1
Saturday, 14th May 2022
Lecture
Medicine
Diet and cancer
What causes cancer?
Recognise several reversible diet factors in cancer
50% related to reversible risk factors
How does this relate to other causes of ill health?
Impact of diet on all forms of ill health is significant
What dietary factors cause cancer?
Obesity
o Significant risk factor
o Well established
Alcohol
Processing
o Preserved foods associated with cancer
o Related to ingestion of salt
Salt
Diet
o Talking generally about a healthy diet
o Good proportion of fruits and vegetables
o More general approach and the risk of cancer
5 a day campaign
Establishing causality
Nailed down association between tobacco smoke and lung cancer
Difficult to establish causality in a scientific way
Criteria
o Eight main criteria
o May not have positive signals for all
Temporal
Risk factor and outcome
o Risk precedes outcome
Specific
Consistent
Strong
Strong association
Dose response
Shows strong relationship
Modifiable
By modifying risk factor, you can change the outcome
Plausible
Mechanism
Coherent
Causality needs to be consistent with the science
o Correlation is not causality
Need to look at individual points to build a case for causality
More evidence there is the more likely it is
Issues with causality: confounding
, 2
Saturday, 14th May 2022
Lecture
Medicine
One risk factor associated with a second risk factor associated with an outcome
o Oral contraceptives linked to cervical cancer
o Use of the pill associated with decrease of barrier methods, increase in
passing on of HPV which causes cervical cancer
Issues with causality: reverse causality
Risk factor associated with an outcome which leads to further risk factors incorrectly
associated with the outcome
o Low body weight thought to be a risk factor for lung cancer
o Increased weight a risk factor
o Obesity leads to lung cancer leads to weight loss
Behaviour
o Increased alcohol leads to angina leads to decrease in alcohol
o Alcohol decrease associated with angina even though it is due to increased
alcohol intake
Case control study
Establish cases of outcome of interest
Establish controls without the outcome
Match for age and gender and other factors which may confound
Go back in time and establish if there was a risk factor in the past
Pros
o Fast
o Cheap
o Good for rare disease
Able to gather all with a rare condition and establish common risk
factors
Cons
o Recall bias
Not always remember accurately
o Difficult to measure the amount of risk
Suggestion of a signal but unable to say how strong it is
o Impact of disease on risk
Impact how they behave and how they recall they behaved
Possible reverse causality
Cohort study
Whole populations of people
o Various ways of doing this
Interest in particular outcome
o Would not include those with the outcome
Monitor those who are outcome free
Information collected in real time
See if there is a difference in outcome between those with and without risk factors
Pros
o Near definitive data of association
o Range of risks and outcomes can be measured
Cost efficient
Saturday, 14th May 2022
Lecture
Medicine
Diet and cancer
What causes cancer?
Recognise several reversible diet factors in cancer
50% related to reversible risk factors
How does this relate to other causes of ill health?
Impact of diet on all forms of ill health is significant
What dietary factors cause cancer?
Obesity
o Significant risk factor
o Well established
Alcohol
Processing
o Preserved foods associated with cancer
o Related to ingestion of salt
Salt
Diet
o Talking generally about a healthy diet
o Good proportion of fruits and vegetables
o More general approach and the risk of cancer
5 a day campaign
Establishing causality
Nailed down association between tobacco smoke and lung cancer
Difficult to establish causality in a scientific way
Criteria
o Eight main criteria
o May not have positive signals for all
Temporal
Risk factor and outcome
o Risk precedes outcome
Specific
Consistent
Strong
Strong association
Dose response
Shows strong relationship
Modifiable
By modifying risk factor, you can change the outcome
Plausible
Mechanism
Coherent
Causality needs to be consistent with the science
o Correlation is not causality
Need to look at individual points to build a case for causality
More evidence there is the more likely it is
Issues with causality: confounding
, 2
Saturday, 14th May 2022
Lecture
Medicine
One risk factor associated with a second risk factor associated with an outcome
o Oral contraceptives linked to cervical cancer
o Use of the pill associated with decrease of barrier methods, increase in
passing on of HPV which causes cervical cancer
Issues with causality: reverse causality
Risk factor associated with an outcome which leads to further risk factors incorrectly
associated with the outcome
o Low body weight thought to be a risk factor for lung cancer
o Increased weight a risk factor
o Obesity leads to lung cancer leads to weight loss
Behaviour
o Increased alcohol leads to angina leads to decrease in alcohol
o Alcohol decrease associated with angina even though it is due to increased
alcohol intake
Case control study
Establish cases of outcome of interest
Establish controls without the outcome
Match for age and gender and other factors which may confound
Go back in time and establish if there was a risk factor in the past
Pros
o Fast
o Cheap
o Good for rare disease
Able to gather all with a rare condition and establish common risk
factors
Cons
o Recall bias
Not always remember accurately
o Difficult to measure the amount of risk
Suggestion of a signal but unable to say how strong it is
o Impact of disease on risk
Impact how they behave and how they recall they behaved
Possible reverse causality
Cohort study
Whole populations of people
o Various ways of doing this
Interest in particular outcome
o Would not include those with the outcome
Monitor those who are outcome free
Information collected in real time
See if there is a difference in outcome between those with and without risk factors
Pros
o Near definitive data of association
o Range of risks and outcomes can be measured
Cost efficient
