Vaccine Efficacy: Understanding Clinical Trial Design

simulator intermediate ~10 min
Loading simulation...
VE ≈ 70% (95% CI: 62-77%)

A trial with 15,000 participants per arm, 5% placebo attack rate, and true 70% efficacy would yield an observed VE of approximately 70% with a 95% confidence interval of roughly 62-77%, providing strong statistical evidence of protection.

Formula

VE = 1 - (ARV / ARP) = 1 - RR (relative risk reduction)
SE(log(RR)) = sqrt(1/a - 1/n₁ + 1/c - 1/n₀) (standard error of log relative risk)
CI = 1 - RR × exp(±z × SE(log(RR))) (confidence interval for VE)

Measuring the Shield

Vaccine efficacy quantifies how well a vaccine protects against disease in a controlled clinical trial. The gold standard is the randomized, placebo-controlled, double-blind trial: participants are randomly assigned to receive the vaccine or a placebo, and neither they nor the investigators know which they received. By comparing disease rates between the two groups, researchers can isolate the vaccine's protective effect from confounding factors.

The Arithmetic of Protection

Efficacy is calculated by a deceptively simple formula: VE = 1 - (ARV/ARP), where ARV and ARP are the attack rates in the vaccinated and placebo groups respectively. If 50 out of 15,000 vaccinated participants get sick compared to 150 out of 15,000 placebo recipients, VE = 1 - (50/150) = 66.7%. But this single number hides important nuances: efficacy against infection may differ from efficacy against severe disease or death, and it may vary by age, sex, and time since vaccination.

Sample Size and Statistical Power

Designing a vaccine trial requires careful calculation of the sample size needed to detect a meaningful efficacy signal. The key drivers are the expected true efficacy, the disease incidence in the placebo group, the desired confidence level, and the minimum detectable efficacy. Lower disease incidence means more participants are needed because fewer total events occur. The COVID-19 trials of 2020 enrolled 30,000-44,000 participants each, powered to detect efficacy above 30% with 95% confidence.

From Trial to Real World

Clinical trial efficacy represents vaccine performance under ideal conditions. Real-world effectiveness is typically lower due to factors like imperfect cold chain storage, varying immune responses across diverse populations, waning immunity over time, and viral evolution. Post-licensure surveillance studies track effectiveness continuously, informing decisions about booster doses, updated formulations, and vaccination schedules. The gap between efficacy and effectiveness varies by vaccine but is a critical consideration for public health planning.

FAQ

How is vaccine efficacy calculated?

Vaccine efficacy (VE) is calculated as VE = 1 - (ARV/ARP), where ARV is the attack rate in the vaccinated group and ARP is the attack rate in the placebo group. A VE of 90% means the vaccine reduces the risk of disease by 90% compared to placebo. This is measured in randomized controlled trials where participants are randomly assigned to receive vaccine or placebo.

What is the difference between vaccine efficacy and effectiveness?

Efficacy is measured in controlled clinical trials under ideal conditions with carefully selected participants. Effectiveness is measured in real-world settings with diverse populations, imperfect storage, and varying adherence. Effectiveness is typically somewhat lower than efficacy due to these real-world factors, but both measure the proportional reduction in disease risk.

Why do vaccine trials need so many participants?

Large sample sizes are needed because you are measuring the difference in rare events between two groups. If the disease attack rate is 5%, you need thousands of participants to observe enough cases to distinguish a true vaccine effect from random variation. Lower disease incidence or lower true efficacy requires even larger trials.

What does a confidence interval mean for vaccine efficacy?

A 95% confidence interval means that if you repeated the trial many times, 95% of the calculated intervals would contain the true efficacy. A narrow CI (e.g., 88-94%) indicates precise estimation, while a wide CI (e.g., 40-90%) indicates substantial uncertainty. Regulatory agencies typically require the lower bound of the 95% CI to exceed a minimum threshold (often 30%).

Sources

Embed

<iframe src="https://homo-deus.com/lab/epidemiology/vaccine-efficacy/embed" width="100%" height="400" frameborder="0"></iframe>
View source on GitHub