ARS | Publication request: Assurance of Quality Data for Food Composition Systems

Submitted to: Meeting Abstract
Publication Type: Abstract Only
Publication Acceptance Date: May 9, 2007
Publication Date: June 7, 2007
Citation: Holden, J.M., Bhagwat, S.A., Patterson, K.K. 2007. Assurance of Quality Data for Food Composition Systems. EuroFIR meeting for Workpackages and Data Quality Workshops, June 7-8, 2007, Paris, France.

Technical Abstract: The Nutrient Data Laboratory (NDL), Beltsville Human Nutrition Research Center, provides USDA’s Nutrient Data Base for Standard Reference (SR), the foundation of most other food composition databases and includes information on about 130 components in approximately 7100 foods. Assessing the validity of analytical data to be used in SR is important to maintaining the quality of the information contained in the database. The NDL has developed a series of modules for data quality evaluation. Each module evaluates the quality of data by rating important documentation concerning the analytical method, analytical quality control (QC), number of samples, sampling plan and sample handling. The pertinent information is culled from the published articles to answer specific questions. A set of general questions was developed for each module to objectively assess the information in nutrient data publications. The questions address the points critical to the quality of data which were generated. To date approximately sixty individuals from up to 20 countries have completed the evaluation of two research articles. A preliminary study of the variability of responses between individuals answering these questions about the same publications was conducted to evaluate whether the information on these topics was expressed clearly. Sample handling and analytical quality control modules gave the greatest variability reflecting uncertainty in the subjects’ understanding of the documentation for these categories. Analytical QC as described in manuscripts documenting nutrient data for foods may be incomplete or ambiguous. Even when reference materials (RM), either certified or in-house, are used as analytical controls when determining the composition of foods, the details may not be clearly described in subsequent publications. It is important that standardized procedures be developed and implemented to facilitate the assessment of data quality across systems.