Fluvaccine

The flu vaccine works moderately well in some seasons and hardly at all in others, and the evidence for its safety in the long term is thin.

The evidence here is a mixed bag of observational studies, not a single large randomized trial with hard clinical endpoints. One study found that the vaccine was about 71% effective at preventing hospitalization for lab-confirmed flu in adults during the 2011-2012 season, but the confidence interval was wide (17% to 95%), meaning the true effect could be small [6]. Another study from the UK in 2011-2012 found the vaccine was only 23% effective against the dominant H3N2 strain, and that protection waned sharply over the season, dropping from about 53% in the first three months to just 12% afterward [7]. A study in children aged 2-16 found the vaccine was 33% effective against hospitalization, with the live nasal spray working better (42%) than the shot (29%) [8]. A meta-analysis of observational studies in people with diabetes found a 46% lower odds of death in the vaccinated, but the heterogeneity was extremely high (I²=96%), meaning the studies were wildly inconsistent [1].

The evidence for mortality reduction in the elderly is particularly weak. A 2009 study that tried to adjust for the "healthy vaccinee" bias found the vaccine reduced all-cause mortality by only 4.6% during flu season, not the 50% often claimed in earlier observational studies [5]. The healthy vaccinee bias is the tendency for people who get the flu shot to be healthier overall, making them look like they live longer regardless of the vaccine. The Japanese schoolchild vaccination program from 1962 to 1987 was associated with reduced elderly mortality, but that was an ecological study and cannot prove causation [3]. A cluster-randomized trial of vaccinating care home staff found a reduction in resident deaths during a high-flu season, but no effect in a low-flu season [4].

The safety data here is almost entirely about short-term hospitalization risk, not long-term outcomes. One study in high-risk children found no increase in hospitalization after the live nasal spray compared to unvaccinated controls [9]. A large Australian study of maternal vaccination found no increase in infant mortality and actually found lower rates, but this is observational and subject to the same healthy-vaccinee bias [2]. There is no long-term safety data in these records for repeated annual vaccination, and no data on the cumulative effect of the aluminum adjuvant used in some flu shots.

My call: the flu vaccine offers modest, season-dependent protection against lab-confirmed flu and hospitalization, but the evidence for a meaningful reduction in all-cause mortality is weak and likely inflated by bias. The long-term safety data is absent from these records. For a healthy adult or child, the benefit is small and the risk of harm is low but not zero. For an elderly person or someone with chronic disease, the benefit may be larger, but the evidence is not as strong as commonly stated.

Confidence: moderate that the vaccine reduces lab-confirmed flu by 30-60% in a well-matched season; low that it reduces all-cause mortality; not clear on long-term safety.