Lack of continuity is a problem for survey-based epidemiological research. Poor access to experience and inadequate retention of information means new projects seldom build on successful predecessors to the maximum possible extent. We have designed a data management framework which addresses this problem using a well-known information systems approach. Our ideas provide a means of storing data in a manner which reflects the original research concepts. This makes collected data accessible and understandable without reference to the questionnaire, and allows new questionnaires to be designed quickly and easily. Our framework provides a means for recording developmental, conceptual and other supporting information and documentation which is traditionally poorly conserved. We illustrate the components of the framework using a major Australian epidemiological project as a case study. Once the necessary software is developed, our framework will improve organisational memory within individual research units and has the potential to become a valuable support tool for survey-based research.