Module 10

Question 1

Randomized clinical (control) trial

Answer 1

o Gold standard
o True experimental study design
o Random assignment

Question 2

Randomization

Answer 2

o Isolate effect of intervention
o Want same experimental/control groups
o Randomization avoids confounding
o Biased results if efficacy under/overestimated
o Underlying premise: can’t give subjects choice

Question 3

Comparison/Control group

Answer 3

No treatment at all
Inactive treatment
o Placebo—pharmacologically inactive substance
o Sham—bogus procedure resembles real one
Another active treatment

Question 4

Historical controls

Answer 4

Compare results of patients given new treatment with previous patients (with same diagnosis) given earlier treatment

Question 5

Crossover study (RCT)

Answer 5

Randomize treatment/control groups and switch at some point (washout period between treatments)
Limitations: treatment residual effects; can’t undo surgery

Question 6

Outcome of clinical trials

Answer 6

• Referred to as clinical end points
• May include rates of disease, death, or recovery
• Outcomes must be comparable

Question 7

The treatment effect

Answer 7

If treatment/agent/program has an effect on outcome, change in outcome is the treatment effect
o Survival
o Functional status
o Change in course of illness
o Odds ratio

Question 8

Attributes of RCTs

Answer 8

• Expensive
• Prospective
• Detect smaller differences with larger sample
• Trade-off: cost and precision

Question 9

Blinding

Answer 9

To minimize bias
o Single blinding—subjects don’t know
o Double blinding—subjects/observer don’t know
o Triple blinding—subjects/observer/reviewer don’t know

Question 10

Noncompliance (reasons)

Answer 10

1. Toxic reactions to treatment, side effects
2. Waning interest
3. Inability to meet demands of study
4. Desire to seek other therapies
5. Disease progression
6. Death

Question 11

How to improve compliance

Answer 11

1. Make treatment easy/simple to follow
2. Enroll motivated/knowledgeable subjects
3. Describe treatment demands realistically
4. Take detailed med hist; exclude risk of noncom
5. Mask subjects so don’t know getting placebo
6. Frequent contact with subjects
7. Test period before enrollment/randomization to assess
   tolerance/compliance

Question 12

Unplanned crossovers

Answer 12

Subjects switched from one group to the other

Question 13

How to deal with unplanned crossovers

Answer 13

o Drop crossovers
o Switch groups
o Apply “intention to treat” principle
o Discard study

Question 14

Intention to treat

Answer 14

o Preserves benefits of randomization
o Maintains statistical power of original study population
o Since good/poor compliers differunbiased results
o Gives information on treatment effectiveness under everyday circumstance where some won’t comply

Question 15

If exposure changes

Answer 15

o Issue of compliance
o “Dilution” or “contamination”
o Multiple risk factor intervention trial

Question 16

Phase I clinical trials

Answer 16

Small group of people (20–80) for the first time to evaluate its safety and identify side effects

Question 17

Phase II clinical trials

Answer 17

Larger group of people (100–300) to determine its effectiveness/further evaluate its safety

Question 18

Phase III clinical trials

Answer 18

Large groups of people (1,000–3,000) to:

• confirm its effectiveness
• monitor side effects
• compare it with standard/equivalent treatments
• collect information that will allow the experimental drug or treatment to be used safely

Question 19

Phase IV clinical trials

Answer 19

After a drug is approved by the FDA and made available to the public, researchers track its safety, seeking more information about a drug or treatment’s risks, benefits, and optimal use

Question 20

RCT ethical issues

Answer 20

o Patient/physician relationship: trust
o Sufficient certainty that treatment might be beneficial to give it to some
o Sufficient doubt about treatment benefits to withhold it from others
o Informed consent—make patients aware of risks/benefits *Randomization/blinding makes informed consent difficult

Question 21

Meta-analysis

Answer 21

o Increase power, resolve disagreements

o Improve estimates, pose new questions

Question 22

Community trials

Answer 22

o Larger unit of analysis
o May be randomized
o If easier/cheaper to give treatment through common mode
o Need protocol

Question 23

Community trial process

Answer 23

1. Formal recruitment
2. Baseline measurements
3. Selection of communities
4. Assignment to treatment/control o Follow over time
5. Calculate difference

Question 24

Community trial vs. RCT

Answer 24

o Depends on intervention
o Depends on population size
o Behavioral change may need community trial
o Complex interventions may need community trial

Question 25

Clinical epidemiology and decision making

Answer 25

Use of evidence, derived from observational and experimental studies of human illness or risk factors for illness, in medical decision making

Question 26

Paradigm shift in making clinical decisions

Answer 26

opinions/tradition > scientific evidence