Clinical trials conducted for regulatory approval may include outcomes that are informative but not routinely collected in clinical practice. This situation can be problematic when pragmatic clinical trials (PCT) seek to use electronic health record (EHR) data to test the effectiveness of medical products and services in actual practice settings. We use TIMI bleeding events to illustrate how a complex clinical trial endpoint can be implemented using EHR data. While we were able to demonstrate that our EHR-defined bleeding events were associated with differences in patient clinical outcomes, we are not confident that these measurements could be replicated in other locations with consistent reliability and validity. We believe the development of PCT endpoint definitions is an important issue that should be addressed by medical and informatics professional societies, regulators and the medical products industry.