Cancer and the diet Flashcards
Where is oesophageal cancer most prevalent?
Middle east and china
Where is gastric cancer most prevalent?
Russia
Where is colon cancer most prevalent?
Western world, UK and US
Why do we have a 5 day campaign?
Evidence from observational epidemiology that average fruit/veg intake of less than 200g associated with increased risk of cancer, but possibly little additional benefit beyond 400g/day
Is there evidence that 5 a day has an impact on cancer?
Very little
Explain how the randomisation process in RCTs is carried out
- Block randomisation - Recruit participants into small blocks, randomly allocate equal numbers in each block to treatment A or B
- Stratified randomisation - Equal number of people with a characteristic that may affect prognosis in each group e.g. breast cancer (same numbers of pre and post-menopausal women), keeps groups as similar as possible
What are case-control studies?
Find a population with similar people who have an outcome and a control group who do not. Look back in time to see what their exposure status to a particular risk factor was.
What are problems with a case-control study in nutritional epidemiology?
- Recall bias, not just people’s memories:
- People with disease are likely to recall better due to their liver having been impacted by disease
- People without disease unlikely to make as much of an effort
- So if you don’t have the same level of information for the control and the case group then you might see relationships that aren’t there - Possible early impact of disease on diet - If the cancer impacts on diet early on then it may be difficult to distinguish whether the diet caused the cancer or if the diet was a result of the early cancer stages
What is a cohort study?
Take a population free of disease and study their exposure to a certain risk factor to see if they develop disease.
Problems with a cohort study in nutritional epidemiology?
- Measuring diet in large group
- Maintaining follow-up over long period
- Cancer takes a long time to develop
Problems for all observational studies
- Bias - This is a risk for ALL research
- Confounding - Patterns of behaviour associated with certain types of diet so these behaviours may be impacting
- Hard to establish causal relationship
What are problems with measuring diet
- Random error - Diet varies and difficulties in measurement, people don’t eat the same things everyday and out individual consumptions vary significantly
- Homogeneity of exposure - If you only do your studies in the same types of populations (e.g. continuously only use white males) they are likely to have similar environments and hence diets, so you are not able to apply results to the population
- Bias
- Confounding
What are different measures of diet?
- Food disappearance data
- Household survery - What do you buy and who eats what?
- Individual survery - 24 hour recall, food frequency (very open to bias), diet diary, biomarkers (very rarely have this)
Pros and cons of food frequency questionnaires
Pros - Captures usual diet and less work to code/complete
Cons - Doesn’t record actual diet as eaten, overestimates fruit and vegetables, poor measure of energy intake, less flexible
Pros and cons of diet diaries
Pros - Records diet as eaten (over limited period), better estimate of energy and absolute intake, more flexible
Cons - Required effort to complete and expensive to code
