Are vaccines safe and effective

The short version: the question is too broad to answer honestly. "Vaccines" are dozens of different products for different diseases in different populations. Safety and effectiveness vary by vaccine, by person, and by outcome. There is no single answer.

The evidence you asked for covers many vaccines, but the strongest data comes from COVID-19 mRNA vaccines, where the picture is genuinely mixed. A Norwegian study of 4.6 million adults found that people who received three or more mRNA doses had about a 58% lower adjusted all-cause mortality rate than unvaccinated people over 2021-2023 [2]. That sounds clear. But a UK study using the same kind of government data found the opposite trend over time: vaccinated individuals started with lower mortality but crossed over to higher mortality than unvaccinated by late 2022, with relative risks between 1.7 and 2.3 for first-dose recipients [3]. A US county-level study found that higher vaccination uptake tracked with higher excess mortality in 2022 and 2023, even after controlling for prior deaths [5]. A European country-level study found the same pattern: higher 2021 vaccination uptake predicted larger monthly mortality increases in 2022 [9].

These contradictions matter. The Norwegian study adjusted for medical risk groups, sex, and calendar time, but the authors acknowledge that residual confounding from unmeasured factors like socioeconomic status could explain the large protective effect [2]. The UK and US studies cannot prove causation either, but they show that the simple "vaccinated live longer" story depends on when and how you measure.

For other vaccines, the evidence is thinner. A meta-analysis of six observational studies found that influenza vaccination in people with diabetes was associated with about a 46% lower odds of all-cause mortality [11]. But the heterogeneity was extreme (I²=96%), meaning the studies disagreed with each other, and all were observational, so healthy-user bias is a real concern. A pneumococcal vaccine meta-analysis of 35 studies found no significant reduction in all-cause mortality overall, with I² values of 100% across nearly every subgroup [18]. That means the studies were so inconsistent that pooling them is meaningless.

The safety data from passive surveillance systems like VAERS is structurally limited. A VAERS study of DTaP in children found 329 reported SIDS cases, with 68% clustering within 6 days of vaccination [36]. The authors call this a temporal association, not causation, but the clustering is not random. A VAERS study of hepatitis B vaccine found disproportionate signals for tendon fibrosis (ROR 252) and myofascitis (ROR 108) [46]. These signals do not prove causation, but they are the kind of pattern that should trigger controlled studies, which have not been done.

The honest answer is that for most vaccines, the evidence for all-cause mortality benefit is weak, conflicted, or absent. The COVID-19 vaccines show the strongest signal of benefit in some studies and the strongest signal of possible harm in others. The difference depends on time horizon, population, and how you handle the fact that healthier people get vaccinated first. No vaccine has been tested against a true placebo in a long-term randomized trial with all-cause mortality as the primary endpoint. That is the study design that would settle the question, and it has never been run.

My call: the evidence does not support a blanket "vaccines are safe and effective" statement. It supports a case-by-case evaluation where the burden of proof has not been met for most vaccines on the outcomes that matter most. Confidence: high that the evidence is insufficient for a universal claim.