. The CDC's Advisory Committee on Immunization Practices (ACIP) then makes recommendations on use, including which age groups and schedules.
Both VRBPAC and ACIP meetings are open to the public and the materials are posted in advance, which is the most accessible primary source for anyone who wants to read what the regulators actually saw.
What happens after a vaccine is approved
Post-licensure safety monitoring is continuous. The Vaccine Adverse Event Reporting System (VAERS), co-managed by the FDA and CDC, accepts reports from anyone — clinicians, patients, parents — about events that occur after vaccination. VAERS is a hypothesis-generating system, not a diagnosis system; a report does not establish causation. Signals from VAERS are investigated using more rigorous data sources, including the Vaccine Safety Datalink (VSD), which links electronic health records across millions of patients in several US health systems.
Manufacturers are also required to conduct Phase 4 (post-marketing) studies for many vaccines, and the FDA inspects manufacturing sites on an ongoing basis. Vaccines have been withdrawn or had recommendations changed when post-licensure data revealed previously unrecognized risks; this is the system working as designed, not a sign that the earlier review was inadequate.