Bridging Differences in Outcomes of Pharmacoepidemiological Studies: Design and First Results of the PROTECT Project