Based upon a review and analysis of selected literature, the author presents a conceptual model of discrepancy analysis evaluation for planning, implementing, and assessing the impact of continuing medical education (CME). The model is described in terms of its value as a means of diagnosing errors in the development and implementation of CME. The role of explanation and prediction within the design and implementation of CME is discussed, and implications for the development of practical theories of CME are examined. Finally, the model is used as a vehicle for articulating areas of responsibility and accountability of continuing medical educators in the patient health status/physician performance/physician competence relationship. This article introduces the notion of a discrepancy between what is and what ought to be--a gap analysis--as a means to assess needs and outcomes. Needs and outcomes were related to gaps in competencies, clinical performance, and patient health. The article proposes that gaps in patient health before and after programs can be explained as either due to clinical performance or the effects of other factors (systems). It proposes that gaps in clinical performance can be attributed to gaps in competence or to other factors, such as barriers to change. This inductive and deductive reasoning and the formal explanations associated with gaps uncovered in assessments were viewed as essential elements in explaining success and failure. In theory, the formal explanation of gaps in competency, performance, and patient health status was key to attributing change to CME performance. For three decades these themes have provided a framework for building CPD that integrates assessment, theory, and program design. (Contains 2 figures.)