Introduction to Nutrition Research
Section 1: Types of Nutritional Studies
Section 2: Statistical Concepts and Interpreting Quantitative Results Section 3: Reading the Scientific Evidence: Cohort Studies
Section 4: Reading the Scientific Evidence: Randomized Controlled Trials
Copyright By PowCoder代写 加微信 powcoder
Section 5: Final Comments
Section 1: Types of Nutritional Studies
Finding Nutrition Research information
Two types of studies
u Observational studies
u Prospective cohort study
u Randomized controlled trial (RCT)
This Photo by Unknown Author is licensed under CC BY- NC-ND
Observational studies determine whether there is an association between exposure and outcome
Cohort studies
HYPOTHESIS: Nutrient X reduces the risk of disease Y.
Nutrition Research: Observational studies
Association/Correlation u Direct or positive
u Intake of dietary component à Disease risk
Association/Correlation u Inverse or negative
u Intake of dietary component à u Disease risk
Intake of dietary component
Number of disease cases
Intake of Dietary Component
Number of disease cases
Nutrition Research:
Limitation of observation studies
u Vegetable intake à Cancer risk
Confounding factor:
Related to both the outcome being investigated (cancer cases) and the exposure (e.g. dietary outcome)
Vegetable Intake
Cancer cases
u Confoundingfactors
u What if low vegetable intake is a characteristic of an older population?
u How do you determine if you are measuring the effect of vegetable intake or age? u Statistical adjustment for confounding factors
u E.g. age adjustment
u Problems
u Can only correct for confounders that are identified and measured u Uncorrectedorunknownconfoundingàresidualconfounding
u Becauseofresidualconfounding,onlyassociationsbetweenexposureandoutcomecanbe measured; causal links cannot be established
Nutrition Research:
(Fig 1.18 from text) Prospective cohort study
HYPOTHESIS: Nutrient X reduces the risk of disease Y.
Measuring exposure:
Validation of food frequency questionnaire
u Food Frequency Questionnaire (FFQ)
How often do you eat these foods?
< Once/ once/ month month
Food Records
2-3 times /month
Once per week
Twice/ week
2-4 times /week
5-6 times /week
2 or more times/da y
Carrot Consumptioon based on:
Repeated Daily Food Records
2-4 times/week
3 times/week
I just had a carrot. Let me write this down.
VALIDATION
Food Frequency Questionnaire
u Repeated FFQ measures over course of study u Calculate a cumulative average food intake
u Accounts for changes in diet that might take place over time especially in very long studies
This Photo by Unknown Author is licensed under CC BY-SA- NC
Measuring Outcomes
“Hard” Disease Outcome
Risk Factors
u Risk factor for cardiovascular disease
u Hypertension
u blood glucose
u Inflammation (C-reactive protein) u Blood lipids:
u LDL cholesterol u HDL cholesterol
u Ratio: Total cholesterol: HDL Cholesterol u LDL cholesterol:HDL cholesterol
u ApoB (associated with LDL cholesterol) u ApoA-I (associated with HDL cholesterol)
u Cardiovascular events; u fatal and non-fatal
u Examples of events:
u Myocardial infarction (heart attack) u Angina pectoris
u Pulmonary embolism
u Transcient ischemic attack
u Peripheral arterial disease
Cohort Study
Limitation
u Shows associations, but not causal links, between exposure and outcome
u Vulnerable to confounding & residual confounding
u Usually measures disease outcomes
u E.g. cardiovascular disease, cancer
Nutrition Research:
Randomized Controlled Trial (RCT)
Control group Intervention or treatment group u Diet typical of the population u Substantial change in the diet
Randomization
Assignment to the control group or treatment group is purely by chance
Implications of randomization:
• Confounding factors are equally distributed between control and
treatment group
• The only difference between control and treatment groups is the
treatment or intervention
If treatment is effective in altering outcome,
then causation is indicated or a causal link is made
Nutrition Research: Fig 1.18 (from text)
Intervention Trial
HYPOTHESIS: Nutrient X reduces the risk of disease Y.
Nutrition Research:
Randomized Controlled Trials (Characteristics)
u Blinding of study
u Whether the participant is assigned to control or treatment group is unknown to u Both participant and researcher = double blind study
u Either the participant or researcher = single blind study
u Blinding is not always possible in a diet study
u Dummy treatment given to control group to eliminate the effect of participant expectations
Nutrition Intervention
u Metabolic diet
Here are your meals for today.
Thank you.
u Nutrition education
Nutrition Control or Comparison Group
u Eat regular diet (i.e. don’t change anything)
u Healthy diet – e.g. CFG – but not expected to be as effective as intervention
u Limited nutrition education
Nutrition Research: Randomized Controlled Trials
Strengths Limitations
u Shows causation
u Evidence from intervention trials considered stronger than observational studies
u Difficult to sustain especially if you are making substantial changes to the diet
u Commonly measures risk factors for disease rather than a disease outcome
u E.g. high blood pressure rather than cardiovascular disease
Nutrition Research: RCTs vs Observational Studies
u Shows causation, or causal links, between intervention and outcome
u Evidence from intervention trials considered stronger than observational studies
Limitations
u Usually measures risk factors of disease rather than “hard” disease outcomes
u Difficult to sustain especially if you are making substantial changes to the diet
Prospective cohort studies: Strengths
u Usually measures disease outcomes
Limitations
u Shows associations, but not causal links, between exposure and outcome
u Vulnerable to confounding 22
Nutrition research
u Relies on results from both intervention trials and observational studies to make dietary recommendations
u Systematic reviews look at the totality of the scientific evidence u Summaries of results from multiple studies
1.5. Mediterranean Diet
u Mediterranean diet pyramid
Dietary pattern
Mediterranean Diet: How strong is the scientific evidence that the Med diet reduces the risk of cardiovascular disease?
Interpreting statistical data
Section 2: Statistical Concepts and Interpreting Quantitative Results
Statistical concepts
u Null hypothesis
u This is a somewhat counterintuitive concept. Scientists tend present their hypotheses in a positive light, e.g. increased diet quality results in decreased disease risk. But the null hypothesis describes a situation of no effect e.g. diet quality will have no effect on disease risk; there will be no difference in disease risk between groups, whether consuming high or low quality diets.
u Alternative hypothesis
u This is an hypothesis that is an alternative to the null hypothesis e.g. There will be a difference in disease risk between groups consuming high quality diets compared to low quality diets
Statistical concepts
u Mathematical equations, used in statistical analysis, that best fit the data
u Probability or P-value
u Expressed as a decimal between 0 and 1, where P = 0 means there is no probability of an event occurring and P = 1, which means there is a certainty that an event will occur. P- values can also be expressed as % e.g. p = 0.50 or 50% probability of an event occurring
u Determines the probability that the observed difference (or greater difference) between groups being compared would occur with repeating experiments assuming:
uThe null hypothesis is true
uAll assumptions used in the mathematical model are correct uChance alone is operating
Statistical Concepts
u Consider these study results:
u The high nutrient intake group has a 40% lower risk of developing CVD
compared to the low nutrient intake group (P =0.002)
u Interpretation: assuming the null hypothesis is true, the model assumptions are correct, and chance alone is operating, a difference as large as (or larger than) the one observed, would occur in 1 out of 500 experiments
u i.e. a rare event
u Or the null hypothesis is incorrect; there is probably a real difference between the high nutrient intake group vs low nutrient intake group
Prospective cohort study: Are the differences in disease outcome between low and high nutrient intakes statistically significant?
Randomized control trial: Are the differences between control group and treatment group statistically significant?
“Significant” in science does not mean the same thing as “significant” in common language.
“The results are significant” refers to the results of a statistical analysis.
What is the relationship between P-value and statistical significance?
Or what is the level of significance.
Low Exposure vs High Exposure Observational Study
Control vs Intervention Randomized Controlled Trial
Statistical Analysis:
What is the probability that the observed differences between groups are due to chance, assuming the null hypothesis is true?
P > 0.05 (greater than 5%) There is no statistically significant difference between groups.
P < 0.05 (less than 5%)
There is a statistically significant difference between gr3o2 ups.
Statistically significant: P < 0.05
Not statistically significant: P > 0.05
Evidence favours rejecting the null hypothesis.
But does not prove that the null hypothesis is false, only that the evidence supporting it is weak; the null hypothesis, however, may still be correct.
Evidence favours not rejecting the null hypothesis but does not prove that the null hypothesis is true; only that the evidence supporting it is strong; but the null hypothesis may still be incorrect.
Continues..
Understanding relative risk
N= 10,000 N= 10,000 A = 4000
Similar to Relative Risk:
Hazard Ratio (HR) Odds Ratio (OR)
RR > 1 increased risk RR < 1 decreased risk
Does Following a Healthy Dietary Pattern Reduce the Risk of Disease? (pg. 85 of your textbook)
Cohort Study:
u Dietary intake for 100,000 people obtained
u Evaluated diet quality using scoring system
u Health info and disease occurrence recorded for several years
u Determined if diet quality effected disease risk
Highest AHEI score-best diet
Lowest score vs highest score
McCullough ML, C. Evaluating adherence to3r7ecommended diets in adults: the Alternate Healthy Eating Index. Public Health Nutr. 2006 Feb;9(1A):152-7.
Population divided into 5
Quartiles: 4 groups Tertiles: 3 groups
95% Confidence Intervals (CI) for relative risk:
u Definition of a 95% confidence interval:
There is a 95% chance that the interval contains the “true” population value.
u Relative risk:
u Reference group: RR = 1
u High intake group: RR = 1.5 (95% confidence intervals: 1.38 – 1.65)
u There is a statistically significant difference from the reference group because interval does not include 1
u High intake group: RR = 1.5 (95% confidence intervals: 0.96 – 1.94)
95% CI indicates
statistical significance and the likely range of the risk.
u Not statistically significant because interval includes 1
Continues...
Relative risk with confidence intervals included
RR: CVD men
Lowest AHEI score-poorest diet
0.85 (0.71 – 1.01)
0.79 (0.66 - 0.95)
0.67 (0.56 – 0.81)
Highest AHEI score-best diet
0.61 (0.49 – 0.75)
McCullough ML, C. Evaluating adherence to recommended diets in adults: the Alternate Healthy Eating Index. Public Health Nutr. 2006 Feb;9(1A):152-7.
Relative risk with confidence intervals included
Lowest AHEI score-poorest diet
Highest AHEI score-best diet
RR: CVD men
0.85 (0.71 – 1.01)
0.79 (0.66 - 0.95)
0.67 (0.56 – 0.81)
0.61 (0.49 – 0.75)
P trend = 0.001
RR: Cancer men
1.10 (0.94 – 1.28)
0.99 (0.85 – 1.16)
0.94 (0.80 – 1.10)
1.03 (0.87 – 1.22)
P trend = 0.68
McCullough ML, C. Evaluating adherence to recommended diets in adults: the Alternate Healthy Eating Index. Public Health Nutr. 2006 Feb;9(1A):152-7.
Section 3: Reading the Scientific Evidence: Cohort Studies
Consider the following:
Based on the studies presented, would you conclude that there is convincing evidence that adhering to the Mediterranean diet reduces the risk of cardiovascular disease?
Evaluating cohort studies
1. Study design: Prospective cohort study:
2. Population; age, gender, healthy, etc.
3. Study location
4. Exposure
5. How was dietary intake assessed? e.g FFQ; was FFQ validated? Repeated?
6. Outcome(s): disease or risk factors
7. Duration of study
8. Result: What is the relationship between exposure and outcome? Statistically significant association i.e. p for trend <0.05
9. Adjustment for confounders
10. Strengths of study: Disease outcome
11. Limitations of study: Measures association not causation; residual confounding
12. Additional comments
Buckland G, González CA, Agudo A, Vilardell M, , Amiano P, Ardanaz E, Arriola L, Barricarte A, Basterretxea M, Chirlaque MD, Cirera L, Dorronsoro M, Egües N, M, Larrañaga N, , Martínez C, , , Quirós JR, , J, Tormo MJ, Moreno-Iribas C.
Adherence to the Mediterranean diet and risk of coronary heart disease in the Spanish EPIC Cohort Study. Am J Epidemiol. 2009: 170(12):1518-29. doi: 10.1093/aje/kwp282.
This study examined the relation between Mediterranean diet adherence and risk of incident CHD events in the 5 Spanish centers of the European Prospective Investigation into Cancer and Nutrition (the EPIC study). Analysis included 41,078 participants aged 29-69 years, recruited in 1992-1996 and followed up until Dec 2004 (mean follow-up:10.4 years). Food intake was evaluated using a validated FFQ, measured once at the beginning of the study. Incident fatal and nonfatal CHD events were analyzed according to Mediterranean diet adherence, using an 18-point Mediterranean diet score. A total of 609 participants (79% male) had a fatal or nonfatal confirmed acute myocardial infarction (n = 468) or unstable angina requiring medical intervention (n = 141). After adjustment for recognized confounding factors such as education, physical activity, BMI, smoking status, diabetes, hypertension, hyperlipidemia, total calorie intake, the following association was observed:
Medium score
0.86 (0.70-1.04)
High score
0.60 (0.47-0.77)
Buckland G et al
Adherence to the Mediterranean diet ❹ and risk of coronary heart disease ❻ in the Spanish EPIC Cohort Study. ❶
Am J Epidemiol. 2009: 170(12):1518-29. doi: 10.1093/aje/kwp282.
This study examined the relation between Mediterranean diet adherence and risk of incident CHD events❻ in the 5 Spanish centers ❸of the European Prospective Investigation into Cancer and Nutrition (the EPIC study). Analysis included 41,078 participants aged 29-69 years,❷ recruited in 1992-1996 and followed up until December 2004 (mean follow-up: 10.4 years)❼. Food intake was evaluated using a validated FFQ,❺ measured once at the beginning of the study. Incident fatal and nonfatal CHD events were analyzed according to Mediterranean diet adherence, measured by an 18-point Mediterranean diet score ❹....
1. Study design – assn or causation
2. Population; age, gender, healthy,
3. Study location
4. Exposure
5. How was dietary
intake assessed e.g FFQ; validated, repeated
6. Outcome(s): disease or risk
7. Duration of study
8. Result: What is
the relationship between exposure and outcome e.g. SS ass’n
9. Adjustment for confounders
10. Strengths
1. Study design –assn or causation
2. Population; age, gender, healthy, etc.
3. Study location
4. Exposure
5. How was dietary
intake assessed e.g FFQ; validated, repeated
6. Outcome(s): Disease or risk factor
7. Duration of study
8. Result: What is the
relationship between exposure and outcome e.g. SS ass’n
9. Adjustment for confounders
10. Strengths 11. Limitations
A total of 609 participants (79% male) had a fatal or nonfatal confirmed acute myocardial infarction (n = 468) or unstable angina requiring medical intervention (n = 141). After adjustment for recognized confounding factors such as education, physical activity, BMI, smoking status, diabetes, hypertension, hyperlipidemia, total calorie intake,❾ the following results were obtained:
Medium score
0.86 (0.70-1.04)
High score
0.60 (0.47-0.77)
Tektonidis TG, Åkesson A, , , C. 2015. A Mediterranean diet and risk of myocardial infarction, heart failure and stroke: A population- based cohort study. Atherosclerosis. 2015 Nov;243(1):93-8. doi: 10.1016/j.atherosclerosis.2015.08.039.
BACKGROUND AND AIMS: The Mediterranean diet, which is palatable and easily achievable, has been associated with lower all-cause and cardiovascular disease (CVD) incidence and mortality. Data on heart failure (HF) and stroke types are lacking. The aim was to examine a Mediterranean diet in relation to incidence of myocardial infarction (MI), HF and stroke types in a Swedish prospective cohort.
METHODS: In a population-based cohort of 32,921 women, (age 48 to 83 years), free of CVD and cancer at the beginning of the study, diet was assessed through a validated food frequency questionnaire, measured only once at the start of the study. The modified Mediterranean diet (mMED) score was created based on high consumption of vegetables, fruits, legumes, nuts, whole grains, fermented dairy products, fish and monounsaturated fat, moderate intakes of alcohol and low consumption of red meat, on a 0-8 scale. Relative risks (RR) with 95% confidence intervals (CI), were determined and adjusted for confounders, such as family history of heart disease, history of hypertension, history of hypercholesterolemia, history of diabetes, smoking, physical activity, BMI, aspirin use, and energy intake (kcal/day).
1. Study design –assn or causation
2. Population; age, gender, healthy, etc.
3. Study location
4. Exposure
5. How was dietary
intake assessed e.g FFQ; validated, repeated
6. Outcome(s): Disease or risk factor
7. Duration of study
8. Result: What is the
relationship between exposure and outcome e.g. SS ass’n
9. Adjustment for confounders
1. Studyddeesisgingn–assn or causation
2. Population;
2. Population; age,
age, gender,
gender, healthy,
healthy, etc.
3. Study location
3. Study location
4. Exposure
4. Exposure
5. Methodology to
5. How was dietary imnteaakseuarsesessed e.g
FFQ; validated,
exposure – how
was dietary
6. Outcome(s):
intake assessed
disease or risk
6. Outcome(s)
7. Duration of study
7. Duration of
8. Result: What is the
relationship
8. Rbetswuelte:nWexhpaotsuisre
and outcome e.g.
the relationship
9. Adjustment for
exposure and
confounders
outcome e.g. SS
RESULTS: During 10 y of follow-up, 1109 MIs, 1648 HFs, 1270 ischemic strokes and 262 total hemorrhagic strokes occurred. The following results were observed:
Quartile 1: Lowest mMed Score
Quartile 2 Intermediate mMed Score
Quartile 3 Intermediate mMed Score
Quartile 4: Highest mMed Score
Myocardial infarction
0.89 (0.76-1.04)
0.75 (0.63-0.89)
0.74 (0.61-0.90)
Heart failure
0.81 (0.71-0.93)
程序代写 CS代考 加微信: powcoder QQ: 1823890830 Email: powcoder@163.com