Final Flashcards
(36 cards)
What is confounding?
Confounding is a potential source of bias in epidemiologic studies
serves as an antecedent of both the exposure and the outcome, third variable distorts the true association between the exposure and the disease
What is effect modification?
The effect of the exposure on the outcome differs for varying levels according to presence of a third variable; related to biologic differences between two groups
How do you deal with confounding and effect modification?
In study design: Confounding - restriction, matching, randomization
In analysis: stratification for both confounding and EMM
Preventative Association
Odds Ratio or Risk Ratio is less than 1, meaning the exposure prevents the outcome
Positive Association
Odds Ratio or Risk Ratio is greater than 1, meaning the exposure leads to greater risk of delevoping the outcome
What is selection bias?
Selection bias occurs when some potential subjects are more likely to be included in the study than others (recruiting patients, can bias OR towards or away from the null)
What is information bias?
Information bias occurs when data is collected differently across subjects (recall bias, outcome misclassification)
What are the two types of outcome misclassification?
Non-differential: probability of individuals being misclassified is equal across all groups of the study
Differential: probability of being misclassified differs between groups in a study
What are the steps of an outbreak investigation?
- Prepare for fieldwork
- Confirm the existence of an outbreak
- Verify diagnosis (define a case)
- Find cases, tabulate and orient data: person, place, time
- Condust descriptive epidemiology
- Formulate and test the hypothesis
- Implement and evaulate control measures
- Initiate or maintain surveillance
- Communicate findings
What are the three parts of the epidemiologic triangle?
Host, Agent, Environment
Formula for incidence of EXPOSURE in an ecological study
(a+b) / [(a+b)+(c+d)]
Formula for incidence of DISEASE in an ecological study
(a+c) / [(a+c)+(c+d)]
What are the advantages of an ecological study?
Large-scale trends through time
Hypothesis generating
Can use aggregate, environmental, or global measures
Data is collected by governments and them available for multiple uses
Inexpensive, simple, fast
What are the disadvantages of an ecological study?
Cannot make casual interpretations
Impossible to control for potential confounder bias
Ecological fallacy: association that exist at a group level may not exist at an individual level
What is the ecological study design?
Examines a group as a unit of analysis to produce correlational measures (trends in population, never individual)
What is the cross-sectional study design?
Data is collected at one point in time and includes exposures and outcomes
Formula for calculating the prevalence of disease in the exposed persons vs unexposed
a / a+b
c / c+d
Formula for calculating prevalence of exposure in persons with disease vs without disease
a / a+c
b / b+d
What are the measures of frequency for a cross-sectional study?
Prevalence of outcome: (a+c) / N
Prevalene of exposure: (a+b) / N
What is the measure of association for cross-sectional studies?
Prevalence odds ratio: ad / bc
POR of outcome: (a/b) / (c/d)
POR of exposure: (a/c) / (b/d)
What are the advantages of a cross-sectional study?
Permit determination of population characteristics
Useful for diseases that are prevalent and of long duration
Good for generating hypothesis and/or initially testing one
What are the disadvantages of the cross-section study design?
Not useful for diseases that are rare and or short duration
Temperoal relationship between expsoure and disease in unclear
What is the case-control study design?
Subjects are selected from a population based on disease status, then it is worked backwards to find who was exposed and unexposed - exposure in determined retrospectively
compares cases to control with respect to exposure
What is the unit of observation and analysis of case-control study?
Individual
