RCT

Question 1

5 Key Characteristics

Answer 1

Manipulation - defined by exposure, but investigator controls the exposure
Comparison Arm
Randomization
Double Blinded
Placebo Controlled

Question 2

Randomization

Answer 2

Useful bc patients then comparable across all characteristics - known and unknown to reduce residual confounding. Limits possibility of bias and confounding. Each pop generalizable to same pop

Question 3

Random Sampling and Randomization

Answer 3

NOT the same. Randomization is for RCT and not observational studies

Question 4

Intent to Treat

Answer 4

All ppl randomized must be analyzed, or else the randomization is broken

Question 5

3 Types of Endpoints

Answer 5

Clinically meaningful - death, morbidity improvements, etc
Surrogate endpoints - EF for CHF instead of mortality, etc
Composite Endpoints - death and MI together - increase event rate

Question 6

Event Rate vs. External Validity

Answer 6

Need enough events to be significant, so may need sicker patients. This may reduce external validity/generalizability, so trade-offs

Question 7

Number Needed to Treat

Answer 7

Inverse of absolute difference (NNT to save 1 pt)

Question 8

Publication Bias

Answer 8

Big Pharm publish studies that work and bury the rest

Question 9

Huge Trial Sample Sizes

Answer 9

Often used to cover up small effect size

Question 10

Unit of Analysis in Meta Analysis

Answer 10

Study

Question 11

Goal Finding of Meta Analysis

Answer 11

Looking for summary statistic of effect size (using weighted effect sizes of the studies)