Maintaining national food databases to enhance precision: A report from the preconference workshop 1 of the 2025 international diet and activity methods conference, Toronto, Canada, April 27, 2025

National food composition databases are essential for data quality and reliability in dietary monitoring, nutrition research, and policy. Nevertheless, they face growing challenges from rapidly evolving food markets, new data demands, and limited analytical resources. This paper summarizes outcomes of a preconference workshop at ICDAM 2025 that brought together experts from seven countries to share experiences in developing and maintaining databases. Presentations highlighted strategies to keep databases fit for purpose, including standardized and harmonized food descriptions, inclusion of biodiversity and regional foods; maintaining strong analytical anchors with continuous updates; systematic handling of missing values (recipe calculations, preparation factors, and imputation), and AI-based linking of branded with generic foods to address nutrient gaps and capture market changes. These developments also underline the value of food composition databases: supporting national dietary surveys, monitoring reformulation, informing front-of-pack labeling, and enabling evaluation of nutrition policies. Breakout discussions emphasized challenges such as resource constraints and reliance on secondary data, as well as demands for new data types beyond traditional nutrients, alongside opportunities in branded data, automation, FAIR data standards, and harmonization. The workshop concluded that sustained investment, modernized infrastructures, and collaborative governance are critical to ensure food composition databases remain robust, policy-relevant, and globally interoperable. • National food composition databases support nutrition research and policy. • Seven countries reported shared challenges and efforts to strengthen data. • New strategies for building food composition databases are emerging globally. • Branded food data help monitor reformulation, but open access is still limited. • Missing nutrients remain a challenge, tackled with data and machine learning. • Linking to other databases adds data on bioactives, contaminants and more.