Copyright © National Academy of Sciences. All rights reserved.Adverse Effects of Vaccines: Evidence and Causality
3evidence represent the committee’s assessment of the quality and quantity of evidence. The twoweights-of-evidence assessments contributed to the third assessment, a conclusion about thecausal relationship.
Weight of Epidemiologic Evidence
Each peer-reviewed epidemiologic study was evaluated for its methodologic limitations(e.g., flawed measurement of either vaccine administration or adverse event, or failure to adequately control confounding variables) and for the precision of the reported results (e.g., thewidth of the 95% confidence interval around an effect estimate, reflecting the statistical power todetect a significantly increased risk of an adverse event). A specific study involving multiple outcomes or vaccines could have fewer limitations for the analysis of some vaccines or some outcomes than for others. Small clinical studies can be well conducted but the low number of subjects may limit the ability to detect most adverse events. Although most efficacy studies include a safety component, the results are often nonspecific (e.g., “no serious adverse events were detected”). The committee was rigorous in assessing the strengths and weaknesses of each epidemiologic study. Some studies reviewed are likely the most reasonably methodologically sound given the nature of the exposure and the outcomes, even if the studies have some residual limitation due to the challenges that often attend such research. Summary paragraphs describe the epidemiologic evidence (as well as the mechanistic evidence and in some circumstances thecausality conclusion) more fully than can be captured with the formal and consistent wording of the assessments used in this report.The committee used a summary classification scheme that incorporates both the qualityand quantity of the individual epidemiologic studies and the consistency of the group of studies in terms of direction of effect (i.e., whether the vaccine increases risk, decreases risk, or has no effect on risk). Integral to the assessment is the confidence the committee has that the true effectlies close to the average overall effect estimate for the body of evidence (i.e., collection of reports) reviewed (Schunemann et al., 2010)