Epidemiology 3 Flashcards
Why is Bradford Hill used?
Enables us to infer causation from both observational and interventional methods
What are the criteria in Bradford hill?
- Strength
- Consistency
- Specificity
- Temporality
- Biological gradient
- Plausibility
- Coherence
- Experiment
- Analogy
What is the strength criteria?
The stronger association increases the confidence that an exposure causes an outcome
What is the consistency criteria?
Consistent findings across setting tended to rule out error or fallacies that might befall one or two studies
What is the specificity criteria?
Specificity describes an association between specific causes and specific effects
What caveats are in the specificity criteria?
A specific disease arising among specific workers is valuable in supporting the argument for casualty, but conceded a lack of specificity does not necessarily invalidate a casual relationship
Is specificity necessary?
Can be informative when present but its absence conveys very little
What is the temporality criteria?
It is insufficient for exposure A and outcome B to coexist A but precede B
How is temporality used in observational study?
In observational study design take a cross sectional approach that determines the presence of exposure A and outcome B
What is the biological gradient criteria?
A dose response effect (in ‘right’ direction) is a compelling argument for casualty (more cigarettes greater risk of lung cancer)
Is biological gradient important?
- Quantification often difficult
- When present it is suggestive but tis absence may be of no value
What is the biological plausibility criteria?
- Relationship should be biologically plausible where science is ‘understood’
- When deficient understanding assessing whether a relationship is plausible or not may not be possible
What is the coherence criteria?
The association ought to be consistent with the existing theory and knowledge
What is the difference between coherence and biological plausibility?
- Seeks to ensure that the association is in keeping withe existing science
- About broader picture, goof when existing scientific theory is correct but if most scientist are wrong, this criteria may support he states quo
What is the experimentation criteria?
- Evidence from experimentation should be supportive of the proposed link
- Observational studies will nerve conclusively prove causation
Why is experimentation sometimes not possible?
Ethics especially in public health issues
What is the analogy criteria?
Drawing upon analogous finding may make inference on the relationship
When is analogy important?
- Understanding emergent disease and new association
- Rest on existing knowledge
What is correlation?
Linear relationship between two variables
Does Bradford Hill give a yes or no?
No, need to have own idea
What is internal validity?
- if believe that the association truly exists in that group of patients
- Extent to which findings accurately describe the relationship between exposure and outcome in the context of the study
What is external validity?
if applicable to other groups of cancer patients e.g. outside of the population
What is external validity the same as?
generalisability
What is bias?
any trend in the collection, analysis, interpretation, publication or review of data that can lead to conclusions that are systematically different from the truth
What does systematic error lead to?
systematic error in the design or conduct of a study can result in bias - which means observed results may be different from the truth - and hurts validity
How do you stop random error?
- Always error expected in measurements
- over or under estimate height when measuring due to chance have random error - problem to precisions
- large enough sample error cancels each other out
What happens if systematically overestimate height?
lead to inaccurate results regardless of sample size
What are the three types of validity?
- Selection bias
- Information bias
- Confounding
Without internal validity can you consider external validity?
Rarely
What is selection bias?
Individuals chance to be included in study sample is related to BOTH the exposure and the outcome - leads to bias estimate of the association between the exposure and outcome
What is Berkson’s bias?
when hospital based case controlled study and controls are selected among the hospital’s patients (selection bias)
What is healthy worker effect?
in occupational studies where active workers are more likely to be healthy compared to those that have retired or stopped working
What type of study is really susceptible to section bias?
Case controlled study
How do you minimise section bias?
- Controls representative of target population: in terms of exposure to factors in the study
- Minimise non-response: if more people decline to participate more likely selection bias introduced
- Compare respondents with non respondents: to see if systematic differences
What is information bias?
Misclassification of the exposure or the disease status or both
How can people be misclassified?
some heavy smokers classified as light smokers or some lung cancer patients may not receive correct diagnosis
Why does misclassification happen?
- study variables are not properly defined
- flaws in data collection
What are two types of flaws in data collection?
- Interviewer bias
2. Recall bias
What is interviewer bias?
- interviewer will ask lung cancer or no lung cancer if they have been smoking
- interviewer may be more thorough interviewing people about past smoking who have been diagnosed with lung cancer or not because they expect lung cancer patients to have been smokers
- lead to misclassification of exposure status and eventually to a bias odds ratio
How do you prevent interviewer bias?
- Interviewer does not know the disease status of the individual
- Collection process has been carefully standardised so that the interviewer follows a strictly defined protocol when collection data from participants
What is recall bias?
the differential inaccurate recall of past exposure between cases and controls - when all participants have trouble remembering their exposure status but has nothing to do with disease no recall bias
How can you prevent recall bias?
- using objective ways to assess exposure for example
1. medical records
2. biomarkers
What is non-differential misclassification?
- When exposure status is misclassified but equally among cases and controls
- Same term applies with errors determining outcome but occur equally among exposed and non-exposed individuals
What is the bias like in non-differential misclassification?
Odds ratio obtained bias towards the null
What is misclassification differential?
when errors in determining an individuals exposure status occur unevenly among cases and controls or when errors in the diagnosis of the disease which occur unevenly among exposed and non exposed individuals
What is the bias like in differential misclassification?
bias towards or away from the null
What is response bias?
describes the tendency of participants to respond to questions in a way that is more socially acceptable
Is response bias a type of selection bias?
NO
What is non-response bias?
process through which people who do not respond are systematically different to the people who do respond
Is non-response bias a type of selection bias?
YES
What is Hawthorne effect?
- usually described as the consequence of participants realising they are being observed and therefore acting differently
- For example, if you knew you were being studied, you might behave in a slightly different way to normal
What are types of information bias?
- Recall bias
- Response bias
- Interviewer bias
- Diagnostic bias
What is confounding?
the effect of an extraneous variable that wholly or partially accounts for the apparent effect of the study exposure, or that masks an underlying true association
What can confounding lead to?
bias estimates and misleading results
What are the four ways to identify confounding?
- Knowledge of subject matter
- Three conditions for confounding
- Stratification
- Compare crude and adjusted estimate
- Only need one
What are the three conditions for confounding?
- Associated with the exposure in the source population
- Associated with the outcome in the absence of the exposure
- Not a consequence of the exposure
How do you find knowledge of the subject matter?
Could undertake an evidence review and see what other people have proposed or found
How can you stratify?
Look at difference of apparent effect within different population strata
What is an effect modification?
After controlling for confounding variable may still be third variable the impact of which on the association between exposure and outcome is so important cannot be ignored
When does effect modification exist?
when the strength of the association varies over different levels of a third variable
How does effect modification impact the study?
take into account with the way that you report results e.g. so present the stratified results with the antibiotic
Should you find a way to control effect modification?
No, natural phenomenon so is is not a problem with the study
How can you test for effect modification?
- Breslow-Day test
- Q test
- Interaction terms in regression models
What is synergism?
effect modifier potentiates the effect of the exposure
What is antagonism?
the effect modifier diminishes the effect of the exposure
Why is it valuable to know if something is confounding or effect modification?
- Addressing a confounded relationship by addressing the exposure exclusively is very unlikely to yield a gain. In example (3), stopping a mother from smoking is unlikely to reverse the childhood mental health outcome.
- Addressing an exposure where effect modification is apparent may be useful: it may be possible to target an intervention into a more homogeneous pool of participants where a greater impact will be yielded
What is a crude model?
- univariate analysis of exposure vs outcome
- simply looking at the impact of the exposure on the outcome – with no consideration of anything else
Why is an adjusted model used?
How to identify and account for potential confounding
