Summary – Nutrition Research Methodologies – Pre-MNH – P1
In order to prepare for the exam, learn the summary below, read notes, re-make the weekly
assignments and learn the answers, read important topics of the book and make relevant exercises,
re-watch the lectures and Q&A sessions and practise all E-modules.
Week 1 (check notes and information published on Brightspace)
- Three different search strategies:
1. Simple search = quick and dirty
2. Systematic search = building query
3. Following a thread = snowballing
- PICO = Patient (or population) Intervention Control (or comparison) and Outcome.
Week 2
- Aims of dietary assessment = research or patient care
- Research aim = experimental and clinical studies, surveillance and monitoring and
epidemiology studies.
- Problems during measuring dietary intake =
1. Preparation
2. Novel foods and supplements
3. Variation intake
4. Age groups
5. Characteristics (high weighted people tend to underestimate portions)
6. Estimation portion sizes
7. Period not always clear
- Essentials for each dietary assessment = reference period, portion sizes and food composition
table.
- Food pattern = can be measured best with self-reports (not a biomarker)
- Multiple food components = can be measured best in self-reports because biomarkers will
only give you a single component.
- Information in self-reports is subjective and direct and in biomarkers it is more objective and
indirect.
- Quality control = can be measured best in a biomarker (easier).
- Self-report methods:
1. 24h dietary recall (short term)
2. Food record diary (short term)
3. FFQ (long term)
4. Diet history (long term)
- 24h dietary recall = interview about recording everything of the past 24 hours. It is easy and
has less errors than an FFQ. It can be long term if it is repeated. It is pre-coded.
- Food record diary = reporting everything immediately after it is consumed. It is a prospective
method. The problem can be estimating portion sizes but isn’t necessary because the food is
reported direct after intake. People become aware of intake, that can result in socially
desirable reports.
- FFQ = Food Frequency Questionnaire = used in large scale epidemiology studies and like a
questionnaire about usual or long-term intake. It is pre-coded.
- FFQ screeners = short questionnaire for more qualitative assessment of diet only provide info
on a few foods or only one or two nutrients.
- Dietary history = interview but long term or usual intake and can take in 2 hours of
questionnaires.
- Steps to choose a method
1. What is the aim of the assessment?
2. What info do you need?
, STEP 1: Aim STEP 2: Type of info
Mean intake population Group mean One day record/recall
% <Cut-off value Group mean distribution Repeated record/recall
Association intake-health Ranking or classification FFQ/diet history
Intake individual Absolute intake Multiple records/recalls and
if possible, a biomarker.
% <Cut-off value = prevalence below another value (distribution)
Group mean = when you are only interested in an average of a population without references
or comparisons (one day record or 24h recall).
Group mean distribution = when there is a reference/comparison mentioned and you want to
compare it with an average or population group (prevalence). You need here a minimum of
two days (repeated record/recall)
3. What is the target population?
24h recalls, FFQ, food records and dietary history for healthy elderly and observed
record, picture sort and biomarkers for disabled elderly.
4. What is the reference period?
Depends if the aim is habitual or actual intake? Memory problems?
🡪 Within person-variation-coefficient (CVw) = SD/mean x 100%
You need a minimum of two days information to determine the within person
variation.
Z CVw 2
🡪 Number of days needed (n) = ( )
Do
Do = defined limitation (determine yourself, you can say I want my outcome
between 20% of the true intake.
Zχ = Z-value, for normal distribution is that 1.96.
The more day-to-day variation is present, the more days you want to
measure.
5. Methods to compare
6. How much time/finances/experience are available?
- Response error = are mostly present when the respondent has to report. But is not that much
in duplicate portion.
- Coding error = if the researcher has to code everything that is reported, there is more chance
on making these errors. But when the method is pre-coded the error is less likely because it
doesn’t depend on one person.
- Portion size errors = present when they have to be estimated, which is not for a weighed food
record and duplicate portion. Also, when the method is observed it is less likely to make these
errors.
- Errors in FCT = are present when you use these tables. So, not in duplicate portion and when
observed, the observer knows better what the composition of the food is. He can check it on
the package.
- Duplicate portion = a dietary assessment method when people don’t have to report but where
everything is put into a basket and it will be calculated and chemically analysed (for example
used in an FCT).
Random errors:
- Make the assessment less precise.
In order to prepare for the exam, learn the summary below, read notes, re-make the weekly
assignments and learn the answers, read important topics of the book and make relevant exercises,
re-watch the lectures and Q&A sessions and practise all E-modules.
Week 1 (check notes and information published on Brightspace)
- Three different search strategies:
1. Simple search = quick and dirty
2. Systematic search = building query
3. Following a thread = snowballing
- PICO = Patient (or population) Intervention Control (or comparison) and Outcome.
Week 2
- Aims of dietary assessment = research or patient care
- Research aim = experimental and clinical studies, surveillance and monitoring and
epidemiology studies.
- Problems during measuring dietary intake =
1. Preparation
2. Novel foods and supplements
3. Variation intake
4. Age groups
5. Characteristics (high weighted people tend to underestimate portions)
6. Estimation portion sizes
7. Period not always clear
- Essentials for each dietary assessment = reference period, portion sizes and food composition
table.
- Food pattern = can be measured best with self-reports (not a biomarker)
- Multiple food components = can be measured best in self-reports because biomarkers will
only give you a single component.
- Information in self-reports is subjective and direct and in biomarkers it is more objective and
indirect.
- Quality control = can be measured best in a biomarker (easier).
- Self-report methods:
1. 24h dietary recall (short term)
2. Food record diary (short term)
3. FFQ (long term)
4. Diet history (long term)
- 24h dietary recall = interview about recording everything of the past 24 hours. It is easy and
has less errors than an FFQ. It can be long term if it is repeated. It is pre-coded.
- Food record diary = reporting everything immediately after it is consumed. It is a prospective
method. The problem can be estimating portion sizes but isn’t necessary because the food is
reported direct after intake. People become aware of intake, that can result in socially
desirable reports.
- FFQ = Food Frequency Questionnaire = used in large scale epidemiology studies and like a
questionnaire about usual or long-term intake. It is pre-coded.
- FFQ screeners = short questionnaire for more qualitative assessment of diet only provide info
on a few foods or only one or two nutrients.
- Dietary history = interview but long term or usual intake and can take in 2 hours of
questionnaires.
- Steps to choose a method
1. What is the aim of the assessment?
2. What info do you need?
, STEP 1: Aim STEP 2: Type of info
Mean intake population Group mean One day record/recall
% <Cut-off value Group mean distribution Repeated record/recall
Association intake-health Ranking or classification FFQ/diet history
Intake individual Absolute intake Multiple records/recalls and
if possible, a biomarker.
% <Cut-off value = prevalence below another value (distribution)
Group mean = when you are only interested in an average of a population without references
or comparisons (one day record or 24h recall).
Group mean distribution = when there is a reference/comparison mentioned and you want to
compare it with an average or population group (prevalence). You need here a minimum of
two days (repeated record/recall)
3. What is the target population?
24h recalls, FFQ, food records and dietary history for healthy elderly and observed
record, picture sort and biomarkers for disabled elderly.
4. What is the reference period?
Depends if the aim is habitual or actual intake? Memory problems?
🡪 Within person-variation-coefficient (CVw) = SD/mean x 100%
You need a minimum of two days information to determine the within person
variation.
Z CVw 2
🡪 Number of days needed (n) = ( )
Do
Do = defined limitation (determine yourself, you can say I want my outcome
between 20% of the true intake.
Zχ = Z-value, for normal distribution is that 1.96.
The more day-to-day variation is present, the more days you want to
measure.
5. Methods to compare
6. How much time/finances/experience are available?
- Response error = are mostly present when the respondent has to report. But is not that much
in duplicate portion.
- Coding error = if the researcher has to code everything that is reported, there is more chance
on making these errors. But when the method is pre-coded the error is less likely because it
doesn’t depend on one person.
- Portion size errors = present when they have to be estimated, which is not for a weighed food
record and duplicate portion. Also, when the method is observed it is less likely to make these
errors.
- Errors in FCT = are present when you use these tables. So, not in duplicate portion and when
observed, the observer knows better what the composition of the food is. He can check it on
the package.
- Duplicate portion = a dietary assessment method when people don’t have to report but where
everything is put into a basket and it will be calculated and chemically analysed (for example
used in an FCT).
Random errors:
- Make the assessment less precise.
