Throughout the world there is a growing recognition that health care should be research-led. This strengthens the requirement for expenditure on health services research to be justified by demonstrating the benefits it produces. However, payback from health research and development is a complex concept and little used term. Five main categories of payback can be identified: knowledge; research benefits; political and administrative benefits; health sector benefits; and broader economic benefits. Various models of research utilization together with previous assessments of payback from research helped in the development of a new conceptual model of how and where payback may occur. The model combines an input-output perspective with an examination of the permeable interfaces between research and its environment. The model characterizes research projects in terms of Inputs, Processes, and Primary Outputs. The last consist of knowledge and research benefits. There are two interfaces between the project and its environment. The first (Project Specification, Selection and Commissioning) is the link with Research Needs Assessment. The second (Dissemination) should lead to Secondary Outputs (which are policy or administrative decisions), and usually Applications (which take the form of behavioural changes), from which Impacts or Final Outcomes result. It is at this final stage that health and wider economic benefits can be measured. A series of case studies were used to assess the feasibility both of applying the model and the payback categorization. The paper draws various conclusions from the case studies and identifies a range of issues for further work.