Nutrition research methodologies
Week 2 Dietary assessment
Problems and aims of dietary assessment
Because of an measurement error an association between diet and a disease can not be found.
Why is this topic important for you?
- Assessment of dietary intake is essential in nutrition research
- Awareness of errors allows improvement, correction, better interpretation
- Awareness of characteristics of new technologies
- For dieticians: learn about application in research
The purpose of a dietary assessment can be different, while it can be about research or patient care.
Research can be needed in experimental and clinical studies, surveillance and monitoring studies and
epidemiological studies. Patient care is needed by dietary advice.
Diet is complex to measure, while there are some problems during the measurements;
- Way of preparation
- Novel foods and supplements
- Variation in intake (some products are daily induced, some not)
- Different age groups
- Characteristics of persons (overweight people tend to underestimate their intake)
- Estimation of portion sizes
- Period is not always clear (short term or long term)
For example what did you eat yesterday evening?
- Is it easy or difficult to report this?
- Can you identify all ingredients?
- Do you know all amounts within 5 grams accuracy?
- Was this meal representative for your evening meal?
- Did you consume ingredients that would lead to social desirable answering?
o Overreporting of healthy food and underreporting the unhealthy food
Assessment of dietary intake
By different dietary assessment methods
- To identify the type of foods that are consumed
- To define the amounts of the consumed food
- Sometimes times or location where/when the food is taken
You can use a food composition table or chemical analysis to convert this into intake of energy and
nutrients.
Essentials for each dietary assessment method
- Reference period
o E.g. previous month, today (depending on your research question)
- Portion sizes
o Weighing foods, household measures or photos
, - Food composition table
o Local table? Recent table?
o Are there specific components available in your food composition table?
Available methods: traditional techniques
A dietary assessment can be done via traditional methods like self reports, but you can also use
markers of exposure.
Information Self-reports Biomarker
Food pattern Yes No
Food components Multi Single
Information Subjective, direct Objective, indirect
Quality control Difficult Easier
Collecting data Time consuming Invasive
Self-reports: different methods
- 24 hour dietary recall (short term)
- Food record or food diary (short term)
o Observation
o Duplicate portion
- Food frequency questionnaire (long term)
- Diet history (long term)
It can be written, oral or technology-based.
24 hour dietary recall
- Principle
o Recall of past 24 hours (retrospective method)
o Interview
o Telephone or face-to-face
- Application
o Monitoring studies
o Easy for the respondent
o Comparison between cultures
o Less errors than food frequency questionnaire
- Disadvantages
o Day-to-day variation
o Response errors
Food record
- Principle
o Record current intake after the food is consumed (prospective method)
o Amount and time eaten
- Application
o Experiments
o Small studies
o Awareness of intake
- Disadvantages
o Become aware of their intake → change of intake (reactivity)
, o Burdensome (weighing)
Dietary history method
- Principle
o Interview
o Habitual intake (long term)
o Meal-based, much detail
- Different applications
o More or less extended (couple of hours)
o Research or dietary advice
- Disadvantage
o Burdensome
o Experienced interviewers needed
Food frequency questionnaire
- Principle
o Frequency or habitual consumption; few nutrients or comprehensive list (long term)
- Application
o Easy
o In large epidemiological studies
- Disadvantages
o Culturally-based
o Difficult to recall past intake
o No information on single foods
Screeners
- Kind of food frequency questionnaires, but short questionnaires tailored for qualitative
assessment of diet
o Only information on one or a couple of nutrients like fruits and vegetables, energy-%
from fat, dietary patterns (e.g. Mediterranean Diet Adherence)
- Needed when very limited room for questions on diet
- Useful in situations that do not require assessment of total diet
- Estimates of intake are not as accurate as those from more detailed methods
- Easy method for clinical practice
Web-based: example of short FFQ
De eetscore is developed by Wageningen University
- Estimate the Dutch Healthy Diet Index
- Assesses adherence to the Dutch dietary guidelines
- Implementation for use by dieticians, family doctors, etc.
- Valid to identify risk groups
- www.eetscore.nl
Available methods: technology-based instruments
In the last couple of years there are a lot of different tools developed. There are different wearable
electronic devices which contain sensors, data processing and storage elements, i.e. miniature
cameras and microphones.
There are three types of tools:
, - 15 smartphone apps
- 2 wearables; eButton and MSSenseCam
- 26 PC or web-based tools
Plus, there are app/wearable combinations:
- FoodNow, Australia – wearable armband for energy expenditure
- Diabetics Diary, Norway – android smartphone paired with Pebble smartwatch
Other developments are sensors and scan technology, like a bite counter (for example with a
chewing sensor).
Different aspects of methods can be improved with technology
- Food identification e.g. automatic based on photos
- Estimates of portion sizes
- Calculating composition
- More objective assessment
- Capturing other determinants such as time and location
- Feasibility for researcher and respondent
Some challenges for new technologies are:
1. Data entry
o 79% of tools relied on self-report or self-entry
o 91% used text entry & 33% used digital images to help identify foods
▪ Integrated tools for diet and health is still lacking
o Photography:
▪ Most semi-automatic, still need text
▪ Fiducial marker needed
▪ Before and after photo
▪ Angles important
▪ Challenges with recipe dishes – segmentation needed
▪ Smartphone or wearable
2. Food identification (what is it?)
• Nearly all (91%) relied on manual identification of foods
• Only 65% had integrated databases for estimating energy of nutrients
3. Food description (details)
o Innovations: food image identification technologies are improving, as are quantity
estimations, but cannot replace some self-entry or checking
o Most challenging to estimate volume from 2D image without depth information
o Novel classification strategies for image recognition:
▪ Traditional machine learning using handcrafted features
• Visual features extracted from food image, used to train prediction
model
▪ Deep learning-based approach
• A large number of connected layers can learn features (e.g. colour,
texture or shape), final layer for classification
4. Food composition (databases)
a. Different styles:
i. National databases – limited
Week 2 Dietary assessment
Problems and aims of dietary assessment
Because of an measurement error an association between diet and a disease can not be found.
Why is this topic important for you?
- Assessment of dietary intake is essential in nutrition research
- Awareness of errors allows improvement, correction, better interpretation
- Awareness of characteristics of new technologies
- For dieticians: learn about application in research
The purpose of a dietary assessment can be different, while it can be about research or patient care.
Research can be needed in experimental and clinical studies, surveillance and monitoring studies and
epidemiological studies. Patient care is needed by dietary advice.
Diet is complex to measure, while there are some problems during the measurements;
- Way of preparation
- Novel foods and supplements
- Variation in intake (some products are daily induced, some not)
- Different age groups
- Characteristics of persons (overweight people tend to underestimate their intake)
- Estimation of portion sizes
- Period is not always clear (short term or long term)
For example what did you eat yesterday evening?
- Is it easy or difficult to report this?
- Can you identify all ingredients?
- Do you know all amounts within 5 grams accuracy?
- Was this meal representative for your evening meal?
- Did you consume ingredients that would lead to social desirable answering?
o Overreporting of healthy food and underreporting the unhealthy food
Assessment of dietary intake
By different dietary assessment methods
- To identify the type of foods that are consumed
- To define the amounts of the consumed food
- Sometimes times or location where/when the food is taken
You can use a food composition table or chemical analysis to convert this into intake of energy and
nutrients.
Essentials for each dietary assessment method
- Reference period
o E.g. previous month, today (depending on your research question)
- Portion sizes
o Weighing foods, household measures or photos
, - Food composition table
o Local table? Recent table?
o Are there specific components available in your food composition table?
Available methods: traditional techniques
A dietary assessment can be done via traditional methods like self reports, but you can also use
markers of exposure.
Information Self-reports Biomarker
Food pattern Yes No
Food components Multi Single
Information Subjective, direct Objective, indirect
Quality control Difficult Easier
Collecting data Time consuming Invasive
Self-reports: different methods
- 24 hour dietary recall (short term)
- Food record or food diary (short term)
o Observation
o Duplicate portion
- Food frequency questionnaire (long term)
- Diet history (long term)
It can be written, oral or technology-based.
24 hour dietary recall
- Principle
o Recall of past 24 hours (retrospective method)
o Interview
o Telephone or face-to-face
- Application
o Monitoring studies
o Easy for the respondent
o Comparison between cultures
o Less errors than food frequency questionnaire
- Disadvantages
o Day-to-day variation
o Response errors
Food record
- Principle
o Record current intake after the food is consumed (prospective method)
o Amount and time eaten
- Application
o Experiments
o Small studies
o Awareness of intake
- Disadvantages
o Become aware of their intake → change of intake (reactivity)
, o Burdensome (weighing)
Dietary history method
- Principle
o Interview
o Habitual intake (long term)
o Meal-based, much detail
- Different applications
o More or less extended (couple of hours)
o Research or dietary advice
- Disadvantage
o Burdensome
o Experienced interviewers needed
Food frequency questionnaire
- Principle
o Frequency or habitual consumption; few nutrients or comprehensive list (long term)
- Application
o Easy
o In large epidemiological studies
- Disadvantages
o Culturally-based
o Difficult to recall past intake
o No information on single foods
Screeners
- Kind of food frequency questionnaires, but short questionnaires tailored for qualitative
assessment of diet
o Only information on one or a couple of nutrients like fruits and vegetables, energy-%
from fat, dietary patterns (e.g. Mediterranean Diet Adherence)
- Needed when very limited room for questions on diet
- Useful in situations that do not require assessment of total diet
- Estimates of intake are not as accurate as those from more detailed methods
- Easy method for clinical practice
Web-based: example of short FFQ
De eetscore is developed by Wageningen University
- Estimate the Dutch Healthy Diet Index
- Assesses adherence to the Dutch dietary guidelines
- Implementation for use by dieticians, family doctors, etc.
- Valid to identify risk groups
- www.eetscore.nl
Available methods: technology-based instruments
In the last couple of years there are a lot of different tools developed. There are different wearable
electronic devices which contain sensors, data processing and storage elements, i.e. miniature
cameras and microphones.
There are three types of tools:
, - 15 smartphone apps
- 2 wearables; eButton and MSSenseCam
- 26 PC or web-based tools
Plus, there are app/wearable combinations:
- FoodNow, Australia – wearable armband for energy expenditure
- Diabetics Diary, Norway – android smartphone paired with Pebble smartwatch
Other developments are sensors and scan technology, like a bite counter (for example with a
chewing sensor).
Different aspects of methods can be improved with technology
- Food identification e.g. automatic based on photos
- Estimates of portion sizes
- Calculating composition
- More objective assessment
- Capturing other determinants such as time and location
- Feasibility for researcher and respondent
Some challenges for new technologies are:
1. Data entry
o 79% of tools relied on self-report or self-entry
o 91% used text entry & 33% used digital images to help identify foods
▪ Integrated tools for diet and health is still lacking
o Photography:
▪ Most semi-automatic, still need text
▪ Fiducial marker needed
▪ Before and after photo
▪ Angles important
▪ Challenges with recipe dishes – segmentation needed
▪ Smartphone or wearable
2. Food identification (what is it?)
• Nearly all (91%) relied on manual identification of foods
• Only 65% had integrated databases for estimating energy of nutrients
3. Food description (details)
o Innovations: food image identification technologies are improving, as are quantity
estimations, but cannot replace some self-entry or checking
o Most challenging to estimate volume from 2D image without depth information
o Novel classification strategies for image recognition:
▪ Traditional machine learning using handcrafted features
• Visual features extracted from food image, used to train prediction
model
▪ Deep learning-based approach
• A large number of connected layers can learn features (e.g. colour,
texture or shape), final layer for classification
4. Food composition (databases)
a. Different styles:
i. National databases – limited
