Investigating a safety issue is a complex process, often requiring the evaluation of adverse event (AE) and laboratory data. While these data-streams provide distinct, yet supportive information for the safety issue at hand, there may be a desire to combine them. For instance, one may want to adjust the comparative risk summary for an AE (e.g., odds ratio) by accounting for whether there are differences for a laboratory parameter that serves as marker for the safety issue. Several features of the data make such an endeavor difficult, including AE data being categorical whereas laboratory measurements are continuous. In this talk these two data-streams are integrated using a Bayesian model selection framework. This work is motivated and illustrated using a drug for the bowel preparation for colonoscopy, where the safety issue of interest is electrolyte changes.