Module 15 Flashcards
What are the nine criteria for assessing causality, according to Bradford Hill?
- Temporality or time sequence
- Consistency or repetition across studies
- Biological or social plausibility
- Strength of association
- Dose response or biological gradient
- Specificity
- Coherence
- Experimental evidence
- Analogy
How to decide if an exposure is causal
The general question:
-Is there a cause and effect relationship between the presence of factor X and the development of disease Y?
The answer:
-is made by inference and relies on a summary of all valid evidence
What are the five most important Bradford Hill criteria?
Temporality
Consistency across studies
Biological or sociological plausibility
Strength of association
Dose response (or biological gradient)
Temporality
In order for an exposure to cause a dz, the exposure must precede the dz in time
Consistency of association or repetition across studies
The same results are replicated by different investigators in different settings with different methods. They are repeatedly observed in different places and times.
Biological credibility, biological plausibility, or sociological possibility
Is there a logical causal pathway by which this exposure could cause the outcome? Is the association biologically supported? Or sociologically, if social factors are being studied?
But suppose the exposure is only causal under certain conditions or in certain populations but not others?
Strength of association
The point estimate or measure of association between the exposure and dz is strong
But is the state of our knowledge always sufficient to identify a causal pathway?
Dose-dependent relationship of biological gradient
An increasing amount of exposure is associated with the risk of dz in an increasing or decreasing pattern
But do some exposures have thresholds?
Judgment
Once a statistical association is deemed valid, then the totality of the evidence needs to be considered before making a judgment about whether causality exists for a specified hypothesis
Health related states or events typically have a ______ etiology
Multi-factorial
Combinations of necessary/sufficient causes
Necessary and sufficient
Necessary but not sufficient
Sufficient but not necessary
Neither sufficient nor necessary
Necessary and sufficient
The factor always causes dz and is enough to cause the dz alone
Very few causes are necessary and sufficient
Necessary but not sufficient
The factor MUST be present to cause the dz, but other factors are also required
Sufficient but not necessary
The factors, when present always causes the dz, but there are other ways of getting the dz without the factor
Neither sufficient nor necessary
The factor is a contributing cause to the dz but is not always present in all cases of the dz and cannot cause the dz alone
Most risk and protective factors are this type
Necessary and Sufficient Causes and causal pies
Component cause (piece of the pie)
Sufficient cause (the whole pie)
Necessary cause (an always present piece of pie)
Component cause
Any one of a set of conditions which are part of the sufficient cause
Sufficient cause
The whole group of component causes together, which inevitable leads to the dz
Necessary cause
A component cause that is part of every sufficient cause
3 steps to evaluating study results
- Consider the role of chance
-Statistical significance - Consider the role of bias
-Selection bias
-Information bias
-Confounding - Consider criteria for causality
-Sir Austin Bradford Hill’s 5 most important criteria
Koch’s postulates for infectious dz
To be considered a CAUSE of a dz a microorganism:
-Should be found in all pts with the dz
-Should be grown in vitro (in a suitable medium)
-Should be capable of reproducing the dz in some animal species
-Should be then recovered and grown again in a pure culture
Example of temporality, time sequence
HTN precedes a stroke in time
Example of consistency of association or repetition across studies
Many different studies in many parts of the world have shown an association between low SES and mortality…not just one study…not just one group of ppl
Example of biological credibility, biological plausibility, or sociological plausibility
Gambling may be statistically associated with cirrhosis of the liver, but is it biologically plausible? Excessive alcohol consumption is also associated with cirrhosis of the liver, and is biologically credible, since the liver processes the alcohol
Example of strength of association
In some studies, heavy smokers of many years were 20-30 times more likely than ppl who never smoked to develop lung CA, not just slightly more likely
Example of dose-response relationship or biological gradient
The overweight are more likely to develop type 2 DM than those with nl weight. The moderately obese are more likely to become diabetic than the overweight. The very obese are still more likely to become diabetic than the moderately obese. The more the weight, the greater the chance of diabetes.
