Helaine R H Rockett1, Catherine S Berkey, Graham A Colditz. 1. Channing Laboratory, Department of Medicine, Brigham and Women's Hospital, Harvard Medical School, Boston, MA, USA. helaine.rockett@channing.harvard.edu
Abstract
INTRODUCTION: Our purpose was to design and evaluate a shorter version of our 126-item food frequency questionnaire (long FFQ) for use with adolescents. A shorter FFQ is needed that can reliably rank research subjects according to their intakes of energy, macronutrients and selected micronutrients. METHODS: Dietary data were collected annually, for 3 years, using the full-list FFQ from 16 882 participants of the Growing Up Today Study (GUTS). From this full-list FFQ data, the top ten food contributors for energy and each macronutrient, and the top five food contributors for eight other selected micronutrients were compiled to create a 26-item (short-list) FFQ. Arithmetic means and Pearson correlations were computed to assess relationships between nutrient intakes estimated from the short- and full-list FFQs. We further compared both FFQs with three 24-hour recalls (approximately every 4 months over a 1-year period). Linear regression models were fitted, using energy and nutrients estimated from the short-list FFQ and separately from the full-list FFQ, to see how results may differ. RESULTS: As expected, mean nutrient values from the short-list FFQ were substantially below those from the full-list FFQ. Pearson correlations >0.85 between the short- and full-list FFQs were found for most nutrients. However, correlations between nutrients from the short-list FFQ and the three 24-hour recalls were lower (mean correlation =0.40) than the full-list FFQ. Linear regression models suggested that the short-list FFQ performed nearly as well as the full-list FFQ, for studying associations between energy and several nutrients (trans fatty acids, saturated fat, and glycemic load) and the non-dietary factor, TV viewing. Model betas for energy and nutrients from the short-list FFQ were slightly smaller than betas obtained from the full-list FFQ, but all were statistically significant. CONCLUSION: The short-list FFQ can assess nutrient values of a population for analytic research purposes, such as studying associations between certain dietary intakes and non-dietary factors.
INTRODUCTION: Our purpose was to design and evaluate a shorter version of our 126-item food frequency questionnaire (long FFQ) for use with adolescents. A shorter FFQ is needed that can reliably rank research subjects according to their intakes of energy, macronutrients and selected micronutrients. METHODS: Dietary data were collected annually, for 3 years, using the full-list FFQ from 16 882 participants of the Growing Up Today Study (GUTS). From this full-list FFQ data, the top ten food contributors for energy and each macronutrient, and the top five food contributors for eight other selected micronutrients were compiled to create a 26-item (short-list) FFQ. Arithmetic means and Pearson correlations were computed to assess relationships between nutrient intakes estimated from the short- and full-list FFQs. We further compared both FFQs with three 24-hour recalls (approximately every 4 months over a 1-year period). Linear regression models were fitted, using energy and nutrients estimated from the short-list FFQ and separately from the full-list FFQ, to see how results may differ. RESULTS: As expected, mean nutrient values from the short-list FFQ were substantially below those from the full-list FFQ. Pearson correlations >0.85 between the short- and full-list FFQs were found for most nutrients. However, correlations between nutrients from the short-list FFQ and the three 24-hour recalls were lower (mean correlation =0.40) than the full-list FFQ. Linear regression models suggested that the short-list FFQ performed nearly as well as the full-list FFQ, for studying associations between energy and several nutrients (trans fatty acids, saturated fat, and glycemic load) and the non-dietary factor, TV viewing. Model betas for energy and nutrients from the short-list FFQ were slightly smaller than betas obtained from the full-list FFQ, but all were statistically significant. CONCLUSION: The short-list FFQ can assess nutrient values of a population for analytic research purposes, such as studying associations between certain dietary intakes and non-dietary factors.
Authors: James L Nodler; Holly R Harris; Jorge E Chavarro; A Lindsay Frazier; Stacey A Missmer Journal: Am J Obstet Gynecol Date: 2019-09-14 Impact factor: 8.661
Authors: Catherine L Davis; Sheldon E Litwin; Norman K Pollock; Jennifer L Waller; Haidong Zhu; Yanbin Dong; Gaston Kapuku; Jigar Bhagatwala; Ryan A Harris; Jacob Looney; Celestine F Williams; Aubrey Armento; Michael D Schmidt; Reda Bassali Journal: Int J Obes (Lond) Date: 2019-11-21 Impact factor: 5.095