Modules 12 and 13: Randomized Control Trials Part 1 and 2 Flashcards
Identify 4 sampling techniques.
- Simple random sampling
- Systematic sampling
- Stratified sampling
- Convenience sampling
Identifying a collection of urban or rural hospitals, then selecting from each of these, is an example of _________ sampling.
stratified
Identify 3 randomization techniques.
- Simple randomization
- Stratified randomization
- Block randomization
What are strengths of the RCT?
- Safety and efficacy
- Strong evidence for casual relationship
- High degree of internal validity
- Inclusion/exclusion criteria
Why do RCTs have a high degree of internal validity?
Randomization and manipulation of intervention
What are weaknesses of the RCT?
- Conducted in artificial environments
- Few potential casual factors to explore
- Does not reveal mechanism of action of therapy - drug works, but how?
RCTs being conducted in artificial environments is a weakness because ….
- They may not be generalizable
2. Lower external validity
How is sample size influenced by alpha?
As alpha DECREASES:
N needs to increase,
How is sample size influenced by effect size?
As effect size DECREASES, N needs to increase.
As effect size INCREASES, N needs to decrease.
How is sample size influenced by variability (standard deviation)?
Larger variability requires a higher N.
Smaller variability requires only a smaller N (in comparison).
How is sample size influenced by the homogeneity of the population?
N decreases with the homogeneity of the population.
10 deaths (of 100) occur in the old therapy. 8 deaths (of 100) occur in the new therapy.
Calculate relative risk reduction.
RRR = (old-new)/old RRR = 0.20 or 20%
10 deaths (of 100) occur in the old therapy. 8 deaths (of 100) occur in the new therapy.
Calculate absolute risk reduction. How is this reported?
ARR = old-new ARR = 0.10-0.08 = 0.02
2 lives are saved out of 10 deaths associated with the new drug. OR - 2 lives are saved due to the drug.
How is the RRR misleading compared to the ARR?
20% is great marketing! But only 2 are really saved.
10 deaths (of 100) occur in the old therapy. 8 deaths (of 100) occur in the new therapy.
Calculate “number needed to treat”.
NNT = 1/ARR NNT = 50
50 patients would have to receive the new drug for 1 to benefit.
Concluding that there is a difference between groups when on does not in fact exist is a type ____ error.
Type I error
Concluding that there is no difference between groups when one does in fact exist is a type _____ error.
Type II error
Which is NOT a benefit of blinding?
A. Reduces assessment bias especially in subjective outcomes
B. Reduces possibility of co-interventions especially in those randomized to placebo
C. Promotes the placebo effect
D. Improves adherence
C
Which is NOT true in regard to composite outcomes?
A. One outcome that is not the most important can make up the majority of the events within a composite outcome.
B. They can summarize multiple clinically important outcomes in one.
C. A statistically significant difference cannot be found with composite outcomes when no statistically significant difference can be found for any of the outcomes.
D. Because of the increased number of overall patients with the outcome, its easier to detect a difference.
C
Which is NOT a quality of an ideal outcome?
A. Short latency from intervention to outcome
B. Subjective
C. Easy to measure
D. Patient oriented
B
Which change below would NOT increase the required sample size for a trial?
A. Increased variability in the population
B. Increased desired alpha
C. Increased desired power
D. All of the above would increase the required sample size
B
