Vaccine efficacy and evaluating preventive measures
- Measure
Vaccine efficacy is the percentage by which a vaccine lowers the risk of disease in vaccinated people, compared with unvaccinated people.
What it is
Vaccine efficacy (VE) measures the proportional reduction in the risk of disease among vaccinated versus unvaccinated people: it is the relative risk reduction, VE = 1 − RR. Measured in a controlled trial it is called efficacy; measured in the population under real conditions, effectiveness. Outside trials it is estimated with cohorts, with case-control studies (VE = 1 − OR) or with the screening method, which needs only the coverage and the proportion of cases vaccinated.
Formula and reading
- VE = (ARU − ARV) / ARU = 1 − ARV / ARU = 1 − RR.
- Screening method: VE = 1 − [PCV / (1 − PCV)] × [(1 − PPV) / PPV]; PCV = proportion of cases vaccinated, PPV = proportion of the population vaccinated.
- VE = 0 % no protection; VE = 100 % complete protection; negative VE: more cases among the vaccinated (look for bias).
Worked example
- Trial: 10 000 vaccinated with 20 cases and 10 000 on placebo with 100 cases.
- ARV = 20 / 10 000 = 0.2 %; ARU = 100 / 10 000 = 1.0 %.
- VE = (1.0 − 0.2) / 1.0 = 0.80 = 80 %; the same as 1 − RR = 1 − 0.2 = 0.80.
- In the population (screening method): coverage PPV = 90 % and cases vaccinated PCV = 60 %.
- [0.60 / 0.40] × [0.10 / 0.90] = 1.5 × 0.111 = 0.167; VE = 1 − 0.167 = 0.833, rounded to 83 %.
- Reading: vaccinated people have 80 % less risk; 6 in 10 cases being vaccinated is no failure, since 9 in 10 people are vaccinated.
How vaccine efficacy or effectiveness is estimated
| Method | Data needed | Formula | Limitation |
|---|---|---|---|
| Randomised trial | Cases among vaccinated and placebo | 1 − RR | Costly; selected population |
| Cohort | Follow-up of vaccinated and unvaccinated | 1 − RR | Confounding (people who get vaccinated look after themselves more) |
| Case-control (or test-negative) | Vaccination in cases and controls | 1 − OR | Control selection |
| Screening method | Coverage and % of cases vaccinated | 1 − [PCV/(1 − PCV)] × [(1 − PPV)/PPV] | Rough estimate; needs accurate coverage |
Markers
- VE
- ARU
- ARV
- 1 − RR
- 1 − OR
How to read it
- Vaccination programmes monitor effectiveness in the population to detect waning protection or cold-chain failures.
- The 1 − RR formula serves to evaluate any preventive measure (bed nets, prophylaxis, campaigns).
- Indirect protection of unvaccinated people (herd immunity) is not captured by individual VE.
Common pitfalls
- Concluding a vaccine does not work because many cases are vaccinated, without accounting for coverage.
- Confusing efficacy (trial) with effectiveness (real practice).
- Reading VE as an absolute reduction: 80 % of a 1 % risk prevents 0.8 cases per 100 vaccinated.
Test yourself
Think of the answer, then tap the question to see it.
If the attack rate is 5 % in unvaccinated and 1 % in vaccinated people, what is vaccine efficacy?
VE = (5 − 1) / 5 = 0.80 = 80 %; also 1 − RR = 1 − 0.2.
In a case-control study, the odds ratio for vaccination is 0.30. What effectiveness does it estimate?
VE ≈ 1 − OR = 1 − 0.30 = 70 %.
Why, with very high coverage, may most cases be vaccinated?
Because almost the whole population is vaccinated: even though each vaccinated person has lower risk, there are so many that they contribute many cases.
What is the difference between vaccine efficacy and effectiveness?
Efficacy is measured in controlled trials; effectiveness in the population under real conditions of use.
Elsewhere in DurgiPedia
- Vaccine types: live, inactivated, subunit, toxoid, mRNA, vector
- Herd immunity and outbreak control
- Colombia's PAI schedule (which vaccines, at what age)
Connections
Educational content on epidemiological and statistical methods; the examples use made-up numbers to teach the calculation. Not a clinical or legal guide: current Colombian rules are in the official sources.
Sources: Principles of Epidemiology | Lesson 3 - Section 6 (Centers for Disease Control and Prevention, 2012); Chapter 6: Choosing effect measures and computing estimates of effect (Cochrane, 2024).
Spotted a mistake?We are always improving for the community. Tell us what is wrong and our team will review it.Report a mistake
Representative model, for educational purposes only.