Lecture 18/19 Flashcards
Bias and Diagnostic Test Summary (27 cards)
Selection Bias
occurs when the measure of association calculated from your sample population is meaningfully different from the measure of effect which would have been estimated if you had data from all eligible subjects in the source population
Sources of selection bias
- participant loss to follow up
- selection processes
- survival bias: occurs when participants are not able to be selected and enrolled because they have died
- participation
- participant nonresponse
Reducing selection Bias
- Maximizing follow-up in longitudinal studies
- maximal participation
- minimizing participant nonresponse
- selecting incident cases in case control studies
Study Design Stage
- representing sampling strategies
- random sampling (simple random sampling, stratified random sampling, cluster sampling)
- nonrandom sampling (convenience sampling, voluntary response sampling, snowball sampling)
Analysis Stage
- selection probabilities
- sensitivity analysis
Internal validity
whether the study results calculated from the sample population are valid (correct) with respect to the source population
- selection Bia is related to internal validity
External validity (generalizability)
related to whether the results calculated from the sample populations are valid with respect to some target population
- study that lacks internal validity cannot have external validity
Information Bias
resulted from the incorrect measurement of exposure and/or outcome data based on information we’re getting from participants
Sources of information bias
- incorrect measurement devices or laboratory assays
- incorrect self-reported data: problems with data collection / participation barriers
- investigator error: observer bias
- data management error
misclassification of the exposure
individuals are only moving between cells A and B and between cells C and D in 2x2 table
Misclassification of the outcome
individuals are only moving between cells A and C and between Cells B and D in a 2x2 table
misclassification of the exposure and the outcome
individuals can move from their true cell in an R x C table to any of the other cell
non differential misclassification
occurs when the extent of misclassification of the exposure or outcome does not depend on the status of the other variable
non-differential exposure misclassification
exposure is misclassified, indecent of outcomes status
non differential outcome misclassification
outcome is misclassified, independent of exposure status
differential misclassification
occurs when the extent of misclassification of the exposure or outcome DOES depend on the status of the other variable
differential exposure misclassification
exposure is misclassified to a different degree for those who have the outcome than it is for those who do not have the outcome
differential outcome misclassification
outcome is misclassified to a different degree for those who are exposed than it is for those who are not exposed
Analysis Stage
- validation studies
- sensitivity analyses
study design stage
- masking
- validated, ideally gold standard exposure and outcome assessment techniques
- strategies to suppose accurate self report
- strategies to support accurate recall
- multiple, independent assessors
- validating studies
screening tests
used in individuals who appear to be healthy but may be at-risk of disease identity or rule out the possibly of disease
- ex) mammogram, colonoscopy
diagnostic tests
are used in individuals with abnormal signs or symptoms (or individuals with a positive screening test) to either establish or rule out disease
ways to quantify test performance
sensitivity, specificity, PPV, and NPV
sensitivity
proportion of those with the outcome or exposure who are correctly classified
