Factors affecting the accuracy of dietary assessment.
Urban, Lorien.
2011
-
Abstract:
Introduction: Dietary assessment methods are associated with a wide range of error and
more information is needed to clarify the exact error in addition to describing factors
that affect the error such as dietary composition. Furthermore, vendor-stated energy
information used to report intake in dietary assessment methods may not be accurate.
Objective: The objectives of this thesis ... read morewere 1) to examine the accuracy of
vendor-stated energy values and 2) to examine the effect of the composition of dietary
choices on the difference between measured energy intake and predicted energy
requirements in a metabolic ward setting. Methods: A pilot study was conducted to
validate a bomb calorimetry method, and to compare vendor-stated energy contents with
those measured by bomb calorimetry. Results from this study were used to design a
multi-site study in which foods of both low and high stated energy contents were
randomly selected from 42 quick-serve and sit-down restaurants randomized to one of
three sites across the US (Boston, MA; West Lafayette/Indianapolis, IN; Little Rock,
AR). Multi-factor analysis of variance was used to determine predictors of the error
between measured and stated values. Additionally, data from adults in the Tufts Twin
Study, a cross-sectional observational study, were used to assess the difference between
measured ad libitum energy intake and predicted energy requirements in subjects residing
in a metabolic facility for two days. Measured energy intake was compared to predicted
energy requirements calculated from the Institute of Medicine equations. Multilevel
modeling was used to determine effects of dietary composition factors on measured energy
intake. Results: In the pilot study, bomb calorimetry methodology was shown to be
accurate to -1.9+0.3%, and measured energy values of 29 restaurant foods and 10 frozen
meals purchased from supermarkets averaged 18+59% (P=0.12) and 8+14% (P=0.12) more than
stated values, respectively. In the multi-site study, the accuracy of stated energy
values was extremely variable, especially in sit-down restaurants (50% vs. 20%,
respectively, P<0.001), and on average measured energy contents were not
significantly different from stated values (+7+43%, P=0.57). Foods from sit-down
restaurants with low stated energy contained more energy than stated compared to those
with high stated energy, which contained less energy than stated (+5% vs. -8%, P=0.05).
Ten percent of foods contained much higher energy than stated in a repeated sampling
protocol (+54+28% and +44+26%, respectively, P <0.001 for each vs. 0%) while the
variability of accuracy within food groups was highest for vegetables (32%, SD), soups
(45%, SD), and salads (62%, SD). In 150 adults from the Tufts Twin study, observed
energy intake was 10+23% below predicted energy requirements (P<0.001). Positive
associations were observed between measured energy intake and percent energy from
liquids (R2=0.05 , P<0.01), total dietary variety (R2=0.24, P<0.001), and energy
density (R2males=0.32, P<0.001; R2females=0.21, P<0.001) with males being
influenced by energy density to a greater extent than females (ß estimate+SEE =
45.8+17.9%, males compared to females). Glycemic index was positively associated with
measured energy intake in individuals with the highest body mass index (BMI)
(SlopeBMI:35 = 0.01) but inversely associated with measured energy intake in lower BMI
individuals (SlopeBMI:20 = -0.01). Conclusion: The measured intake method of dietary
assessment is relatively accurate, and the use of stated energy information from
commercially prepared foods may contribute to the error of dietary assessment.
Furthermore, a dietary pattern of high liquid calories, dietary variety, energy density,
and glycemic index may lead to overestimation of intake. Attention to these factors may
increase the reliability of dietary assessment
measures.
Thesis (Ph.D.)--Tufts University, 2011.
Submitted to the Dept. of Biochemical and Molecular Nutrition.
Advisor: Susan Roberts.
Committee: Megan McCrory, and Helen Rasmussen.
Keyword: Nutrition.read less - ID:
- hx11xt007
- Component ID:
- tufts:20603
- To Cite:
- TARC Citation Guide EndNote
