Epi in the News

1. What was the exposure? Healthy plant-based foods
2. How was this coded? Using a Likert’s scale. What was the outcome? Bowel cancer diagnosis
Is this an incidence or prevalence figure? Incidence
3. What was the hypothesised mediating factor? Antioxidants in plant-based food
4. What is the study design? Prospective Cohort study

A large study that involved 79,952 US-based men found that those who ate the largest amounts of healthy plant-based foods had a 22% lower risk of bowel cancer compared with those who ate the least.

5a. What is the name of the relative risk figure that was estimated? Risk ratio
5b. Who was in the numerator for the RR, who was in the denominator? Numerator: Risk of bowel cancer in men who ate the least plant-based foods
Denominator: Risk in men who ate the most plant-based foods
5c. What was the value of the RR? 0.78

...The researchers found no such link for women... The authors found the link among men also varied by race and ethnicity.
6. What epidemiologic concept is this? Confounding

7. Comment on the possibility of differential misclassification of exposure by outcome status.
I do not think that differential misclassfication is relevant in this study as the outcome occurred after the measurement of the exposure.

8a. What would be the key advantage and disadvantage of investigating this research question with a cross-sectional study? Advantage: No long period of follow-up
Disadvantage: cannot measure incidence and reverse causality will come in as exposure and outcome are measured simultaneously

8b. What would be the key advantage and disadvantage of investigating this research question with a randomized control trial?
Advantage: Confounding and biases are minimised since study participants are randomised
Disadvantage: ethics- will it be ethical to randomly allocate participants to consuming less healthy foods when there is some evidence that it could lead to bowel cancer?

What was the exposure? How was this coded?

Quintile of consumption of plant based food - ordinal variable

2. What was the outcome? Is this an incidence or prevalence figure?

Incidence of colorectal cancer

3. What was the hypothesised mediating factor?

nutritional quality of food

4. What is the study design?

Historical cohort study using Cox regression

A large study that involved 79,952 US-based men found that those who ate the largest amounts of healthy plant-based foods had a 22% lower risk of bowel cancer compared with those who ate the least.

5a. What is the name of the relative risk figure that was estimated?

Incidence rate (thanks Onyinyechi!) ratio (I do not think it was a risk difference)

5b. Who was in the numerator for the RR, who was in the denominator?

1 - (incidence rate of CRC in highest quintile of plant intake/incidence rate of CRC in lowest quintile of plant intake)

5c. What was the value of the RR?

0.78

...The researchers found no such link for women... The authors found the link among men also varied by race and ethnicity.
6. What epidemiologic concept is this?

effect modification

7. Comment on the possibility of differential misclassification of exposure by outcome status.

This seems very possible. People struggle to remember what they eat and social desirability bias means they tend to report more optimistically than the truth. It is likely that some people are more prone to social desirability bias than others and this will adversely effect the reliability of the study findings.

8a. What would be the key advantage and disadvantage of investigating this research question with a cross-sectional study?

Advantage - more accurate exposure if only ask what people ate on a specific day
Disadvantage - would have to rely on prevalence of outcome which would be prone to survival bias and also people who have already had CRC once may well have changed their diet as a result

8b. What would be the key advantage and disadvantage of investigating this research question with a randomized control trial?

Advantage - much better for establishing causal effect
Disadvantage - extremely difficult for people to stick to any diet for long. Not ethical - plant based diets are healthier, so could not randomise to low plant diet.

Thanks for asking the questions Sujit

Best wishes

Judith