Epidemiology: 1) Study Designs

Front 1

Classes of Mistakes

Back 1

External Validity
Internal Validity
Execution

Front 2

External Validity

Back 2

Narrow research question not Clinically relevant. Inappropriate population or outcome

Front 3

6 issues of Internal Validity

Back 3

Fails to address narrow question.
Selection bias
Information bias
Misclassification
Confounding
Power
Surrogate Endpoint

Front 4

Execution

Back 4

missing data
inaccurate assays
data entry error

Front 5

Synonyms: Exposure

Back 5

Etiologic factor
Treatment, stimulus
Risk factor
Independent Variable
Prognostic Variable

Front 6

Synonyms: Disease

Back 6

health status
response variable
endpoint
dependent variable
outcome

Front 7

Primary Study designs

Back 7

Clinical Trial
Cohort
Case Control

Front 8

Clinical Trial:
Risk
Risk Ratio
Risk Difference

Back 8

Risk (E+): p1=a/N1
(E-): p0=b/N0
Risk Ratio: p1/p0
Risk Difference: p1-p0

Hint 8

Use vars from data layout

Front 9

Strengths of RCT

Back 9

Random
Strong Inferential Validity

Front 10

Limits of RCT

Back 10

Impractical:
rare disease
long follow-up
difficult to manipulate
Unethical
Impossible

Front 11

Cohort:
Risk
Risk Ratio
Odds Ratio
Risk Difference

Back 11

Risk (E+): p1=a/N1
(E-): p0=b/N0
Risk Ratio: p1/p0
Odds Ratio: (ad)/(bc)
Risk Difference: p1-p0

Hint 11

Use vars from cohort data layout

Front 12

Cohort (variable follow-up)
Incidence Rate
Rate Ratio
Rate Difference

Back 12

Incidence rate (E+) = I1 = a/T1
Incidence rate (E-) = I0= a/T0
Rate Ratio: IRR = I1/I0
Rate Difference: IRD = I1-I0

Front 13

Strengths of Cohort Study

Back 13

"Natural expriment"
Easy to understand

Front 14

Limits of cohort study

Back 14

Other differences betwen E+, E-

Impractical:
rare disease
long follow-up

Front 15

Strengths of Case Control

Back 15

Logistically Effecient
Good for rare disease
Efficient for sporadic exposures

Front 16

Limits to case-control

Back 16

More susceptible to bias
Difficult to understand
Unless nested: no disease rate
cannot always determine E after D
Inefficient for infrequent exposures
Differences between E+ and E-

Front 17

Types of Ecological Studies

Back 17

Geographic/Static exposure
Temporal or changing exposure

Front 18

Limits to static ecologic studies

Back 18

Temporal precedence
Variation in other factors
confounding
Ecological Fallacy

Front 19

Ecological studies best when:

Back 19

E is dominant risk factor for D
E is manipulated externally (change in policy) or other temporal variation

Front 20

Secondary designs

Back 20

partial cohort: E+ only
partial case-study: D+ only
Cross sectional

Front 21

Why use E+ only

Back 21

Very rare D and information needed quickly. Can be augmented with external controls.
Often use pre/post design

Front 22

good and bad about D+ only?

Back 22

limit of no D- group.
good if need info quick.

Front 23

Limits of cross-sectional study

Back 23

temporality:
E is altered by D+
Survival post D+ is affected by E

Front 24

Inception Cohort

Back 24

all individuals assembled at a given
point based on some factor, e.g. where they live or work