Environmental and Occupational Epidemiology

Question 1

Is there an interest in causation?

Answer 1

Yes; if yes, how important

Question 2

Types of association?

Answer 2

• no association
• artificial association: die to chance or bias
• indirect association: confounding
• direct association: causal

Question 3

Observational studies?

Answer 3

-cohort (prospective and retrospective)
-case control
cross-sectional

Question 4

analytical observational studies do what?

Answer 4

observe events as they happen (or did happen) w/ no active role in events

• measure the exposures, outcomes, & potential confounders that might effect the relationship
• statistical tests

Question 5

Another name for cohort studies?

Answer 5

• incidence studies
• longitudinal studies
• prospective studies

Question 6

Types of errors?

Answer 6

• random error (chance)

- systematic error

Question 7

Random errors?

Answer 7

more likely to cancel out as the measurements increase

Question 8

Confounding variable?

Answer 8

a variable that distorts the association between the exposure and outcome
-either strengthen or weaken the association; Type I/II error

Question 9

What makes a confounding variable?

Answer 9

• must be independently associated w/ the outcome
• must be associated w/ the exposure
• must not be in a causal pathway between exposure and disease

Question 10

Alcohol leads to (arrow) lung cancer. What’s the confounding variable?

Answer 10

Smoking

-correlated w/ alcohol and a risk factor for lung cancer even if you don’t drink alcohol

Question 11

Controlling confounding factors? (at design stage)

Answer 11

• restricting
• randomization
• matching

Question 12

Controlling confounding factors? (at analysis stage)

Answer 12

• stratification

- multivariable adjustment

Question 13

Randomization?

Answer 13

an attempt to evenly distribute potential (unknown) con founders in the intervention and control groups in a randomized controlled trial

Question 14

Restriction?

Answer 14

prevents confounding, but may be reduce the study size reducing statistical power
Ex: Concern that maternal smoking may be a confounder in associations between childhood lead exposure and reduced IQ. Restrict study population to kids w/ non smoking moms

Question 15

Multivariable adjustment?

Answer 15

can adjust (control) for the effects of many variables simultaneously

• linear regression
• logistic regression

Question 16

Selection bias?

Answer 16

selective differences between the comparison groups that distorts the relationship between exposure and outcome
-Ex: healthy worker effect; only picking ppl who are considered healthy to work or be in the study

Question 17

Information bias?

Answer 17

measurement bias; Investigators who know exposure status may be more or less likely to ascertain the outcome (diagnostic bias)
-to fix this, blinding of investigator and subjects

Question 18

Controlling bias?

Answer 18

selection bias and measurement bias must be controlled at the design stage

Question 19

How to limit selection bias?

Answer 19

Representative sampling from the same population

Question 20

How to limit measurement bias?

Answer 20

blinding

Question 21

Strengths of cohort studies?

Answer 21

• good when exposure is relatively rare
• can look at multiple outcomes from a single exposure
• can minimize measurement bias

Question 22

Weaknesses of cohort studies?

Answer 22

• inefficient for rare outcomes
• long time to complete
• expensive
• risky following people at risk

Question 23

Weaknesses of cohort studies?

Answer 23

• inefficient for rare outcomes
• long time to complete
• losses to follow up
• expensive
• potential ethical issues. may be a risk

Question 24

Case control studies?

Answer 24

a type of observational study commonly used to look at factors associated with diseases or outcomes
-one group that has the disease (cases) are tested while another group w/o the disease (controls) are given the same exposure

Question 25

What is the measure of association for case control?

Answer 25

odds ratio (AD/BC)

Question 26

Strengths of case control studies?

Answer 26

• useful for diseases that are rare, have long induction periods, like cancer, and to explore a wide range of exposures
• cheap
• quick
• safe

Question 27

nested case control studies?

Answer 27

select cases and controls for the study

-the study is “nested” w/in the cohort

Question 28

Recall bias?

Answer 28

cases tend to closely scrutinize their past history looking for explanations for their illness
-the disease status of subjects affects their likelihood of reporting the exposure.
Ex: a patient with cancer may be more likely to recall being a smoker

Question 29

How to avoid recall bias?

Answer 29

• avoid using records of exposures prior to disease
• choose new incident cases (for better memory,
• choose appropriate control group (those w/ a different disease not related to the exposure)
• blind study subjects
• verify exposure from another source (spouse, friend, etc.)