A hospital's stroke care team believes that for patients with acute cerebral ischemia, a CT angiogram with a perfusion study, rather than simply a routine CT scan, will better direct clinical care. The team will collect data prospectively in a database to document the effect of this new practice. A hospital implements a new program and collects process and outcome data to monitor and improve performance on stroke-related care, including timely receipt of thrombolytic therapy, dysphagia screens for all patients, and deep venous thrombosisprophylaxis for at-risk patients. Are these activities research? This is an increasingly common question, arising most frequently in 2 contexts. In quality improvement (QI) circles, there is considerable concern that the current regulatory definition of research used by institutional review boards (IRBs) may be too sensitive, creating unnecessary burdens on important QI activities.1 In the surgical or procedural setting, in which the history of improvements has followed the path of “surgical innovations” rather than systematic research, there is increasing discussion about the appropriate level of oversight of such activities.2,3 Given the increasing emphasis on evidence-based medicine and the public pressure to improve the quality of health care, the question “Is it research?” will become even more common. The current US regulations define research as a “systematic investigation … designed to develop or contribute to generalizable knowledge” (45CFR46.102d). Since a well-organized QI project involves a systematic change and measurement (of the process or outcome), the question becomes whether it is also “designed” or intended to contribute to generalizable knowledge. The Office of Human Research Protections (OHRP) guidance document (which does not have …