Week 3: LEC 4&5 Flashcards
(22 cards)
Using Food and Nutrient Databases
LEC 4
Food Analysis Programs
are…
NOTE: the prog is only as goof as the nutrient database it links to
What are nutrient databases?
->
Who makes nutrient databases?
->
-
-
->
Where to find these nutrient databases?
- Online searchable databases / publications / international network of food data systems
are… computer programs that link your dietary intake with nutrient databases
- ASA24 / Cronometer etc
-> List of foods with their nutrient compositions
-> Usually government-funded
- USDA
- Health CAN
-> Some university-based databases
- widely used for research collection
Canadian Nutrient File (CNF)
-
-
-
* can always access online *
= Derived from USDA Nutrient Database for Standard Reference:
- with the addition of info based on the chemical analyses of some Canadian foods and food products,
- with adjustments for differences in the cuts of meat and their fat-to-lean ratio, and for Canadian enrichment and fortification practices,
- and with exclusion of foods included in the USDA database but that are known not to be on the Canadian market
-
-
-
-
-
- Proximate components: moisture (water), protein, total lipid, CHO, ash
- Other CHO: fibre, starch, glucose, fructose etc.
- Minerals / Vitamins
- Amino Acids
- Fatty Acids
- Other components: caffeine & theobromine
Number of Observations
> To make one entry in the database, you:
1
2
3
4
1 Choose a food to analyse
2 Gather different samples of the food item to analyse -
3 Do chemical analyses for EACH nutrient in the food item
a) water, CHO, Fat, FA, fibre, vitamins, minerals, caffeine etc.
4 Calculate the weighted mean for each nutrient
a) the greater the # of observations is, the more accurate and the more representative is the data (& less variation)
Standard Error
- Indicates…
-
- How big is the standard error for your food?
-
Do standard error and number of observations change in same or opposite directions?
[If no standard error or number of samples is included, the values have been imputed or calculated from another form of the food, from a similar food, are based on calculated recipe or there was no change from USDA]
Why are there differences between CAN and USA data?
- Indicates variability between samples
- measures dispersion of values around the mean
? - Micros have lots of variation b/c of sensitivity to heat, light, oxidation etc
? as standard error decreases / observations increase [inverse]
?
Food and Nutrient databases around the world
-> Why is it important to have local data?
-
-
-
-
- Different food available in the food supply
- Differences in food fortification programs
- Differences in processing, preparation, or cooking
- Differences in soil nutrient characteristics
- levels of selenium, iodine, and zinc depend on trace element levels in soil
Limitations of food and nutrient databases
-> Indicates…
-> Many
-> Some foods are not…
-> Data for some nutrients are…
*** -> Intakes of most nutrients calculated from food composition data represent the maximum available to the body
- does not reflect the amount actually absorbed by the body **
-> …avg amounts of nutrients available based on food sources across the country
- soil characteristics (vegetables) or methods of raising cattle (meat products) can affect nutrient content
-> many foods in the database represent a generic product
- There’s no coke or pepsi, only colas / for brand specific nutrient data, it is usually best to look on the food packaging
- For mixed dishes, values are based on a ‘typical’ recipe and method of preparation
-> are not found in the database
ie. data for mixed dishes, fast foods, specific brands or ethnic foods may not be available
-> are inaccurate or not available -
ie. Vitamin K, trans fat, choline, sugar breakdown, vitamin E, biotin
Why don’t things add up?
Food energy (kcals)
-> The calorie value for food will often differ from that calculated by the standard conversion factors of 4kcal (CHO/Pro) & 9 kcal (fat) OR the food-specific conversion factors
WHY?
Why don’t things add up?
Totals for macronutrients
-> Different components are measured differently
EX:
- Some CHO components may not be measured
Ex.
?
-> EX. Fatty Acids are analyzed by different methods than total fat - their sum will be close enough but not identical to total fat
- Same is true for protein and amino acids
WHY? b/c conversion factors = utilizable
& food-specific conversion factors = what you’re actually metabolizing
EX: Fatty Acids are analyzed by different methods than total fat, same for PRO and AA
Ex. Dietary Fibre
VITAMIN A
3 values for Vit A in the CNF:
1
2
3
1 RAE =
-> Other units you may encounter:
- International Units (IU) (used in the USA)
- 1 IU =
1 Beta-carotene (mcg) - inactive forms which can be converted to retinol in the body
2 Retinol (mcg) - the active form
3 Retinol Activity Equivalents (RAE) (mcg)
1 RAE = mcg retinol + mcg beta-carotene/12 + mcg other carotenoids/24
[/12 is due to 12mcg of beta-car = 1 mcg retinol AND /24 due to 24mcg of other for 1 mcg retinol]
- 1 IU = 0.3 mcg Retinol
VITAMIN D
-> Values of Vitamin D in CNF account for 2 forms of vitamin D in foods:
1
2
?3?
CNF reports Vit D in ____
{CNF as well as USA and many analysis programs use International Units}
1 mcg =
VITAMIN E
-> There are several isomers of Vit E, but the only one with significant activity in the body is ___
-> Units for Vit E in the CNF are ___
1 Vitamin D3 (cholcalciferol)
2 Vitamin D2 (ergocalciferol)
?3? 25 hydroxyl-D3 = not accounted for in CNF
___ in mcg
1 mcg = 40 IU Vitamin D
___ alpha-tocopherol (only active form)
are mg of alpha-tocopherol
NIACIN
-> Niacin is expressed in the CNF as:
-
-
2 ways of calculating NE:
- If values for preformed niacin and tryptophan are available:
- If Tryptophan value is not available, it is imputed to be…
EXAMPLE - SLIDE 35**
- Preformed niacinamide (mg) - present in food
- Niacin equivalents (NE): includes preformed niacin from foods + niacin formed in the body from tryptophan
(the conversion factor is 60mg tryptophan to 1 mg niacin) - : NE = (tryptophan x 1000/60) (mg) + preformed niacin (mg)
- … to be 1.1% of total protein:
NE = (0.011 x protein) x 1000/60 (mg) + preformed niacin (mg)
EX: Tryp is in g so x1000 for mg (which is in eqn)
ex answer: 0.043 x 1000/60 + 1.221mg = 1.937 mg = 1.94 mg rounded
FOLATE
2 forms in food:
1
2
-
-
-
1 DFE =
1 Naturally occurring folate in food
2 Synthetic form is folic acid (more active than natural food folate)
2 x more bioavail than folate in fasting person
1.7 x more bioavail than folate if taken with food
- Food folate or naturally-occurring folate (mcg)
- Folic Acid (Synthetic form) (mcg)
- Total folacin (mcg) - this is the sum of food folate and folic acid in mcg (used on CAN food labels)
[doesn’t take different bioavailability into account] - Dietary Folate Equivalents (DFE) (mcg)
1 DFE = (mcg folic acid x 1.7) + mcg food folate
[1.7 is the factors of bioavailability]
Accuracy, Precision and Validity of Dietary Assessment
LEC 5
How do you know if your assessment tool is of high quality?
VALIDITY =
EX
How well does a 24 hour recall measure yesterday’s intake?
Absolute Validity would look like:
Relative Validity:
- it is too pricey to feed people and ask them what they ate yesterday to validate the 24 hour recall
INSTEAD…
Exs of testing relative validity SLIDE 7**
= describes the degree to which a dietary method measures what it is intended to measure
EX - 24 hour recall - is it measuring food in the last 24 hours or?
: [you supply the food] & [ask them what they ate]
You would compare what they actually ate and they said they ate
Instead… relative validity is examined by comparing the “test” dietary method to a “reference” standard dietary method performed on the same subjects
Considerations when testing relative validity
-> The 2 methods should…
-
-> Need to carefully sequence and space administration of each method:
- & why
-
- Can you assess the dietary intake on the same day with 2 methods?
*** need to account for gender and age:
-> Should measure the same thing (same objective ie. usual intake, past intake, current intake) and time frame (ie. day/ month / year)
- Usually administer the test method first: WHY? if you start with control (reference method) then could impact how participant behaves during the test method
- Not too long or short btw admin of each method
? burden to participant and may influence each method so they are not truthful
*** : (sex) respond differently due to physiological biology
(age) under 18 interact differently with assessment tools & 65+ accuracy & over reporting
- Keep in mind *
-> Poor agreement btw 2 methods
-> Good agreement btw the test and the reference methods…
A new way of assessing validity - using biomarkers
-> Using a 24 hour…. (1)
-
-
-
-
-> Using 24 hour… (2)
-> … at least one of the dietary methods is invalid
-> … not necessarily indicate validity of the test method
- may indicate a similar degree of errors in both methods, so agreement on outcomes but same errors - be critical
-> (1) Using a 24 hour urinary nitrogen excretion to validate 24 hour recorded/recalled protein intake
- Stable N balance [ + N balance = too much N intake // - N balance = excreting more than intaking ]
- Extra-renal losses
- Complete 24 hour urine collections
- Repeat 24 hour urine collections (within-subject variation - difference in water intake day-to-day)
-> (2) Using a 24 hour urinary sodium excretion to validate reported sodium intake
- within-subject variation [suggested is 15- 24 hour sodium excretion days (not consecutive) to get an accurate picture]
Characteristics of an ideal biomarker
->
->
-
VS
-
->
- short term biomarkers
- med
- long term biomarkers
->
-> Valid and Reliable
-> High sensitivity and specificity
- Sensitivity = ability to discriminate btw various levels of intake
VS
- Specificity = differences in levels of the biomarker are specific to intake of the nutrient of interest
-> Temporal relations with the dietary intake
- Serum, plasma and urine (Vit C; B vitamins): short term biomarkers
- Erythrocytes (fatty acids; folate; selenium): medium-term biomarkers
- Hair, fingernails, toenails (selenium) and adipose tissue (fatty acids): long-term biomarkers
-> Inexpensive to collect and analyze / non-invasive and only need small amounts
RELIABILITY
AKA…
=
-
-
-
-
-
AKA… Reproducibility
= Reproducibility is when a dietary assessment method gives very similar results when used repeatedly in the same situation (or similar situations)
- Variety in an indiv’s food choices
- Day of the week
- Season
- Holidays and special occasions
- Appetite
- Food sources of the nutrient of interest (some nutrients have greater daily variation than other (ie. Vit A))
What are Accuracy and Precision?
(aka reliability)
A =
P =
Ex’s:
A & P are affected by 2 types of error:
1 =
2 =
A = how close a measured value is tot he actual value
P = how close the measured values are to each other
: A+P = measuring height [little sources of error]
: A = measuring BP [ A - close to real value / not P cause so many factors ]
: P = measuring weight on an incorrectly calibrated scale [Precise in wrong way]
: neither = self reported body weight
1 Systematic error = error that is not determined by chance but is introduced by an inaccuracy (as of observation or measurement) inherent in the system
ie. omitting supplement use for all respondents [product of your design - maybe didn’t train interviewers -prevented by better design]
{P not A}
2 Random error
ie. Recording a person’s weight incorrectly
{A not P}
Sys and Random together = neither A or P
Sources of systematic error (Bias)
- Non-responsive bias =
- Respondent bias =
- Interviewer bias =
- Omitting supplement use
Prevented?
N-R = sample is not representative of the population of interest
R = systematic underreporting or over reporting
I = probing for information to varying degrees, omitting certain questions, recording responses incorrectly
? yes by training interviewers, trust participants so no under/over reporting / better design of study
-
-
-
-
How can it be minimized?
- Respondent memory lapses - omitting or adding foods
- Incorrect estimation of portion size
- Coding errors - can happen when converting portion sizes or inputting food items into nutrients analysis programs
- Mistakes in the handling of mixed dishes
- Inaccurate food composition values
? by having a larger sample size
Underreporting food intake
- is a common problem, evident in:
- problem may be due to:
-
-
-
-
ensure a safe space for participants to avoid this
No simple solution for underreporting
-> You cannot just multiply energy or nutrient intakes by a correct factor (for all indiv and all food items) for 2 main reasons:
-
-
So what do you do?
: in 24 hour recalls, food records, and FFQs
due to - eating less than usual on days when food intake is recorded
- inaccurate recall, portion size estimate
- omitting or forgetting certain types of foods
- social desirability!
- probability of underreporting is not the same among individuals
- some foods are more likely to be underreported than others
? Build a good space for participants, good design, good tools at use
Improving Validity and Reliability
-> To obtain a valid estimate of usual nutrient intake:
-
-
-> To obtain reproducible estimate of usual nutrient intake:
-
-
** Review her in class example of research article on dietary adequacy of Inuit - SLIDE 29 **
- anticipate and prevent systematic error
- take steps to increase reliability
- increase the # of days of intake data (no more than 7 days)
- increase the # of ppl in the sample
