a1 Division of Preventive Oncology, Cancer Care Ontario, 620 University Avenue, Toronto, Ontario, Canada, M5G 2L7
a2 Department of Public Health Sciences, University of Toronto, Toronto, Ontario, Canada
a3 Department of Nutritional Sciences, University of Toronto, Toronto, Ontario, Canada
a4 Block Dietary Data Systems, Berkeley, CA, USA
a5 Department of Public Health, University of California, Berkeley, CA, USA
Objective To assess the validity and reliability of the most recent adaptation of Block's full-diet food-frequency questionnaire (FFQ) among a sample of Canadian women.
Design Participants completed a self-administered FFQ (FFQ1), two unannounced 24-hour recalls (weekday and weekend) and a second FFQ (FFQ2) between October 2003 and February 2004. FFQs and recalls were analysed for 32 nutrients using Block Dietary Data Systems and the University of Minnesota's Nutrient Data System. Mean and median intakes were computed, along with crude and deattenuated Pearson correlation coefficients between FFQ1 and the average of two recalls (validity) and between FFQ1 and FFQ2 (reliability).
Setting Ontario, Canada.
Subjects A random population-based sample (n = 166) of women aged 25 to 74 years.
Results One hundred and fifteen (69%) women completed FFQ1, 96 completed FFQ1 and both recalls, and 93 completed both FFQs, about 56 days apart. Mean intakes were similar for most nutrients. FFQ reliability was high, with Pearson correlation coefficients having a median of 0.75, ranging from 0.57 to 0.90 (macronutrients) and from 0.65 to 0.88 (micronutrients from supplements and food). FFQ validity was moderate to high, with deattenuated Pearson correlation coefficients having a median of 0.59, ranging from 0.11 to 0.73 (macronutrients) and from 0.50 to 0.76 (micronutrients from supplements and food). Our micronutrient correlations were similar to or higher than those of other studies that included supplements. Two correlations <0.40 were associated with fats.
Conclusions The validity and reliability of this full-diet version of the Block FFQ were moderate to high, supporting its use in future studies among Canadian women.
(Received December 14 2004)
(Accepted May 18 2005)