FOM 7.4.3 Flashcards
How does increasing study size affect chance for random error?
As sample size increases chance of random error goes down
How does increasing study size affect chance for systematic error?
It does not affect it
What is selection bias?
Distortion in effect estimate resulting from manner in which people were selected or selected losses
What is an example of selection bias?
Study to see how brushing teeth can affect cavity. Study uses volunteers. This is selection bias because volunteers tend to be more hygienic than the normal population so this would not be accurate picture
What is internal validity?
True situation of study population
What is external validity?
How does our group apply to general population
If patients develop side effects during a study and patients want to quit what happens?
This creates selection bias in the data
What is information bias?
Measurement of exposure or disease is systematically inaccurate.
What is misclassification bias?
result of error in design. putting patients in the wrong cell
What is recall bias?
The inability to recall perfectly. When looking at babies with birth defects and interview parents, parents with a baby that has birth defects are more likely to remember possible exposures to harmful risks
What is interviewer bias?
The interviewer may probe one person harder for information than another
How can you reduce information bias?
Blinding, visual aids, validate exposures, and standardized data form
What is confounding bias?
Third variable that distorts exposure/ outcome relationship
What is association?
Identifiable relationship between exposure and disease but there is no directionality
What is causation?
Implies that there is a mechanism from exposure to disease and these tend to be multifactorial
What are the titles of these two?
Strength of association - strong association is more likely causal
Consistency - results should be able to be repeated
What is the title of this?
Specificity - requires that cause lead to single effect
What is this title
Temporality - cause precedes effect
What are the titles of these?
Biological gradient - Presence of a monotonic (unidirectional) dose-response curve; increasing trend in disease frequency with increasing dose.
Biological plausibility - Hypothesized relationship between exposure and health outcome makes sense in context of current biological knowledge.
What is the title of this?
Cessation of exposure
How to practice EBM?
