Lecture 7, part 2 Flashcards
What are sources of exposure info?
Pre-existing records
Self-report
Physical/biological observations
Examples of pre-existing records
Medical records
Insurance records
Employment records
Vital stats
Tax or property records
Advantages of pre-existing records
Inexpensive
Relatively easy to work with
Usually unbiased since data were collected for non-study purposes
Disadvantages of pre-existing records
Exposure info may not be precise enough to address the research question
Records frequently do not contain data on potentially confounding factors
Examples of self-report
Interviews
Surveys
Diaries
Advantages of self-report
Opportunity to question subjects on as many factors as necessary
Good for collecting info on exposures not routinely recorded
Disadvantages of self-report
Rely on participants’ ability to know and recall info
Potential response bias due to stigma, response expectations, observation, etc.
Participants may not be sufficiently aware of their exposure status
Examples of physical/biological observations
Physical exams
Lab tests of biological specimens
Environmental monitoring
Advantages of physical/biological observations
Biomarkers of exposure often more objective
Allow ascertainment of internal dose/exposure
Higher quality data
Disadvantages of physical/biological observation
Can be invasive and expensive to collect
Appropriate lab tests may not be available
May be subject to bias if interpretation or clinical observation is influenced by exposure
______of the exposure is crucially important
Timing
Relevant etiologic period defintion
Time during which an exposure will affect dz occurrence
How should study design be grounded?
In knowledge of natural hx of dz as well as hypothesized biological mechanism
What should be considered in terms of f/u?
Length required to observe the outcome of interest
Induction period definition
Interval between causal action and dz initiation
Latent period definition
Interval between dz onset and detection
What should be considered to determine the length of f/u?
Minimum latency
Maximum latency
Minimum latency considerations
How soon can dz result after exposure has first occurred
Maximum latency considerations
Once exposure has ceased, when does the alteration in risk disappear (if at all?)
Characteristics of good measurement of outcome
Should be as precise and objective as possible
What types of measurements of outcome are possible?
Categorical
Continuous
Dichotomous
What should be be done in measuring outcome when the measures are weaker/more subjective?
Use triangulation, or supplement
Sources of outcome data
Self-report
Employee records
Medical/pharmacy/insurance records
Physical exams/blood testing
Death or dz registries
Multiple sources
Definition of information bias
Systematic error in the measurement of the exposure and/or outcome
Key characteristics of information bias
Occurs after participants have entered study
Pertains to how data are collected
Can bias results away or toward the null
Can occur in any type of study design
What is the most common type of information bias?
Misclassification bias
