Epidemiology: 4) Exposure - Definition Of Groups

Front 1

Definition of groups

Back 1

Appropriate definition, classification and measurment of expsure. Analogous to randomization in CT

Front 2

Examples of exposure groups

Back 2

Genetics
Age
Demographics
Lifestyle
Geographoc
Environmental
Disease
Meds
Devices
Education
Policies

Front 3

6 - Exposure checklist

Back 3

1) External E-
2) Internal E-
3) Zero time
4) Classification
5) Measurement
6) Change w\ follow-up

Front 4

Internal E-

Back 4

cohort is single group of subjects with uniform procedures.
Classification (E+ or E-) done at same time (t0).
Can be matched or entire E-

Front 5

Increasing comparability of E- group

Back 5

try and restrict E- to subjects most similar to E+ (some covariate that makes them very similar to E+)

Front 6

Using alternative exposure for E-

Back 6

careful of counfounding. Can show erranous protective effect. Should try and use plocebo or no exposure.

Front 7

Temporal E- comparisons: Natural crossover

Back 7

uses person time of exposure compared to person time of unexposed.
t0 should be at initial intervention..not follow/up

Front 8

External E-

Back 8

Group from a different population.
Different procedures to identify and collect data.

Front 9

3 Fundamental weakness of E-

Back 9

1)Effect of different procedures cannot be separated from the effect of E.
2)Entire group risk of D may be different (healthy worker effect)
3)Usually E+ data more complete

Front 10

External E - Comparison at Different institutions

Back 10

Must take care that patients are treated exactly the same.

Front 11

External E: Control group membership influenced by D

Back 11

E- Should have same chance of getting exposure an getting outcome as E+

Front 12

Zero time

Back 12

inception cohort - patients should be identified at an early and unambiguous point in the course of their disease. It is a challege for E-

Front 13

Reference category

Back 13

reference group with which others are compares in calculating measured effect (often lowest quintile of exposure).

Front 14

3 ways to Constructing Categories

Back 14

1)Intrinsically meaningful divisions
2)Percentiles (if no a priori H)
3)Don't torture the data