Epidemiology Flashcards
Differentiate
epidemiologist vs doctor
- Doctor: sees 1 person at a time, very individual level
- Epidemiologist: disease on a population level
Define
Epidemiology
Study of the distribution and determinants of health & disease in human populations
When is epidemiology better for correlation or causation?
Better for correlation bet. Adverse effect & substance
Better for causation, depending on how complex epidemiological study is
sinu itech
Karl Popper
Repeated observations that event A is followed by event B
* Hypothesized proposition that A causes B
What assumption do we have in epidemiology?
Theory of Causation
Event A causes Event B
DO NOT RELY ON IT!
THEY R TENTATIVE!
may be disproved by further testing
Which type of study establishes correlation, and which establishes causation?
Ecological study establishes correlation
More human studies establish causation
sinech itey
Sir Austin Bradford Hill
- Pioneered randomized clinical trial
- Demonstrated connection bet cigarette smoking & lung cancer
List
Criteria for Plausibility
5
- Proper temporal relationship (absolutely required)
- Consistency of association
- A large effect size
- Positive dose-response relationship
- Biological plausibility
Explain
Proper temporal relationship
Exposure must be before the disease
* Need a logical timeline
* If this doesnt exist → scrap assumption
Explain
Consistency of association
Same time frame, same thing happening every time
Is epidemiology statistics heavy? Why?
yes
bc we need a large effect size
bc we r looking at the population level of the disease!
Explain
Positive dose-response relationship
↑ exposure = ↑ likelihood of getting disease/adverse effect
Explain
Biological plausibility
Must make sense visavis exposure, dose, etc.
Where toxicological studies happen
List
Epidemiological Studies
3
- Descriptive studies
- Correlational or ecological studies
- Etiologic or analytical studies
Describe
Epidemiological Studies
Largely observational study
- Not like clinical tests where you manipulate a variable
- “Variable manipulation” is in selecting your cohort
- Statistics lang talaga, kasi looking at outcome based on your selected participants