Background: Online sources such as blogs and news feeds provide timely information about public attitudes towards vaccination and are a potentially valuable source for surveillance to guide public health programming. Current approaches to extracting information from these online sources, however, tend to identify only the general sentiment (e.g., positive or negative) or independent mentions of terms, such as the name of a vaccine or disease. This information identifies what is being mentioned, but it does not help public health personnel to understand attitudes towards vaccines. For example, knowing that a blog post refers to MMR is noteworthy, but knowing that the post asserts MMR vaccination causes autism is considerably more useful. This type of information should help public health agencies to identify and understand prevalent concerns regarding specific vaccines and to identify effective interventions.
Objective: The purpose of this study is to develop, evaluate, and apply on a large scale an automated method for Vaccine Attitude Surveillance using Semantic Analysis (VASSA). We will apply this method to online sources such as blogs and news feeds.