Overview of Study Designs Flashcards
In this type of study, we determine an association between exposure risk and disease of interest, usually NOT causation. **Association is not equal to causation.
Observational Study
A lack of random error, which leads to a false association between exposure and outcome just by “chance”.
Precision
Lack of bias and confounding.
Validity
This is an error committed by the investigator in the design or conduct of a study that leads to a false association between exposure and outcome
Bias
A situation in which the relationship between the exposure and outcome is distorted because of the association of exposure with other factor(s) that also influence the outcome under study.
Confounding
_____ relates to: The larger the association, the more likely the exposure is causing the disease. For example, the relative risk of lung CA in smokers vs non-smokers is 9 times higher; with further evidence that a relative risk of lung CA in heavy smokers vs non-smokers is even greater at 20 times higher.
Strength of Association
The investigator may have inadvertently selected the wrong sample of participants
Selection Bias
An error in the assessment process. For example, recall bias of someone who developed food poisoning and cannot remember the food items they ate over the past few days.
Information Bias
Name the 3 categories of bias
- Selection Bias
- Information Bias
- Confounding
The association is observed repeatedly in different persons, places, times, and circumstances. For example, smoking has been associated w/lung CA in at least 29 retrospective and 7 prospective studies among different population samples.
Consistency
_____ is linked to the notion that a single exposure should cause a single disease.
Specificity
causes a SPECIFIC disease
_____ means the causal factor must precede the disease in time
Temporality
word is defined as: have a relationship w/time; ex: temporary
______ is a DOSE RESPONSE relationship between exposure and disease. Persons who have increasingly higher exposure levels have increasingly higher risks of disease. Ex: lung CA death rates rise w/the number of cigarettes smoked. Some exposures might not have a “dose-response” effect but rather a “threshold effect” below which there are no adverse outcomes.
Biological Gradient
_____ and _____ means there is a biological or social model that exists to explain the association. The association does not conflict with current knowledge of the natural history and biology of the disease. For example: we know that cigarettes contain many substances which have been shown to be carcinogenic on their own.
Plausibility and Coherence
Investigator-initiated intervention that modifies the exposure through prevention, treatment, or removal should the result show less disease. Example: Smoking cessation programs result in lower lung CA rates.
Experiment
This refers to having a similar relationship being observed with another exposure and/or disease. Example: effects of Thalidomide and Rubella on the fetus provide _____ for effects of similar substances on the fetus.
Analogy
The 3 most common types of epidemiological study designs are: ____, with ____ having the highest level of causal inference, just below clinical trials
In order from weakest to strongest causal inference:
- Cross-sectional
- Case-control
- Cohort (strongest)
**This is an observational study in which exposure and disease are measured AT THE SAME POINT IN TIME in a given population. Cannot establish causation. Highly susceptible to reporting bias.
Cross-sectional study
Hint: The disease and exposure are CROSSING paths/overlapping at the same point in time
**This is an observational study in which persons with disease and persons without disease who had the same chance of exposure are studied.
Case-Control Study
Persons with disease (cases) and persons without disease (controls) are studied.
**With this type of observational study design, the disease occurs first then the investigator looks back (retrospectively) to determine exposures or risk factors.
Case-control study
Those with the disease (cases) are compared against those who were exposed but did not develop disease (controls)
**This is an observational study of persons exposed who develop and persons not exposed who develop the disease. Incidence can be calculated because the exposure comes FIRST and the entire population at risk can be included.
Cohort Study
**In this type of study, the exposure occurs first then the determination of disease occurs after the exposure.
Cohort Study
For rare diseases, a ____ study may be a poor choice because it may take too long for enough incidences of the disease to occur. Thus, rare diseases might be best studied using a _____ study.
Poor choice: Cohort
Better choice: Case-control
In a 2x2 Table, the _____ is across the top (x- axis), and the _____ is along the left (y-axis)
Disease across the top
Exposure along the left
These 2 types of study designs utilize a 2x2 table:
- Case-Control
2. Cohort
**With a case-control study, the measure of association is the ____
Odds Ratio (OR)
To calculate OR with the 2x2 table:
Multiply across the cells (A x B) and (C x D), then divide those 2 numbers (AxB)/(CxD)
With a cohort study, the measure of association is _____
Relative Risk (RR)
[A / (A+B)] divided by [C/(C+D)] is the calculation to determine ____
Relative Risk (RR) the measure of association for a Cohort Study
**If a risk in exposed group is greater than the risk in unexposed group, the RR would be:
> 1
**If the risk in exposed group is less than the risk in unexposed group, the RR would be:
<1
**Odds of exposure in cases (those who have disease) is greater than odds of exposure in controls (those who do not have the disease) would show an OR of ___
OR >1
**Odds of exposure in cases (those who have disease) is LESS than odds of exposure in controls (those who do not have the disease) would show an OR of ___
OR <1
This measures the likelihood the effect is a result of the exposure or by chance alone.
Confidence Interval
What would you look for in a CI to determine if the results are significant?
If the CI includes 1, then it is not significant. Ex: a CI of 0.3 to 3.5 (includes 1 within that range, so it is not significant)
For p-Value, we are stating there is only a 5% chance of the outcome occurring by chance alone, so we look for a p-Value of ____
0.05 or less represents significance
This test is considered the “gold standard” for a 2x2 table
Fisher Exact Test
Fisher gives you the EXACT results you need
GOLD FISH
This is the test of choice when the numbers in the 2x2 table are small (any cell less than 5)
Fisher Exact Test
small fish in a big pond
This test can be used when there are at least 30 subjects in the study and at least 5 in each cell of the 2x2 table
Chi-Square Test
There are many people in CHIna
How would you write an OR of 1.65 in a sentence?
The odds of having the disease is 65% higher if exposed
How would you describe an RR of 3.3 in a sentence?
The risk for disease is 3.3 times higher in those who were exposed.
