Spontaneous reporting databases are convenient sources of information for identifying potential toxicities of medical products, but are subject to deficiencies that limit the interpretability of analyses based on the data they contain. This chapter describes approaches that are variations of traditional epidemiologic and pharmacovigilance approaches for systematically exploring conventional databases of accumulated health information. These approaches differ from conventional analyses of spontaneous reporting databases and analyses of observational databases primarily in the source of the informational substrate used for the analyses and in the techniques needed to put this material into a form amenable to signal-detection analysis. An evaluation of various detection methods, reporting odds ratio, temporal pattern discovery, incidence rate ratio (IRR), Longitudinal Gamma Poisson Shrinker (LGPS), hierarchical Bayes, case–control, and self-controlled case series (SCCS) applied to seven European electronic health databases with prespecified positive and negative controls demonstrated area under the curves (AUCs) in the range of 0.7–0.8.