The committee’s bibliographic retrieval was posted on the project website with a request for public comment regarding missing articles.2 The committee received one submission, which was reviewed. The bibliography was separated into two sections. Section I contained those articles on which the committee focused its initial review. Section II contained those citations for articles that did not meet the committee’s criteria (i.e., original research, vaccine used in the United States, adverse event within the committee’s scope, animal or in vitro studies of relevance).
The committee made three assessments for each relationship reviewed. The first assessment applies to the weight of evidence from the epidemiologic literature; the second applies to the weight of evidence from the biological and clinical (mechanistic) literature. The third assessment is the committee’s conclusion about causality. In assessing the weights of evidence, each individual article (or findings within an article if more than one outcome or vaccine was studied) was evaluated for its strengths and weaknesses. The committee then synthesized the body of evidence of each type (epidemiologic or mechanistic) and assigned a “weight of evidence” for each. These weights of evidence are meant to summarize the assessment of the quality and quantity of evidence. The committee then reviewed the two weight-of-evidence assessments in order to make a conclusion about the causal relationship. The committee’s approach to each of these three assessments will be discussed in the following sections.
Experimental studies (trials) are generally considered more rigorous than observational studies; controlled studies are generally considered more rigorous than uncontrolled studies. A brief description of major study designs and methodological considerations can be found in Appendix A. Surveillance studies were reviewed, but the absence of a control group limited their contribution to the weight of epidemiologic evidence; studies that included individual case descriptions were reviewed for their contribution to the evaluation of mechanistic evidence (discussed in subsequent sections). Small clinical studies that were not controlled for vaccine administration were generally reviewed for contributions to the mechanistic weight of evidence.