Vaccination: What the Evidence Actually Shows — and What Doubt Costs

Vaccines are victims of their own success. Earlier generations saw children paralysed by polio and babies die of whooping cough. We don't — precisely because the vaccines worked. What we do see is the jab, the fever that evening, and stories on social media. The risk of the disease has become abstract and the risk of the vaccine concrete, when in reality it is the other way round.

Hesitancy is not a fringe position. In a survey of 5,736 parents across eighteen European countries, 56% described themselves as not at all hesitant and another 24% as somewhat hesitant; confidence was lowest in Bulgaria and Poland, and highest hesitancy was found among parents who consulted homeopaths. In the US, exemptions among kindergartners rose to 3.3% in the 2023–24 school year, and 14 jurisdictions reported that more than 5% of kindergartners were exempt from at least one vaccine.

So if you are reading this asking “should we or shouldn't we?”, you are neither unusual nor “anti-vax”. You want to understand. What follows is not a campaign but the studies: what they showed, with what limitations, and where the edge of current knowledge lies.

The immune system learns from experience. The first time a pathogen gets in, it needs days to recognise it and produce the antibodies and cells that will destroy it. In that time the disease has room to develop — sometimes severely, sometimes fatally. After recovery, a memory remains: next time the response kicks in within hours, and most people don't even notice they were infected.

A vaccine provides that memory without the disease. It shows the immune system a “wanted poster” of the pathogen: a weakened or inactivated virus, a single protein from its surface, a neutralised toxin, or — in mRNA vaccines — instructions for the body's own cells to briefly make one viral protein. The immune system reacts as if to a real threat — hence the fever, the sore arm, the tiredness — and remembers.

Two consequences follow. First, a vaccine does not work instantly — protection builds over weeks. Second, the closer the “wanted poster” is to the real pathogen and the more stable the pathogen, the better the protection. Measles changes little, and two doses of vaccine protect against it by 96%. Flu changes every year, so the vaccine has to be updated and protects more modestly. More on that below.

The myth that vaccines cause autism was born in Europe — in a 1998 paper in The Lancet by the British doctor Andrew Wakefield — and Europe still lives with the consequences. In Romania, an EU member state, measles vaccination coverage was above 97% before 2010; by 2022 it had fallen to 83.4% for the first dose and 71.4% for the second. In 2023 the WHO European Region recorded 60,860 measles cases and 13 deaths, 95% of them in six countries including Romania.

The US declared measles eliminated in 2000. A 2026 analysis in Nature Medicine found that average susceptibility to measles roughly doubled after the pandemic, from about 5% to 10%, and that at the level of individual schools transmission had crossed the epidemic threshold by 2022–2023 — a shift invisible in county-level averages. In 2025–2026, more than 3,500 cases put the country's elimination status under formal review. Nationally, kindergarten MMR coverage stood at 92.7% in 2023–24; four years earlier, coverage across all required vaccines had been about 95%.

One more European lesson — about coincidence. In 2008, during a measles vaccination campaign in Ukraine, a person died. The death was temporally but not causally related to the vaccine. The mishandled response collapsed public trust: polio vaccine coverage fell from 91% to 15% by 2015, when two children in western Ukraine were paralysed. Timing is not causation. Every day, people fall ill for reasons that have nothing to do with medicine; when it happens the day after a vaccine, the mind joins the dots. Only a comparison with the unvaccinated can tell coincidence from cause — and that is what the rest of this article is about.

The second common fear is the ingredients. Many vaccines contain aluminium salts: an adjuvant, a substance that strengthens the immune response so that less antigen is needed. Whether it does harm is a fair question, and it has been answered at scale.

A 2025 Danish study followed 1,224,176 children, taking advantage of the fact that the aluminium content of the childhood schedule changed over the years. It examined 50 chronic conditions — autoimmune, allergic, asthma, autism and ADHD. No increased risk was found for any of the 50. For neurodevelopmental disorders, risk per additional milligram of aluminium was even slightly lower (0.93). The authors are honest about the limit: a very small effect for some rare conditions cannot be statistically ruled out.

Now “too many vaccines for a small child”. A review in Pediatrics brought together evidence on this from several angles: infants can respond to vaccines, a mild cold does not stop them building protection, combination vaccines work as well as separate ones, and vaccinated children do not get more infections from other pathogens. And a surprising fact: children today receive fewer antigens from vaccines than 40 or 100 years ago, because modern vaccines are far more precise. Every day a child meets thousands of foreign proteins in food, on hands and toys. A few antigens in a syringe are a drop in that ocean.

Measles is one of the most contagious infections known. In a fully unprotected population, one person infects on average 12 to 18 others — though a systematic review found real estimates vary even more widely depending on the setting. That is why stopping it takes coverage of about 93–95%, well above what most other diseases need. Gaps hide inside averages: in a large Ohio primary care network, even 20 months after a measles outbreak, timely first-dose MMR coverage was 53.6%.

Nor is it a harmless rash. Even in Europe, far from the world's poorest health systems, the first half of 2018 brought more than 41,000 cases and 37 deaths across the WHO European Region — roughly one death per 1,100 cases. In low- and middle-income countries, a 2023 systematic review put 2019 mortality at 1.32% among patients up to 35 years old outside hospital and 3.03% among babies under one.

The least-known part is that measles keeps doing damage after recovery. A 2019 study in Science compared the blood of 77 unvaccinated children before and two months after measles: the virus wiped out between 11% and 73% of the “library” of antibodies they had built before — part of their protection against other infections they had already survived. This did not happen in vaccinated infants. Another paper in Science showed at population level that after measles epidemics, deaths from other infections rise for a further two to three years. “Better to catch it naturally” is not natural protection; it is a risk of losing some of what you already had.

Not all vaccines are equally strong, and pretending otherwise costs trust. The best example is the annual flu vaccine.

According to a 2018 Cochrane review (52 trials, more than 80,000 people), in healthy adults it lowers the chance of confirmed flu from 2.3% to 0.9%. To spare one person the flu, 71 people need to be vaccinated. The effect on hospital admissions is small and uncertain, and on days off work close to zero. The authors also note that 29% of the included trials were funded by industry.

That does not make the flu vaccine useless. It gives healthy adults moderate protection. The review covered healthy people aged 16 to 65, so for older people, pregnant women and people with chronic illness the question is worth discussing with a doctor separately. What it does mean is this: “vaccines” are not one product. The measles vaccine protects 95% after one dose; the flu vaccine, far more modestly. An honest conversation about vaccination starts by not lumping them together — in praise or in criticism.

For contrast, consider whooping cough vaccination in pregnancy. When England introduced it during an outbreak in 2012, protection for babies under three months was 91% — the age group most likely to die of whooping cough, because they are too young for their own vaccines.

COVID-19 vaccines are the most argued-over, so here is only what can be checked. In the randomised trial of the Pfizer vaccine in 43,548 participants, symptomatic COVID-19 was confirmed in 8 vaccinated people and 162 who received placebo — 95% efficacy. Of 10 cases of severe disease, 9 were in the placebo group.

The limits are documented too. Over six months of follow-up, efficacy gradually declined — to 91.3% on average over the period — though against severe disease it stayed at 96.7%. And there is the real risk of myocarditis in young men discussed above.

The overall effect was estimated by modelling: in the first year of vaccination across 185 countries, vaccines prevented 14.4 million deaths based on official death counts — or 19.8 million based on excess mortality. That is a model with assumptions, not a direct count, but the order of magnitude is clear.

Real contraindications exist, but they are few, and they are decided by a doctor, not a forum. The most important is a severe allergic reaction to a previous dose of the same vaccine or one of its components. Live vaccines (such as MMR) are usually not given during pregnancy or to people with severe immune deficiency. With an acute illness and high fever, vaccination is often postponed until recovery — but a mild cold, according to the review in Pediatrics, does not stop a child building protection.

If a previous vaccine caused seizures or a strong reaction, or if a child has a chronic illness, that is not a reason to refuse — it is a reason to talk to a doctor about timing and setting. That is exactly what paediatricians and family doctors are for, and they are a better guide than anonymous stories online. In the European parent survey, parents who saw paediatricians were less hesitant than those who saw general practitioners, and those who consulted homeopaths were the most hesitant of all.