[3A] The Scientific Study of Nutrition

one of the first to link nutrition (diet) with the incidence of chronic disease (CVD)

Farmington Heart Study

How old is nutrition as a form and evidence-based scientific discipline

Relatively very young (last 100 years)

The scientific method emphasizes creating a _________ hypothesis and then?

Falsifiable (the more the better, whatever is left must be true)

then gathering evidence to support or not support that hypothesis.

order of scientific method

observation and question > hypothesis > experiment > theory (when hypothesis supported) OR new observations and questions (hypothesis not supported)

theories are very reliable (they just don’t have final confirmation to make it a law)

Pellagra

disease characterized by dermatitis (“sunburn-like” damage in exposed areas), diarrhea, and dementia

What 3 observations did scientists observe about pellagra?

• More common amongst poor and institutionalized (hospital patients, orphans in state homes, prisoners). 

• Staff who worked in hospitals, orphanages, and prisons did not suffer pellagra, even if the residents did.

• Most common in SE USA, which economically and nutritionally depended on cotton and corn.

Based on these observations for pellagra, what hypothesis was proposed? Why?

pellagra was due to a dietary deficiency in the affected populations

• More common amongst poor and institutionalized (could be diet, access to HC, racial disparities?)

• Staff who worked in hospitals, orphanages, and prisons did not suffer pellagra, even if the residents did (suggests non infectious)

• Most common in SE USA, which economically and nutritionally depended on corn and cotton (could imply something wrong with corn, working outside all the time)

How did they test the pellagra hypothesis? What were the results?

nutritious foods such as milk, fresh meats, and vegetables were added to the diets of children in two orphanages.

RESULTS: Symptoms disappeared with supplementation, supporting the hypothesis that pellagra is due to a deficiency of something in the diet > theory that pellagra is caused by a dietary deficiency of niacin (vit B3)

What indicates that treatment has an effect?

A difference between treatment and control

Control groups should be what to treatment group

similar (i.e. age, sex, diet) to remove confounding effects

What is the purpose of placebo

Anticipation can lead to physiological effects, prevents participants from knowing which group they are assigned to

2 main types of study design in nutrition science

1. observational (epidemiological)

2. experimental (interventional)

Observational studies

 do not attempt to affect outcomes (e.g., provide a treatment).

Often large pop, look for correlation, can suggest mechanism but doesn’t give you the mechanism

Difference between no association, inverse association, direct association

Inverse association: as one variable increases the other decreases

Direct association: as one variable increases the other also increases

cross-sectional study

researchers analyze a population at a single, specific point in time  ("snapshot“) to identify factors that make that population unique.

e.g., People in the Mediterranean region drink lots of wine, eat plenty of fat from olive oil, and have a lower incidence of heart disease compared to North America and Europe.

cohort study

researchers follow a group of people who do not have an outcome at baseline to estimate the incidence of the outcome over time

e.g., How many people born in 1995 in China developed iodine deficiency by 2015?

case-control study

researchers compare people with and without an outcome (e.g., diabetes) to find factors associated with an outcome

(end of cohort study)

Experimental (interventional) studies

Find causation and relationship

Attempt to uncover a direct, causal relationship between two variables by comparing groups with (treatment) and without (control-control, placebo) an intervention applied to them.

Generally more powerful than observational studies, but they are also much more difficult to conduct (costs, ethical considerations, and approval processes)

Gold standard of human interventional studies

double-blind randomized control trial (RCT)

Data dredging “p-hacking”

abuse of data mining where large compilations of data are examined to find any correlation, without any pre-defined hypothesis (e.g., to explain the mechanism of the correlation).

What is the “File drawer” problem

Papers that have more significant results get more attention – someone will do a good study with larger P, data gets put into “file drawer”, if a study doesn’t have a huge stat significance why do we need to publish it?

Fabricated trends on graph - line

If you put a line through it that suggests a significant trend which can sometimes be misleading (drawing trendlines through uncorrelated P > 0.05)

What are the best measurements for effect size and relationship

R² [scatter] or Fstat

NOT P

What does each part of CRAAP stand for

C - currency: timeliness of information

R - relevancy: importance of info for your needs

A - authority: source of information

A - accuracy: reliability, truthfulness, correctness of content

P - purpose: reason information exists