Interpreting Epidemiological Findings (1) Flashcards
What are the terminology of Bradford hill criteria?
association vs causation
- criteria used to enable us to infer causation from observational and interventional methods
Strength Consistency Specificity Temporality Biological Gradient Plausibility Coherence Experiment Analogy
What does strength mean?
A stronger association increases the confidence that an exposure causes an outcome
What does consistency mean?
Consistent findings across settings tends to rule out errors or fallacies that might befall between studies
What does specificity mean?
describes an association between specific causes and specific events
If not specific – doesn’t mean invalidity! (esp. with multifactorial issues)
Specificity can be informative when present but its absence convey very little
What is temporality?
Measuring over time
the exposure should commonly precede the outcome
Hard to assess with cross sectional studies
What is biological gradient?
A dose-response effect is a compelling argument for causality
e.g. consuming more cigarettes increases likelihood of disease
What is plausibility?
Relationship should be biologically plausible where the science is understood
What is coherence?
The association ought to be consistent with the existing theory and knowledge
What is experiment?
Evidence from experimentation should be supportive of the proposed link
What is analogy?
Drawing upon analogous findings, we many make inference on the relationship
e.g. rubella causes deafness so you could consider that other viruses can cause similar issues
What is bias?
Any trends in the collection, analysis, interpretation, publication or review of data that can lead to conclusions that are systematically different from the truth
What are the three types of bias?
Selection
Information
Compounding
What is internal validity?
The extent to which findings accurately describe the relationship between exposure and outcome in the context of the study
-when association truly exists within study participants
What is external validity?
Generalisability
-when association observed within a study can be extended to the wider population
What can selection bias also be called?
What is selection bias?
How can you avoid selection bias?
Berkson’s bias – hospital based - case control study w/controls selected among the hospital’s patients
Healthy worker effect – active workers are more likely to be healthy than those who have stopped working (occupational studies)
Occurs when individuals or groups in a study differ systematically from target population leading to a systematic error in an association or outcome - When participation in a study is associated with the exposure and the outcome
to miminimse:
- Controls representative of target population
- Minimise non-response (i.e declining to participate)
- Compare respondents with non respondents
What is information bias?
Due to a number reasons we have misclassification of the exposure or disease status (or both)
Can include: interviewer, recall, response, and diagnostic bias
Due to ill-defined variables/flaws in data collection
Flaws in data collection
- Interviewer bias
- Recall bias
What is interviewer bias?
Interviewers ask about exposure status -
= Prevented by interviewer being blind/strict protocol when collecting data
Might be more thorough in interviewing on a related topic if disease status is known
e.g. smoking history in those with lung cancer
What is recall bias?
When people become ill they are more likely to think hard about how past experiences have affected them - – bias when participants self report exposures between cases and controls
= Prevent with objective ways to assess exposure
e.g. more likely to remember a brief period of smoking many years ago if you have been diagnosed with lung cancer
What is non-differential misclassification?
When exposure status is misclassified but equally among controls and cases
Errors in determining outcome in controls and cases
Bias always towards the null
What is differential misclassification?
Misclassification of exposure is not equal between cases and controls (disproportionate)
-Bias towards or away from the null
2 types of misclassification
non differential and differential
Key points
a. An epidemiologically sceptical approach requires us to rule out chance, bias (including
confounding) before we infer a real relationship – whether causal or not.
b. When assessing the validity of a study, we examine both its internal validity and its
generalisability (external validity) when interpreting findings.
c. There are nine Bradford Hill criteria that can be used to infer causality or absence
thereof. Their application is not a precise science but requires judgment and critical
thinking.
d. In addition to confounding, there are two major groups of bias: selection bias and
information bias.
e. Association does not equal causation.
which type of bias is described as Misclassification of the exposure, outcome or both.
information bias
the following are all types of which bias? Recall bias Response bias Interviewer bias Diagnostic bias
information bias
Healthy worker effect is a type of which bias
Selection bias
Where cases and controls equally mis-report their exposures, this would likely be described as:
Non differential misclassification
if cases or controls were likely to report exposures differentially then it would:
differential misclassification.
Non differential misclassification always leads to an association that is biased:
towards the null
Where cases and controls unequally mis-report their exposures AND in such a way that the overall consequence is an association that tends away from the null, this is best described as:
information bias
term to describe:
The extent to which findings accurately describe the relationship between exposure and outcome in the context of the study
Internal validity
External validity is synonymous with:
Genralisability
A study, which is designed very well, and implemented equally well - but which is carried out a single teaching hospital in London can be said to have high:
internal validity
The findings of a study which is designed well but which loses more participants than expected are submitted for peer review. Which of the following might be called into question by the peer reviewers?
internal validity lacking
The findings of a study which is designed poorly but implemented exceptionally well, is submitted for peer review. Which of the following might be called into question by the peer reviewers?
internal validity
external validity
internal validity - due to poor design
An international multicentre randomised controlled trial with triple blinding is reported in the literature. On the basis of this sentence alone, which of the following is more likely?
Good internal validity
Good external validity
good external validity
which type of bias are the following types of:
Healthy worker effect
Non-response bias
Selection bias
What is the Healthy Worker Effect
how employed persons are generally more healthy than those who are unemployed.
What is response bias
the tendency of participants to respond to questions in a way that is more socially acceptable.
What is non-response bias
the process through which people who do not respond are systematically different to the people who do respond. This is a type of selection bias.
describe the lake woebegone effect
- referred to as illusory superiority.
eg. If you ask a class of students how good their driving is, most of them will believe themselves to be better than average.
describe the hawthorne effects
The precise definition is controversial
- 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!
