BackgroundHow to assess the impact of research is of growing interest to funders, policy makers and researchers mainly to understand the value of investments and to increase accountability. Broadly speaking the term "research impact" refers to the contribution of research activities to achieve desired societal outcomes. The aim of this overview is to identify the most common approaches to research impact assessment, categories of impact and their respective indicators.MethodsWe systematically searched the relevant literature (PubMed, The Cochrane Library (1990-2009)) and funding agency websites. We included systematic reviews, theoretical and methodological papers, and empirical case-studies on how to evaluate research impact. We qualitatively summarised the included reports, as well the conceptual frameworks.ResultsWe identified twenty-two reports belonging to four systematic reviews and 14 primary studies. These publications reported several theoretical frameworks and methodological approaches (bibliometrics, econometrics, ad hoc case studies). The "payback model" emerged as the most frequently used. Five broad categories of impact were identified: a) advancing knowledge, b) capacity building, c) informing decision-making, d) health benefits, e) broad socio-economic benefits. For each proposed category of impact we summarized a set of indicators whose pros and cons are presented and briefly discussed.ConclusionsThis overview is a comprehensive, yet descriptive, contribution to summarize the conceptual framework and taxonomy of an heterogeneous and evolving area of research. A shared and comprehensive conceptual framework does not seem to be available yet and its single components (epidemiologic, economic, and social) are often valued differently in different models.