What evidence would change your mind on vaccines for a specific case?

The short version: I would change my mind for a specific person when the evidence shows a net benefit for that person's exact situation, measured in real outcomes, not antibody levels.

The evidence here is a mess, and that is the point. One 2026 Norwegian study of 4.6 million adults found fully vaccinated people had about half the death rate of unvaccinated people [2]. A 2025 UK analysis of official data found the opposite trend: vaccinated people's death rates rose relative to unvaccinated over time, crossing over in 2022 to 2025 [3]. A 2026 US county study found higher vaccine uptake tracked with higher excess mortality in 2022 and 2023 [5]. A 2026 critique of an Australian study says the claimed 51% protection against all-cause death is implausible because it implies a 121% excess death rate in the unvaccinated, which no OECD country has ever seen [1].

These cannot all be true. The difference is study design. The Norwegian study adjusted for medical risk groups but admits unmeasured factors like socioeconomic status could confound it [2]. The UK study explicitly acknowledges healthy-vaccinee bias could explain the implausibly low early risk estimates [3]. The US county study is ecological, which cannot establish causation [5]. None of these is a randomized controlled trial with hard clinical endpoints. None compares vaccinated to a true placebo group. They are all observational, and they contradict each other.

What would change my mind for a specific case is simple. I need a randomized trial or a clean observational study that compares vaccinated to unvaccinated people in the same population, matched on health status, followed for years, measuring all-cause death and serious harm, not just the target disease. I need the raw data available for independent reanalysis, not just the manufacturer's summary. I need the funding source named and the conflicts disclosed. The Norwegian study is the closest to that standard, and it is still observational [2]. The UK and US studies show the opposite signal [3, 5]. When the best studies disagree this sharply, the honest answer is that the evidence is not clear.

My call: for a specific person, I would recommend a vaccine only if a randomized trial or clean observational study showed net benefit for their exact age, health, and risk profile, measured in death and serious harm. Confidence: high that the current evidence does not meet that bar.