The Food and Drug Administration (FDA) can take months to identify, investigate and issue product recalls, often after enough people have gotten sick to trigger negative press. US Health Departments are already using data from Twitter, Yelp, and Google to monitor foodborne illnesses – suggesting that many alternative data sources could provide early signals for product recalls. In this report, we review a public health study that uses a trained NLP engine to predict FDA product recalls using Amazon.com reviews.