Several approaches to the use of health claims on foods have been made around the world, and the common theme is that any health claim will require scientific validation and substantiation. There is also broad consensus that any regulatory framework should protect the consumer, promote fair trade and encourage innovation in the food industry. This paper is based on a critical evaluation of existing international approaches to the scientific substantiation of health claims, with a view to identifying common new ideas, definitions, best practice and a methodology to underpin current and future developments. There is a clear need to have uniform understanding, terminology and description of types of nutrition and health claims. Two broad categories were defined: Nutrition Claims, i.e. what the product contains, and Health Claims, i.e. relating to health, well-being and/or performance, including well-established nutrient function claims, enhanced function claims and disease risk reduction claims. Such health claims relate to what the food or food components does or do. The categories of health claims are closely and progressively related and are, in practice, part of a continuum. Provision is also made for "generic" or well-established, generally accepted claims and for "innovative" or "product-specific" claims. Special attention was paid to reflect the health-promoting properties of a food or food component in such a way as to facilitate the making of risk reduction claims outside the medical scope of the term prevention. The paper sets out basic principles and guidelines for communication of health claims and principles of nutritional safety. The main body of the work examines the process for the assessment of scientific support for health claims on food and emphasises an evidence-based approach consisting of: Identification of all relevant studies exploring the collection of evidence, data searches, the nature of the scientific evidence, sources of scientific data (including human intervention studies, human observational studies, animal studies and in vitro studies, and the use of biomarkers in human studies Evaluation of the quality of individual studies to ensure good experimental design and interpretation Interpretation of the totality of evidence to apply scientific judgement to interpret the weight of evidence as a whole Assessment of significant scientific agreement on a case-by-case basis to agree within the relevant scientific community that an association between a food or a food component and a health benefit is valid. Annexes include an international comparison of regulatory approaches to health claims, suggestions for the documentation and presentation of evidence, and a procedure for reviewing the evidence.