Influenza vaccination: policy versus evidence

No gap between policy and evidence

Jefferson believes that his systematic reviews show that influenza vaccines “have little or no effect on the effects measured.”1 He wonders why there is a gap between evidence and policy.

Jefferson identifies three reasons why his evidence contradicts policy.

Firstly, policy relies heavily on non-randomised studies. Yet his preference for randomised controlled trials is based on historical, pragmatic, and heuristic reasons and cannot be justified on epistemological grounds alone.2 He argues that heterogeneity among studies and study years3 can be overcome by “averaging” outcomes over several years, and finds observational studies yield relative risk reductions that are statistically significant.

Secondly, he cites the lack of evidence of vaccination on effects “at the centre of campaign objectives,” yet he documents such evidence for older adults, barely fails to show it for younger adults, and is unable to consider influenza-related otitis media, a frequent complication in young children.

Thirdly, he is concerned that published studies lack safety data, but 300 million doses are used each year4 and safety problems would not require a systematic review to be detected.

The goal of vaccination policy is not to “prevent seasonal outbreaks” of influenza 1 but to prevent costly influenza-related hospital admissions and deaths. In formulating policy, health officials consider three types of evidence.

Firstly, influenza vaccine must work, and randomised controlled trials provide this evidence.

Secondly, health officials must understand the population burden of disease and know whether it can be reduced by vaccination. Observational studies can help document reductions in attributable (not relative) risk following vaccination.3 5

Thirdly, health officials must know whether vaccination will be economically worth while, and evidence for this must come from other sources.5

There is no “gap between policy and evidence.”1 Systematic reviews can never provide the three types of evidence needed to formulate policy for influenza vaccination.