Statistics for Clinicians Quiz Flashcards
Consider this outcome:
exposed: 23
control:15
risk: 1.5
p value: 0.04
What does a p value < 0.05 mean?
A. False positive and false negative conclusions are ruled out
B. This was a randomized study
C. There was a good match between treatment and reference groups
D. This study has a high quality design without important limitations
E. It is unlikely (< a 5% chance) this finding is due to random variation
F. The finding is clinically significant
E
There were 60 patients enrolled in this study. The outcome M vs F reports 16 males, and 12 females. the p value is 0.31.
A. Reflects a reasonable balance for gender
B. This is an example of level 1 evidence
C. Patient randomization is inadequate; patient allocation is inadequate and intent to treat is inadequate
D. There is no difference between groups for this outcome.
C
The authors concluded there was no difference in the primary and secondary harm outcome.
Primary harm:
exposure 7
control 0
risk ratio 15
p value 0.06
Secondary harm:
exposure 3
control 0
risk ratio 7
p value 0.19
A. The difference is not statistically significant. However: Limited power prevents making a conclusion about the risk of harm.
B. Since only 7 out of 60 patients reported harm we can conclude that harm is not
C. The authors are correct because the p value is > than 0.05
A
Power is often used to estimate sample size. Sample Size is proportional to:
1) The Alpha (0.05) and Beta (0.8 or 0.9) error (Acceptable type I and Type II error). Alpha power protects from accepting an illusion as reality. Beta Power protects us from missing reality.
2) Sample size is further estimated by the Population or Underlying Rate of an outcome, and
3) also by the size of Delta or difference in rate expected for the experimental group.
All of this does make intuitive sense. If you are looking for rare events you need a large sample size to have a decent chance of finding those events. Or when looking for a small change between groups again you need a large sample size to be able to recognize those changes.
What is the correct statement about delta for harms in this study?
Primary harm:
exposure 7
control 0
risk ratio 15
p value 0.06
Secondary harm:
exposure 3
control 0
risk ratio 7
p value 0.19
A. Delta for statistical significance is a risk greater than 15 fold.
B. Cannot be determined
C. Delta is 1.5
D. Since the p value is > 0.05 a 15 fold increase in harm is not a meaningful finding.
E. The p value is 0.06
A
True or False: A placebo group is an adequate comparison to use in a study of treatment of type 2 DM
False – if there is a proven effective treatment against the condition you are researching, a placebo group is inadequate as the new treatment needs to be compared to a group receiving current gold standard treatment
Drug X is a novel compound designed to treat MDD, studied in a phase II clinical trial. Which response provides a clearer meaning for the term novel compound?
A. Drug X is an important breakthrough for the treatment of MDD
B. Drug X is an experimental compound designed to treat MDD. The efficacy and safety of drug X are not yet established.
C. Drug X is a compound from the future superior to all the ordinary and overused medications of the past.
B
This question is about 95% confidence intervals (95% CI)
Statistically significant:
A. The 95% CI of the test medication are entirely outside of the 95% CI of the comparison medication
B. The 95% CI of the test medication are entirely within the 95% CI of the comparison medication
C. The 95% CI overlap between the test, and comparison medication
A
This question is about 95% confidence intervals (95% CI)
The difference is not statistically significant:
A. The 95% CI of the test medication are entirely outside of the 95% CI of the comparison medication
B. The 95% CI of the test medication are entirely within the 95% CI of the comparison medication
C. The 95% CI overlap between the test, and comparison medication
C
This question is about 95% confidence intervals (95% CI)
Equivalent:
A. The 95% CI of the test medication are entirely outside of the 95% CI of the comparison medication
B. The 95% CI of the test medication are entirely within the 95% CI of the comparison medication
C. The 95% CI overlap between the test, and comparison medication
B
