Researchers lack the rich evidence base and benchmark patient outcomes needed to evaluate the effectiveness of medical education practice and guide policy. The authors offer a framework for medical education research that focuses on physician-influenced patient outcomes that are potentially sensitive to medical education. Adapting the concept of ambulatory care sensitive conditions, which provided traction to health services research by defining benchmark patient outcomes to measure health system performance, the authors introduce the concept and propose the adoption of educationally sensitive patient outcomes and suggest two measures: patient activation and clinical microsystem activation. They assert that the ultimate goal of medical education is to ensure that measurement of future physicians' competence and skills is based not only on biomedical knowledge and critical clinical skills but also on the ability to translate these competencies into effective patient- and systems-level outcomes. The authors consider methodological approaches and challenges to measuring such outcomes and argue for large, multiinstitutional, prospective cohort studies and the development of a national Database for Research in Education in Academic Medicine to provide the needed infrastructure. They advocate taking the next steps to establish an educational evidence base to guide the academic medical centers of the 21st century in aligning medical education practice with health care delivery that meets the needs of individuals and populations.