Using Big Data for Risk Quantification of Rare Medical Associations