Development of a matching file of Australian food composition databases (AUSNUT 2007 to 2011–13)

• An AUSNUT 2007 to 2011–13 matching file was created using a systematic approach.
• A total of n = 1070 AUSNUT 2007 foods were matched appropriately via existing files.
• n = 2804 foods required manual matching using conceptual and nutritional similarities.
• The matching file was applied to clinical trial data originally collected in AUSNUT 2007.
• The file allows conversion of data from AUSNUT 2007 to the most recent database.

Changes to the food supply and analytical methods necessitate updating food composition databases over time. In Australia, survey-specific databases include AUSNUT 1999, 2007, and 2011–13. There is no standardized method to match AUSNUT 2007 foods to 2011–13 counterparts. This study aimed to develop an AUSNUT 2007 to 2011–13 matching file, and demonstrate its use in the clinical trial context. Food items aligned with AUSNUT 2007 were back-matched to 1999 and then forward to 2011–13 using food identification codes and existing matching files. Any unmatched AUSNUT 2007 foods were manually matched to appropriate 2011–13 foods based on conceptual and nutritional similarities. The file was then applied to clinical trial data originally collected using AUSNUT 2007. Of the n = 3874 products in AUSNUT 2007, n = 1270 were initially matched to 2011–13 equivalents using existing matching files. Of these foods, n = 1070 were deemed to have an acceptable one-to-one match. A total of n = 2804 AUSNUT 2007 foods required manual matching. Application to clinical trial data found small differences in nutrient intake between original and converted data. The AUSNUT 2007 to 2011–13 matching file will facilitate conversion of dietary data originally collected using AUSNUT 2007 into 2011–13 foods and groups, and to allow re-coding of intake data.