I. Nature of Variation in Diet

1
I. Nature of Variation in Diet

• II. Overview of Dietary Methods

2
I Introduction on dietary variation

• There is day-to-day variation in dietary intake.
• Long-term dietary pattern is used as the exposure
  parameter in epidemiologic applications.
• Period of time
• Years--- atherogenesis or cancer
• Weeks--- fetal malformation
• Days,weeks, or months---Physiologic intermediates
• plasma lipids or endogenous hormone excretion

3

• Feature of the dietary intake
• variation from day to day superimposed on an
  underlying consistent pattern
•  

4

• Factors contribute to daily variation in dietary
  intake
• Day of the week
• ex large meals on Sunday/weekends
• Season
• in developing countries without food preservation
  and transportation systems
• The magnitude of influence is largely determined
  by cultural and ecological factors.
• Most of the variation in an individuals diet is
  without an obvious patternrandom variationtrue
  variation and measurement error(single day)

5

•   The degree of random variation differs
  according
• to nutrient
• Total energy intakeleast day-to-day variation
• Macronutrienta somewhat constrained possibility
  for large degree of variation
• Micronutrientlarge variation
• (Fig 3-1) vitamine Agt total fat,
• related to the level of intake

6
(No Transcript)
7

• (Fig 3-2) The consumption of a specific nutrient
  for a single individual(within-person variation)
• (Fig 3-3) Distributions for mean daily intakes
  of total calories, total fat, and vitamine A for
  a group of 194 women(between-person variation)
• (Fig 3-4) a sample of one or several
  days----artificially increase the standard
  deviation---misleading report

8
(No Transcript)
9
(No Transcript)
10
(No Transcript)
11

• The variation in daily intake of specific
  nutrient
•  
• Daily Nutrient Yijkµ subjecti factor Xij
• day of
  weekijk eijk
• factor Xij factors that influence within person
  variance, such as season and day of the week
• eijk random within person variance (day to day
  within person variance not explained by the other
  independent variables)

12

• (Table3-2, 3-3) major contributors to variance
  subject (between person variance) and residue
  (ewithin-person variance)
•  
• Daily Nutrient Yijkµ subjecti e
• e day to day within person variance
•  
• Within-person variance/between person variance
  gt1(Table 3-4)
• within-person variance depending on populations
• in U.S. (similar degree across age and age
  grouping)high within person varianceTable
  3-4,Table 3-5)

13
(No Transcript)
14
(No Transcript)
15
(No Transcript)
16
(No Transcript)
17

• Seasonal variation
• not apparent in U.S.,
• probably more apparent in Taiwan or China than in
  US.
• Year-to-year variation
• minor in comparison with day-to-day
• Within-person-variation random? dependent?
• --homeostatic mechanisms (probably both
  physiologic and cultural)
• -Autocorrelations(positive,negative)
• --Those with higher mean intakes have greater
• within-person variation.---transform data
• --possible nonindependence of intake on
  consecutive days --- sampling days at random
  intervals.

18
II.  Number of days necessary to estimate true
intake

• Calculation of number of days
• (the degree of accuracy, the variability of the
  studied nutrient)
•  
• n(ZaCV?/D0)2
• n the number of days needed per person
• Za the normal deviate for the percentage of
  times the measured
• value should be within a specified
  limit
• CV? the within-person coefficient of
  variation
• (SD/Mean)100
• D0 the specified limit (as a percentage
  of long-term true intake)
•  
• (table 3-6CV)may vary among different
  populations

19

• (example)
• ---to calculate the number of days needed to
  estimate a persons cholesterol intake to within
  20 of their true mean 95 of the time
• n(1.9662/20)2 37 days
•   -----beyond practical possibilities
• (Table 3-7)
• CVs are different for various nutrients

20
(No Transcript)
21
(No Transcript)
22
III. Implications for developing countries

• More homogeneousdeserves examination(foods
  limitedincome differencesincrease between
  person variation)
•  
• Within developing countries, lower economic
  status has been shown to be associated with
  higher within-person variation.
•  
• Less able to adapt to temporatory short falls in
  income. Influence of season, preservation and
  transportation---provide excellent opportunities
  for epidemiologic studies of the relation between
  diet and health.---measure the important
  components of variation, such as seasonal effect)
   

23
(No Transcript)
24
?????

• ????(??????)
• ????(???????)

25
????(within-person)??????????

• ????(missclassification)
• ?????????????????
• ?????????
• ??????????
• ????????,????????,????
• ???????,???????

26
????

• ??????,???????
• ????????,??????????
• ?????,
• ????????

27
IV. Effects of random within person variation
on measures of associations in epidemiologic
studies

• Reducing the strength of association
• correlation coefficients (example0.620.28),
• regression coefficients (example0.120.01),
• relative risks (example2.361.51)

28
Attenuation of the relations
Physiological variables
R0.96
R0.65








Dietary intake level
29
(No Transcript)
30
(No Transcript)
31

• Methods to estimate true correlation
•  
• 1. To use large number of replicate measurements
  for each individual and use the average.
• ---ex within 10 of the true correlation14
  repeats
• 2. To make a small number of repeated
  measurements per subjects and use knowledge of
  the within-person variance (see Ch12).

32
Summary

• 1. Day to day variation in nutrient intake has
  consistently proved to be large.
• 2. The magnitude of variation varies according
  to nutrient.
• 3. Measurements of dietary intake based on a
  single or small number of 24-hour recalls per
  subject may provide a reasonable estimate of the
  mean for a group, but the SD will be
  overestimated.
• 4. Measurements of association (such as
  correlation and regression coefficients and
  relative risks) are weakened.

33
????/??????

• Page 107
• Chapter 5

34
??????

• ?????
• Duplicate portion technique, aliquote sampling
  technique, equivalent composite
• ?????
• ???precise weighing, weighted inventory, intake
  recorded in household measures
• ???????,????,????

35
??????

• Daily consumption methods
• ?????
• ?????precise weighing, weighed inventory,
  recorded in household measures?
• ?????duplicate portion technique, aliquot
  sampling technique, equivalent composite
• Long-term dietary pattern
• ?????
• ????
• ??????
• ???????

36
?????????

• ?????
• ?????,???????/24????
• ?????
• ?????
• ?????
• ????????24-?????
• ?????????????
• ????
• ?????24-?????????????
• ?????
• 24-?????
• ???? p133, section 6.6