interpreting epidemiological findings Flashcards
what are Koch’s postulates used for?
infer whether a particular pathogen is responsible for a particular disease
what do Koch’s postulates state?
microorganism must be found in abundance in all organism suffering from the disease but not in healthy organisms
must be isolated from a diseased organism and grown in pure culture
cultured microorganism should cause disease when introduced into a healthy organism
must be re-isolated from the inoculated diseased, experimental host and identified as being identical to the original specific causative agent
why do Koch’s postulates sometimes fail?
unethical to re-introduce microorganisms into humans with the intention of causing infection
what does the Bradford-Hill criteria allow?
infer causation from observational and interventional methods
what are the Bradford-Hill criteria?
strength
consistency
specificity
temporality
biological gradient
plausibility
coherence
experiment
analogy
what does ‘strength’ mean with reference to the Bradford-Hill criteria?
strong relationship between an exposure and an outcome makes it more likely that there may be a causative element
what does ‘consistency’ mean with reference to the Bradford-Hill criteria?
stronger association increases confidence that an exposure causes an outcome
consistent findings across settings can rule out errors/biases that may take place in 1 or 2 studies
what does ‘specificity’ mean with reference to the Bradford-Hill criteria?
specificity - describes association between specific causes and effects
specific disease arising amongst specific workers supports argument for causality
what are the criticisms of ‘specificity’ in the Bradford-Hill criteria?
Hill concedes that lack of specificity doesn’t necessarily invalidate the relationship
specificity is difficult to account for in multifactorial issues as many factors interact
therefore specificity is informative when present, but absence of specificity does not mean much
what does ‘temporality’ mean with reference to the Bradford-Hill criteria?
exposure must commonly precede an outcome
which study designs are better to assess ‘temporality’ with reference to the Bradford-Hill criteria?
cross sectional studies - look at many people at same point in time to assess presence of exposure and outcome, cannot assess which preceded which
longitudinal studies can assess temporality better
what does ‘biological gradient’ mean with reference to the Bradford-Hill criteria?
dose-response effect in the ‘right’ direction suggests a relationship between an exposure and outcome
(e.g. higher risk of CVD caused by smoking - the greater the number of cigarettes smoked, the greater the risk)
what are the criticisms of ‘biological gradient’ in the Bradford-Hill criteria?
quantification is difficult
when present it is suggestive but absence of biological gradient has little value
what does ‘biological plausibility’ mean with reference to the Bradford-Hill criteria?
relationship is biologically plausible where the science is understood
how does ‘biological plausibility’ (Bradford-Hill) change over time?
advances in scientific understanding allows previously implausible relationships to be explained
what does ‘coherence’ mean with reference to the Bradford-Hill criteria?
association is consistent with existing theory and knowledge
slight overlap with biological plausibility
what does ‘experiment’ mean with reference to the Bradford-Hill criteria?
evidence from experimentation is supportive of proposed link
what is the issue with ‘experiment’ (Bradford-Hill criteria)?
experimentation within ethical constraints is difficult
what does ‘analogy’ mean with reference to the Bradford-Hill criteria?
analogous findings may provide inferences about relationships
e.g. rubella causes deafness, eye problems, heart problems - by analogy other viral illnesses contracted during pregnancy may cause similar pathology
what is the issue with ‘analogy’ (Bradford-Hill criteria)?
depends on existing knowledge
what is the purpose of validity and bias?
determine whether the results of a study are relevant and should be trusted or not
what is internal validity?
association exists amongst a group of patients within a study
observed results are due to true cause-and-effect
what is external validity/generalisability?
association determined in source population (internal validity) also applies outside of source population - in different places or times etc.
what is the importance of external validity?
determines real-life impact of a finding
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 is the difference between random and systematic error?
random:
- occasional underestimation/overestimation due to chance
- interferes with precision
- errors can be more or less cancelled out with a large enough sample size
systematic:
- consistent underestimation/overestimation of measurements
- significantly alters distribution pattern
- introduces bias
what are the 3 types of bias?
selection bias
information bias
confounding
what is selection bias?
for results to be valid sample must be representative of study population
however, an individual’s chance of being included in the study sample may be related to both exposure and outcome
causes biased estimate of association between exposure and outcome
what is Berkson’s bias?
type of selection bias
hospital based case control study where controls are selected amongst the hospital’s patients
what is the healthy worker effect?
type of selection bias
active workers are more likely to be healthy compared to those who have retired or stopped working
what are 3 ways of minimising selection bias?
select controls which are representative of the study population in terms of exposure to the facts on the study
minimise non-response (if people decline to participate it could introduce bias)
compare those included in the sample with those who declined to respond and explore whether there are any systematic differences.
what is information bias?
misclassification of exposure, disease status or both
why can information bias occur?
study variables not properly defined
flaws in data collection
what is interview bias?
common flaw in data collection that causes information bias
interviewers ask individuals about their exposure status but may be more or less thorough depending on their outcome
causes misclassification of exposure status and eventually biased odds ratio
how can interview bias be prevented?
interviewer does not know individual’s disease status
collection process has
been carefully standardised so that interviews follow a strictly defined protocol when collecting participant data
what is recall bias?
causes information bias
refers to differentially inaccurate recall of past exposure between cases and controls
patients are more likely to recall a brief exposure if they have developed the outcome
how can recall bias be prevented?
use objective ways to assess exposure
e.g. medical records, biomarkers
what is non-differential misclassification?
exposure status is misclassified but equally so amongst cases and controls
e.g. all participants have trouble remembering that exposure status - has nothing to do with their disease (not recall bias)
or - there are errors in determining outcome but they occur equally in exposed and non-exposed individuals
what type of bias does non-differential misclassification result in?
odds ratio biased towards null
what is differential misclassification?
errors in determining an individual’s exposure status occur unevenly amongst cases and controls
or - errors in the diagnosis of the disease occur unevenly amongst exposed and unexposed individuals
what type of bias does non-differential misclassification result in?
biased odds ratio
cannot predict if it is biased towards or away from null