Use of household supermarket sales data to estimate nutrient intakes: a comparison with repeat 24-hour dietary recalls.

Electronic supermarket sales data provide a promising, novel way of estimating nutrient intakes. However, little is known about how these data reflect the nutrients consumed by an individual household member. A cross-sectional survey of 49 primary household shoppers (age [mean+/-standard deviation age]=48+/-14 years; 84% female) from Wellington, New Zealand, was undertaken. Three months of baseline electronic supermarket sales data were compared with individual dietary intakes estimated from four random 24-hour dietary recalls collected during the same 3-month period. Spearman rank correlations between household purchases and individual intakes ranged from 0.54 for percentage of energy from saturated fat (P<0.001) to 0.06 for sodium (P=0.68). Other correlation coefficients were: percentage of energy from carbohydrate, 0.48; and protein, 0.44; energy density of nonbeverages, 0.37 (kcal/oz); percentage of energy from total fat, 0.34; sugar, 0.33 (oz/kcal); and energy density of beverages, 0.09 (oz/kcal; all P values <0.05). This research suggests that household electronic supermarket sales data may be a useful surrogate measure of some nutrient intakes of individuals, particularly percentage of energy from saturated and total fat. In the case of a supermarket intervention, an effect on household sales of percentage energy from saturated and total fat is also likely to impact the saturated and total fat intake of individual household members. Copyright 2010 American Dietetic Association. Published by Elsevier Inc. All rights reserved.