Nutrition Research Methodologies (HNH24306)
INHOUDSOPGAVE
WEEK I: INFORMATION LITERACY (2/09) ..................................................................................2
LEARNING GOAL..................................................................................................................2
LECTURE .............................................................................................................................2
LITERATURE: COURSE BACKGROUND INFORMATION: TWO MAIN TYPES OF HUMAN NUTRITION RESEARCH ...2
A. STUDY DESIGN: POPULATION-BASED STUDIES .............................................................2
B. STUDY DESIGN: INTERVENTION STUDIES .....................................................................4
WEEK II: ANTHROPOMETRY & BODY COMPOSITION (9/09) .........................................................9
LEARNING GOAL..................................................................................................................9
LECTURE .............................................................................................................................9
LITERATURE: CHAPTER II – METHODS FOR ASSESSING NUTRITIONAL STATUS & BODY COMPOSITION ...........9
WEEK III: ENERGY EXPENDITURE & PHYSICAL ACTIVITY (16/09) ................................................20
LEARNING GOAL................................................................................................................20
LECTURE ...........................................................................................................................20
LITERATURE: CHAPTER XII – ENERGY EXPENDITURE & INTAKE METHODS............................................20
LITERATURE: PAPER STRATH ET AL. (2013) ...............................................................................23
WEEK IV: DIETARY ASSESSMENT (23/09) ...............................................................................24
LEARNING GOAL................................................................................................................24
LECTURE: CHAPTER IV – METHODS TO DETERMINE DIETARY INTAKE ..................................................24
LITERATURE: PAPER (MURPHY ET AL) ON REFERENCE VALUES .........................................................34
WEEK V: BIOMARKERS & TRACERS (30/09) ...........................................................................35
LEARNING GOAL................................................................................................................35
LECTURE ...........................................................................................................................35
LITERATURE: CHAPTER VI – BIOMARKERS OF INTAKE....................................................................35
LITERATURE: CHAPTER XVII – STABLE ISOTOPES IN NUTRITION RESEARCH .........................................45
WEEK VI: COMPARING TWO METHODS (7/10) ........................................................................55
LEARNING GOAL................................................................................................................55
LECTURE ...........................................................................................................................55
LITERATURE: …. .................................................................................................................55
SHORT SUMMARY ...........................................................................................................56
1
,WEEK I: INFORMATION LITERACY (2/09)
LEARNING GOAL
- locate & access high-quality scientific papers in the field of human nutrition
LECTURE
See intro PPT
See other word document exercises E-module
LITERATURE: COURSE BACKGROUND INFORMATION: TWO MAIN TYPES OF HUMAN NUTRITION
RESEARCH
A. STUDY DESIGN: POPULATION-BASED STUDIES
Introduction
Epidemiology = study of disease in populations
Types population-based studies:
- Ecological study
- Case-control study
- Cohort study
Key point: researchers has no control over exposure of interest (e.g. diet)
Purpose: useful for generation hypotheses & exploring associations between diet & health
outcomes
Limitation: cannot prove cause-and-effect relationships – only associations
Bias risk: more prone to bias compared to randomized controlled trials (RCTs)
- Selection bias = systematic differences between participant groups can lead to
cofounding results
- Non-response bias = non-random dropout of participants can affect results
- Non-attrition bias = dropping out of participants
- Measurement bias = includes recall bias & social desirability reporting bias, leading to
cofounding results
Cofounding variables:
- Provide alternative explanations for observed association
- Associated with both the dietary exposure & disease/health outcome but are not part of
causal pathway
2
,Cross-sectional studies
- Type of observation or descriptive study
- Provide snapshot of info of a population at one point in time
- Not possible to determine whether exposure & outcome are causally related
- A.k.a. ‘prevalence survey’, they estimate prevalence (rate) of disease in a population at
specific time (prevalence = number of cases of a disease in the population at a particular time)
Analysis of cross-sectional data
- Cross-sectional data can be analysed by using correlations between exposure &
outcomes
- Regression modelling helps with explore the influence of one continuous variable on
another, taking into account potential cofounders
Problem with cross-sectional studies
- Main limitation: exposure & disease are measured at same time, so not possible to say
which is cause & which is effect
Confounding
= variables associated with both exposure (e.g. diet) & outcome (e.g. disease) BUT is not part of
causal pathway
Impact: can influence perceived size, direction… or significance of an association between
dietary factors & health outcomes
Example: relationship between coffee consumption & reduced birth
weight may be confounded by smoking, as smoking is linked to both
coffee intake and lower birth weight
➔ So a confounder can provide an alternative explanation for an
apparent association
Controlling confounders: identify, measured & controlled trough study design (e.g. by matching
or restricting participants) or analysis (e.g. by stratification, restriction or regression adjustment)
Challenges: even after adjustment, some residual confounding remain & true dietary
associations might be masked due to less accurate dietary measurements
In RCTs: less likely to due randomization, but practical issues in nutritional studies (e.g.
compliance, uncontrolled intake, not willing to be randomized to a diet…) exist
Conclusion: due to the limitations of RCTs in nutrition research, results of observational studies
remain useful for exploring the role of nutritional exposures in the causation of disease, despite
confounding and other weaknesses.
Conclusion
Population-based observational studies: use epidemiological methods to explore diet and
health associations:
- Ecological & cross-sectional studies: useful for generating hypotheses but cannot test
causal relationships due to limitations.
- Case-control studies: helpful in studying diet-disease links but face challenges with
control selection and recalling past diet.
3
, - Cohort studies: allow more rigorous testing of diet-disease relationships.
➔ All these methods can be influenced by confounding factors, which are associated with both
the diet (exposure) and disease (outcome).
B. STUDY DESIGN: INTERVENTION STUDIES
1. Introduction
Observational studies Intervention studies
demonstrate association between Demonstrate case & effect by altering
nutrients/foods/diet & health outcomes nutrients/foods/diet in a controlled way &
effect on selected outcome is measured.
cannot prove causality prove cause & effect
e.g.: cross-sectional, cohort, and case- e.g. parallel group, cross-over, factorial
control studies design, cluster randomized design…
lower level of scientific evidence than higher level of scientific evidence than
observational studies observational studies
2. Intervention study types
2.1. Randomised controlled trials: Parallel & cross-over
Purpose
- Test (hypothesises about) cause-and-effect relationships by randomizing participants to
study groups
- In all controlled studies – use of a ‘control group’
o Receive a placebo or no intervention
o Outcomes compared to intervention group
o Without control group, no cause-and-effect statement can be made as other
factors may be responsible
→ Use a control group to compare outcomes & account for other variables (e.g. placebo
effect, seasonal changes…)
Key RCT designs
Parallel group studies
- Each participant receives only one intervention (e.g. low or high intake, nutrient A or B…)
- Comparisons between participants
- Suitable for long-term studies (but shorter over-all time frame) and cases where
returning to baseline is difficult or unethical (e.g., weight, bone density)
- Least suitable to outcomes that show large inter-participant variations
Cross-over studies
- Each participant receives all interventions in a specified order
o More interventions = higher drop-out risk
4
INHOUDSOPGAVE
WEEK I: INFORMATION LITERACY (2/09) ..................................................................................2
LEARNING GOAL..................................................................................................................2
LECTURE .............................................................................................................................2
LITERATURE: COURSE BACKGROUND INFORMATION: TWO MAIN TYPES OF HUMAN NUTRITION RESEARCH ...2
A. STUDY DESIGN: POPULATION-BASED STUDIES .............................................................2
B. STUDY DESIGN: INTERVENTION STUDIES .....................................................................4
WEEK II: ANTHROPOMETRY & BODY COMPOSITION (9/09) .........................................................9
LEARNING GOAL..................................................................................................................9
LECTURE .............................................................................................................................9
LITERATURE: CHAPTER II – METHODS FOR ASSESSING NUTRITIONAL STATUS & BODY COMPOSITION ...........9
WEEK III: ENERGY EXPENDITURE & PHYSICAL ACTIVITY (16/09) ................................................20
LEARNING GOAL................................................................................................................20
LECTURE ...........................................................................................................................20
LITERATURE: CHAPTER XII – ENERGY EXPENDITURE & INTAKE METHODS............................................20
LITERATURE: PAPER STRATH ET AL. (2013) ...............................................................................23
WEEK IV: DIETARY ASSESSMENT (23/09) ...............................................................................24
LEARNING GOAL................................................................................................................24
LECTURE: CHAPTER IV – METHODS TO DETERMINE DIETARY INTAKE ..................................................24
LITERATURE: PAPER (MURPHY ET AL) ON REFERENCE VALUES .........................................................34
WEEK V: BIOMARKERS & TRACERS (30/09) ...........................................................................35
LEARNING GOAL................................................................................................................35
LECTURE ...........................................................................................................................35
LITERATURE: CHAPTER VI – BIOMARKERS OF INTAKE....................................................................35
LITERATURE: CHAPTER XVII – STABLE ISOTOPES IN NUTRITION RESEARCH .........................................45
WEEK VI: COMPARING TWO METHODS (7/10) ........................................................................55
LEARNING GOAL................................................................................................................55
LECTURE ...........................................................................................................................55
LITERATURE: …. .................................................................................................................55
SHORT SUMMARY ...........................................................................................................56
1
,WEEK I: INFORMATION LITERACY (2/09)
LEARNING GOAL
- locate & access high-quality scientific papers in the field of human nutrition
LECTURE
See intro PPT
See other word document exercises E-module
LITERATURE: COURSE BACKGROUND INFORMATION: TWO MAIN TYPES OF HUMAN NUTRITION
RESEARCH
A. STUDY DESIGN: POPULATION-BASED STUDIES
Introduction
Epidemiology = study of disease in populations
Types population-based studies:
- Ecological study
- Case-control study
- Cohort study
Key point: researchers has no control over exposure of interest (e.g. diet)
Purpose: useful for generation hypotheses & exploring associations between diet & health
outcomes
Limitation: cannot prove cause-and-effect relationships – only associations
Bias risk: more prone to bias compared to randomized controlled trials (RCTs)
- Selection bias = systematic differences between participant groups can lead to
cofounding results
- Non-response bias = non-random dropout of participants can affect results
- Non-attrition bias = dropping out of participants
- Measurement bias = includes recall bias & social desirability reporting bias, leading to
cofounding results
Cofounding variables:
- Provide alternative explanations for observed association
- Associated with both the dietary exposure & disease/health outcome but are not part of
causal pathway
2
,Cross-sectional studies
- Type of observation or descriptive study
- Provide snapshot of info of a population at one point in time
- Not possible to determine whether exposure & outcome are causally related
- A.k.a. ‘prevalence survey’, they estimate prevalence (rate) of disease in a population at
specific time (prevalence = number of cases of a disease in the population at a particular time)
Analysis of cross-sectional data
- Cross-sectional data can be analysed by using correlations between exposure &
outcomes
- Regression modelling helps with explore the influence of one continuous variable on
another, taking into account potential cofounders
Problem with cross-sectional studies
- Main limitation: exposure & disease are measured at same time, so not possible to say
which is cause & which is effect
Confounding
= variables associated with both exposure (e.g. diet) & outcome (e.g. disease) BUT is not part of
causal pathway
Impact: can influence perceived size, direction… or significance of an association between
dietary factors & health outcomes
Example: relationship between coffee consumption & reduced birth
weight may be confounded by smoking, as smoking is linked to both
coffee intake and lower birth weight
➔ So a confounder can provide an alternative explanation for an
apparent association
Controlling confounders: identify, measured & controlled trough study design (e.g. by matching
or restricting participants) or analysis (e.g. by stratification, restriction or regression adjustment)
Challenges: even after adjustment, some residual confounding remain & true dietary
associations might be masked due to less accurate dietary measurements
In RCTs: less likely to due randomization, but practical issues in nutritional studies (e.g.
compliance, uncontrolled intake, not willing to be randomized to a diet…) exist
Conclusion: due to the limitations of RCTs in nutrition research, results of observational studies
remain useful for exploring the role of nutritional exposures in the causation of disease, despite
confounding and other weaknesses.
Conclusion
Population-based observational studies: use epidemiological methods to explore diet and
health associations:
- Ecological & cross-sectional studies: useful for generating hypotheses but cannot test
causal relationships due to limitations.
- Case-control studies: helpful in studying diet-disease links but face challenges with
control selection and recalling past diet.
3
, - Cohort studies: allow more rigorous testing of diet-disease relationships.
➔ All these methods can be influenced by confounding factors, which are associated with both
the diet (exposure) and disease (outcome).
B. STUDY DESIGN: INTERVENTION STUDIES
1. Introduction
Observational studies Intervention studies
demonstrate association between Demonstrate case & effect by altering
nutrients/foods/diet & health outcomes nutrients/foods/diet in a controlled way &
effect on selected outcome is measured.
cannot prove causality prove cause & effect
e.g.: cross-sectional, cohort, and case- e.g. parallel group, cross-over, factorial
control studies design, cluster randomized design…
lower level of scientific evidence than higher level of scientific evidence than
observational studies observational studies
2. Intervention study types
2.1. Randomised controlled trials: Parallel & cross-over
Purpose
- Test (hypothesises about) cause-and-effect relationships by randomizing participants to
study groups
- In all controlled studies – use of a ‘control group’
o Receive a placebo or no intervention
o Outcomes compared to intervention group
o Without control group, no cause-and-effect statement can be made as other
factors may be responsible
→ Use a control group to compare outcomes & account for other variables (e.g. placebo
effect, seasonal changes…)
Key RCT designs
Parallel group studies
- Each participant receives only one intervention (e.g. low or high intake, nutrient A or B…)
- Comparisons between participants
- Suitable for long-term studies (but shorter over-all time frame) and cases where
returning to baseline is difficult or unethical (e.g., weight, bone density)
- Least suitable to outcomes that show large inter-participant variations
Cross-over studies
- Each participant receives all interventions in a specified order
o More interventions = higher drop-out risk
4
