The prevailing “segregated model” for understanding clinical research sharply separates it from clinical care and subjects it to extensive regulations and guidelines. This approach is based on the fact that clinical research relies on procedures and methods—research biopsies, blinding, randomization, fixed treatment protocols, placebos—that pose risks and burdens to participants in order to collect data that might benefit all patients. Reliance on these methods raises the potential for exploitation and unfairness, and thus points to the need for independent ethical review and more extensive informed consent. In contrast, it is widely assumed that clinical care does not raise these ethical concerns because it is designed to promote the best interests of individual patients. The segregation of clinical research from clinical care has been largely effective at protecting research participants. At the same time, this approach ignores the fact that several aspects of standard clinical care, such as clinician training and scheduling, also pose some risks and burdens to present patients for the benefit of all patients. We argue that recently proposed learning health care systems offer a way to address this concern, and better protect patients, by developing integrated review and consent procedures. Specifically, current approaches base the need for independent ethical review and more extensive informed consent on whether an activity is categorized as clinical research or clinical care. An ethically sounder approach, which could be incorporated into learning health care systems, would be to base the need for independent ethical review and more extensive informed consent on the extent to which an activity poses risks to present patients for the benefit of all patients.