Lecture 7

Question 1

what are 2 main kinds of analytic?

Answer 1

1. Experimental

2. Observational

Question 2

2 important criteria for validity

Answer 2

1. Validity of comparison
   • possible to make a “fair” comparison
   • Confounded = if not comparable
2. Validity of measurement
   • info is collected as completely & accurately as possible
   • information bias

Question 3

What is it called if 2 groups are not comparable?

Answer 3

Confounding error

Question 4

What are important aspects of a study that are important for validity of comparison?

Answer 4

1. control group necessary
2. control group nearly identical to tx group
3. Randomization **

Question 5

Validity of measurement

Answer 5

• Blinding

Question 6

Advantage/disadvantage to experimental studies

Answer 6

Advantage
+ best for causal effect
+ faster results
+ practical for some clonal dz.’s

Disadvanage

• $$
• ethical
• lack participants
• non-compliance / drop-out

Question 7

Observational studies

Answer 7

• • no control over exposure **
    1. Ecological study
    2. Cross-sectional
    3. Cohort study
    4. Case-control study

Question 8

Ecological study

Answer 8

Done when info is known on ENTIRE GROUP, but NOT individuals

• Ex: country data

Question 9

Cross-sectional study

Answer 9

Dz & exposure status at SINGLE POINT in time
• Cannot measure incidence (ONLY PREVALENCE)
• Cannot know which came first

Question 10

cohort study

Answer 10

LONGITUDINAL
• at risk individuals are followed
• Measure Cumulative incidence OR incidence rate
• Short follow up period makes it more feasible!

2 types

1. Prospective
2. Retrospective

Question 11

Cohort advantages/ disadvantages

Answer 11

Advantages
• common dz
• rare exposure
• studying multiple outcomes

Disadvantages
1. $$$$$
2, If the reason an individual leaves is related to the outcome, then bias may result
- Censoring (Lost to follow-up)

Question 12

Case-control study

Answer 12

“efficient” means of performing cohort study

• Case = INCIDENT cases only
• controls:
  • selected independent of exposure
  • IF they had developed the dz, they would be in the study (i.e. same cohort)

• Calculate:
- Exposure Odds ratio

Question 13

Exposure Odds Ratio

Answer 13

Cases Exposed / Cases NOT exposed
———–Divided by ——————–
NONcases Exposed / NONcase NOT exposed

• cannot be used directly for causal inferences
• mathematically equal to disease odds ratios, which can be used for causal inferences

Question 14

Disease Odds ratio

Answer 14

Cases Exposed / NONCases exposed
———–Divided by ——————–
cases UNExposed / NONcase UNexposed

Question 15

Case-control advantages/ disadvantages

Answer 15

Advantages
+ relatively inexpensive
+ short, if retrospective
+ examining Lg # of risk factors

Disadvantage

• based on O’s recall of info (cannot verify accuracy) = info bias
  - Cases > non-cases to recall exposure