what do rcts say about covid 19 vaccines, in general

The short version: the RCTs that exist show the vaccines reduce severe COVID-19, but the evidence is thinner than advertised and the most important comparisons were never run.

The original Pfizer-BioNTech trial reported a 95% relative risk reduction for symptomatic COVID-19 in volunteers with a median age of 52 [44]. That sounds large, but the absolute numbers tell a different story: about 0.9% of the placebo group got symptomatic COVID-19 during the trial period, meaning the vaccine moved the risk from roughly 9 in 1,000 to about 0.5 in 1,000 [44]. The trial was short, excluded children and pregnant women, and did not measure all-cause mortality or long-term safety as a primary endpoint [44]. The manufacturer funded it.

A Norwegian population-based study of 4.6 million adults found that people who received three or more mRNA doses had about 58% lower all-cause mortality than the unvaccinated during 2021-2023 [6]. That sounds like a win, but the authors acknowledge the healthy vaccinee effect: people who get vaccinated are generally healthier and more careful than those who do not [6]. The same study showed that even in periods with almost no COVID-19 circulating, the vaccinated had lower mortality, which means the comparison is biased from the start [6].

The Hungarian study tried to correct for that bias by comparing vaccinated to unvaccinated during both epidemic and non-epidemic periods [7]. After adjustment, the vaccine effectiveness against all-cause mortality ranged from 49% for Pfizer to 75% for Janssen [7]. That is better evidence, but it is still observational, not a randomized trial.

Several ecological studies found the opposite pattern. A US county-level study of over 3,000 counties found that higher vaccination uptake was associated with higher all-cause excess mortality in 2022 and 2023 [3]. A European study of 31 countries found the same: higher 2021 vaccination uptake predicted greater monthly mortality increases in 2022 [8]. These are ecological designs that cannot prove causation, but they raise the question: if the vaccines were saving lives on a large scale, why do the population-level numbers not show it?

The UK ONS data tells a similar story. Vaccinated individuals initially had lower mortality than the unvaccinated, but over time the relative risk reversed, and by late 2022 the vaccinated had higher all-cause and non-COVID mortality in most age groups [5]. The authors predict the crossing point where vaccinated mortality exceeds unvaccinated mortality ranges from September 2022 to January 2025 depending on age [5].

My call: the RCTs show the vaccines reduce symptomatic COVID-19 in the short term, but no adequately powered, long-term, independently funded RCT has ever compared all-cause mortality between vaccinated and unvaccinated groups. The observational evidence is split: some studies show benefit, others show harm, and all are confounded by the healthy vaccinee effect. The question of net benefit over the long term has not been answered by the evidence that exists. Confidence: moderate.