What is compiled
For each vaccine on Vaccine Data Navigator, three kinds of estimate:
- Disease risks: the acute consequences of infection, such as complications, hospitalisation, intensive care, death and case-fatality, both before vaccines were introduced and today.
- Post-acute and long-term sequelae: what can follow the acute illness, including long COVID, subacute sclerosing panencephalitis, post-herpetic neuralgia, cancers after chronic HBV or HPV infection, congenital rubella and varicella syndromes, post-polio syndrome, hearing loss and neurodisability after meningitis, and myocardial infarction or stroke after respiratory infection.
- Efficacy and effectiveness: every estimate found, from randomised trials and from observational designs (test-negative, cohort, case-control, screening method), including waning and protection in special populations.
PICOS annotation
A single headline percentage hides who it applies to. Every row is annotated with:
- Population: age as reported plus a standard age band, sex, risk group, immune status, pregnancy, prior infection, vaccination status (for disease-risk rows), country, setting, period and sample size.
- Intervention: product, platform, number of doses, schedule and time since vaccination.
- Comparator: placebo, unvaccinated, test-negative controls, fewer doses, or (for post-acute rows) uninfected controls.
- Outcome: the case definition, severity level (infection, symptomatic, hospitalisation, death) and strain, serotype, variant or season.
- Study design, with GRADE certainty where the source reports it.
Verification
No number on this site was written from memory. Each row carries a verbatim quote from its source. For PubMed sources, the title and abstract are fetched from NCBI E-utilities and cached in the repository. For other sources (CDC Pink Book, MMWR, WHO position papers, product labels, full-text articles), a plain-text snapshot of the page is cached. An automated validator rejects any row where:
- the quote does not appear verbatim in the cached source;
- the point estimate or a confidence bound does not appear in the quote;
- a PubMed ID was not retrieved from PubMed.
The validator cannot judge whether the PICOS annotation faithfully reflects the source. That is checked by reading, so open "quote" on any row to check it yourself.
How to read the numbers
- Estimates are shown as the source reports them (VE %, RR, HR, CFR) and are not recomputed. Where a confidence interval is on a different scale from the point estimate (for example a VE with an RR interval), the scale is labelled.
- Disease-risk rows from the pre-vaccine era describe what the disease did before vaccination. Current rows describe today's mostly vaccinated populations. Check the period field.
- Observational effectiveness and post-acute hazard ratios are open to confounding, including healthy-vaccinee bias, differences in testing, and comparator choice. The notes on each row say so where relevant.
- Low and null estimates are included, not selected out.
Gaps (what was searched for and not found) and the full search log are listed at the bottom of each vaccine page.