Interventional Study Design

Question 1

Interventional Retrospective

Answer 1

Post hoc analysis

Question 2

Interventional Prospective

Answer 2

Controlled trials

Uncontrolled trials

Question 3

Interventional

Answer 3

Investigator ASSIGNS TREATMENT

Question 4

Observational

Answer 4

Treatment is predetermined

Question 5

Ultimate goal of clinical trials

Answer 5

Establish causality
If X then Y
If you give the pt this antibiotic then their infection will resolve
If you do not give the antibiotic then their infection will not resolve

Question 6

Research Question

Answer 6

FINER
Feasible, interesting, novel, ethical, relevant
PRIOR to start of study

Question 7

Background and significance

Answer 7

Why is it worthwhile?
What is already known?
What is unclear?
What needs to be answered?
Is it beneficial?

Question 8

Placebo controlled trials indicated when:

Answer 8

No standard therapy available
Pts aware of placebo control and odds of receiving placebo
Placebo may be added to the standard therapy

Question 9

Effective Controls

Answer 9

Proper matching
Same appearance, smell, taste
Comparable dose, regimen, duration
Unmasking: accidental or deliberate (safety concern)

Question 10

Placebo effect

Answer 10

Some kind of intervention = real or preceived positive negative effects “nocebo”
Effects are also possible even if they know they are receiving placebo
Linked to EXPECTATIONS

Question 11

Hawthorne effect (observer bias)

Answer 11

Subjects enrolled in a study change their behavior in response to the study itself (exercise when they normal don’t, etc)

Question 12

Pygmalion effect (expectancy advantage)

Answer 12

Investigator’s expectations of an advantage in one group vs another affects subject response

Question 13

Why Randomize

Answer 13

Minimize bias
Comparable groups
Uncertainty of benefits of a therapy

Question 14

Randomization Methods

Answer 14

Fixed randomization methods

Adaptive randomization methods

Question 15

Fixed Randomization Methods

Answer 15

Simple, block, stratified

Question 16

Adaptive randomization methods

Answer 16

Allocation probabilities change as study progress in order to address imbalances that arise

Question 17

Simple Randomization

Answer 17

Participants randomly and equally assigned
Simple and easy
Potentially imbalanced

Question 18

Block Randomization

Answer 18

Avoids large imbalances from occurring throughout the study period
Can lead to selection bias
Double bind = no bias

Question 19

Stratified Randomization

Answer 19

Smaller study
Stratify pts by variables of concern at baseline ensures comparability btwn grps for that factor
A: ensures even distribution, increases power, balancing only 2-3 baseline variables
D: Large numbers

Question 20

Define Selection Bias

Answer 20

Sample not representative of the population due to pt selection methods

Question 21

Define Sample Bias

Answer 21

Sample does not reflect the spectrum of characteristics in the target population

Question 22

Define Informational bias

Answer 22

Differences in the way that information collected in study occurs

Question 23

Define study design/analysis bias

Answer 23

Unfair comparisons

Fail to take into account confounders

Question 24

Control bias

Answer 24

Randomization

Blinding

Question 25

Parallel Group Pro

Answer 25

Well controlled
Well defined population
Bias and error limited
Well known statistical tools

Question 26

Parallel Group Cons

Answer 26

Expensive
Need a large number of subjects
Volunteer bias
Loss to followup

Question 27

Crossover Design Pros

Answer 27

Difference based upon within subject comparisons (less variability and fewer subjects)
Easier
Subjects must be on meds all times
Good for short-acting drugs, stable/chronic diseases, bioequivalent studies

Question 28

Crossover Design Cons

Answer 28

Not useful for acute or permanent endpoints
Carry over effect
Exaggerated efficacy of second treatment