Relative validity of food frequency questionnaire nutrient estimates in the black women's health study

Shiriki K. Kumanyika, David Mauger, Diane C. Mitchell, Brenda Phillips, Helen Smiciklas-Wright, Julie R. Palmer

Research output: Contribution to journalArticlepeer-review

98 Scopus citations

Abstract

PURPOSE: The Black Women's Health Study (BWHS) was designed to investigate determinants of health and disease in US black women. More than 64,000 women are enrolled in the BWHS cohort. This study assessed the relative validity of the 68-item food frequency questionnaire (FFQ) used in the BWHS baseline questionnaire. METHODS: Four hundred and eight BWHS enrollees were asked to provide three telephone, 24-hour recalls and one written 3-day food diary over a one-year period. Means and Pearson correlations were computed to compare estimates for energy, total fat, saturated fat, protein, carbohydrate, dietary fiber, calcium, iron, vitamin C, folate, β-carotene, and vitamin E from the FFQ, recalls, and diaries. RESULTS: Mean energy intake (kcal) was higher for the diary (1716) than the FFQ (1601) or recalls (1510). Other nutrient estimates (% kcal or per 1000 kcal) were similar across methods, except β-carotene (FFQ higher). Correlations (energy-adjusted, except for energy, and corrected for intraperson variation) between the FFQ and the recalls were higher than for the diary data and were between 0.5 and 0.8, except for energy and vitamin E (both <0.3). CONCLUSION: The BWHS FFQ will support meaningful analyses of diet-health associations for 10 of the 11 energy-adjusted nutrient intake variables analyzed.

Original languageEnglish (US)
Pages (from-to)111-118
Number of pages8
JournalAnnals of Epidemiology
Volume13
Issue number2
DOIs
StatePublished - Feb 2003

All Science Journal Classification (ASJC) codes

  • Epidemiology

Fingerprint

Dive into the research topics of 'Relative validity of food frequency questionnaire nutrient estimates in the black women's health study'. Together they form a unique fingerprint.

Cite this