Lecture 8

Question 1

biases that exist in RCTs

Answer 1

• selection bias

- measurement bias

Question 2

in a ____ study we are comparing two groups that differ in exposure but also differ in a number of other characteristics as well

Answer 2

• cohort

Question 3

difference in cohort studies versus RCTs

Answer 3

• in RCTs we impose our own intervention
• interventions are assigned randomly to participants
• much stronger case for causation

Question 4

the benefit of randomizing in RCTs

Answer 4

• confounding is basically eliminated.

- may be imbalances in factors strictly due to chance

Question 5

difference in analysis of RCTs versus cohort studies

Answer 5

• analysis of RCTs is simpler than analysis of cohort studies because confounding can basically be ignored

Question 6

problems arise when interpreting RCTs when the trial is ____

why

Answer 6

• small
• null hypothesis might not be rejected either because of low power or there being no difference in the groups
• imbalance in intervention groups on important prognostic characteristics, by chance aline

Question 7

type II error

Answer 7

• fails to detect a difference when a difference does exist

Question 8

should we wait for an RCT before using every medical intervention?

Answer 8

• no.

Question 9

once as intervention becomes part of accepted clinical practice, even in the absence of evidence, it is perceived as ____ to perform a randomized clinical trial at that point

Answer 9

• unethical

Question 10

a recent example of a difference in findings between cohort studies and RCTs

Answer 10

• effect of vitamin E on ischemic heart disease
• cohorts said it reduced risk
• RCTs said that it does not

Question 11

what is the name for someone figuring out the randomization scheme and could subvert it in any way

ways to avoid this

Answer 11

• allocation concealment
• randomization done by people not involved in enrolling subjects
• intervention assignment in a sealed envelope opened only after the patient has been enrolled in the study

Question 12

what are crossovers

Answer 12

• people assigned to one treatment but actually receive the comparison treatment

analyze these people based on treatment that they receive not on what they were assigned

Question 13

intention to treat analysis

• depends on

Answer 13

• subjects analyzed according to the group to which they were randomized rather the treatment actually received
• depends on the research question being asked

Question 14

if participants are analyzed by the treatment they actually received, rather than what they were assigned to receive, one loses the advantages of the

Answer 14

• the advantages of the randomization process are lost

Question 15

what is the preferred method of analysis when attempting to determine if a given treatment produces a desired result

Answer 15

• intention to treat

Question 16

what happens in intention to treat analyses if a large number of crossovers occur

Answer 16

• underestimation of true treatment effect

Question 17

what is clinical equipoise

Answer 17

• honest professional agreements exist about the preferred treatment of the condition in question
• neither the experimental treatment nor the standard treatment should be considered inferior or potentially harmful to the patient.
• RCT must produce results which are convincing enough to resolve the dispute over the preferred treatment

Question 18

what is the name for the extent to which the results of the student under review are applicable to a wider group of people

Answer 18

• generalizability (external validity)