Sunteți pe pagina 1din 12

Biomedical Literature Evaluation & Statistics

Exam Study Guide


Michael A. Mancano, Pharm.D.

1. In a randomized controlled clinical trial of a new antihypertensive agent, patients with diastolic blood
pressure readings greater than 90 mmHg on their most recent clinic visit are selected as subjects. A patient
assigned to the experimental treatment group is normotensive at his first clinic visit after initiating drug
therapy. All of the following factors represent plausible explanations for this observation EXCEPT:

a) Simpson’s Paradox
b) Random variation in blood pressure measurements.
c) Regression to the mean.
d) Efficacy of the experimental drug.
e) Measurement error.

2. A study is carried out to compare the effect of two low fat diets on vascular plaque accumulation. The
investigator conducting the study obtained a list of individuals who had been continuously enrolled
between 1985 and 1990 in one of two commercial weight loss programs based on the two diets.
Participants in both diet programs were invited to participate in the study. The protocol described above is
an example of which of the following study designs?

a) Randomized, double-blind
b) Randomized block design
c) Meta-analysis
d) Double-blind, crossover study.
e) Observational study: historical cohort study.

3. A group of patients with coronary heart disease volunteered to participate in a study of the relationship
between diet and vascular plaque accumulation. The subjects were randomized to one of three diets: very
low fat (< 10% total calories from fat), low fat (< 30% of the total calories from fat), or a reduced calorie
diet with no restrictions on fat intake. At the end of a five-year follow-up period, vascular plaque
accumulation in the three diet groups was compared to determine which diet achieved the greatest
reduction in plaque accumulation. The protocol described above is an example of which of the following
study designs?

a) Randomized clinical trial.


b) Case report
c) Meta-analysis
d) Double-blind, crossover study.
e) Observational study: historical cohort study.

4. A pharmaceutical company conducted a double-blind randomized controlled clinical trial of a new


antipruritic drug. One group of subjects received a lotion containing the experimental drug, a second group
received an identical lotion lacking the active ingredient, and a third group received no treatment at all.
Two weeks later, the “itching intensity” scores (the lower the score, the less severe the itching) for the three
groups were 12 (experimental group), 15 (placebo group) and 65 (no treatment group). Which of the
following factors BEST explains the result obtained for the subjects who received the lotion lacking the
active ingredient?

a) The Hawthorn effect


b) Regression to the mean.
c) Selection bias.
d) The placebo effect.
e) Failure to adequately blind subjects and investigators.
Please refer the following scenario to answer questions 5 and 6.

In a clinical trial of a novel antihypertensive agent, 60 patients were randomly assigned to receive the new
drug for 14 days and 60 patients were randomly assigned to receive a placebo for 14 days. Changes in
blood pressure from baseline during the treatment period were recorded for all patients. The primary
research objective was to determine if the mean change from baseline in diastolic blood pressure was
greater for the new drug than for the placebo. The result was that patients who received the new
antihypertensive agent had a reduction in blood pressure of 5 % and the patients who received the placebo
had a reduction in blood pressure of 3%, with a p-value for the comparison between the groups of p= 0.8.
Upon completion of the study the investigator decided to look at the 10 women who were randomly
assigned to each of the two groups and the results were shocking. The 10 women who received the new
antihypertensive agent experienced a reduction in blood pressure of 25% and the 10 women who received
placebo had a reduction in blood pressure of 8%, with a p-value for the comparison between the two groups
of p=0.06.

5. Which of the following statements are TRUE with regard to the study results?

a) The new antihypertensive agent is very effective in women.


b) Further study is warranted in a study designed to look specifically at the female population.
c) Initial patient selection bias probably accounts for the observed results.
d) The decision to analyze the female data is an example of an “a priori” hypothesis.
e) None of the above.

6. The results observed in the female sub-group in the above described clinical trial is an example of:

a) The Hawthorne effect.


b) Bonferroni error.
c) Simpson’s paradox
d) The Greenhouse effect.
e) Regression to the mean.

7. A pediatrician wishes to evaluate the efficacy of an experimental formula in infants who fall below the
tenth percentile in weight for their age group. Fifty such infants will be randomly assigned to receive either
the experimental formula or a standard commercially available formula for three months. Weight gains for
both groups will be recorded at the end of this time period in an attempt to answer the research question, “Is
the average weight gain higher among infants receiving the experimental formula than among those
receiving the standard formula?” After conducting the study, the pediatrician reports at a research
conference that the results are “not statistically significant (p>0.05).” All of the following statements
about this conclusion are true except:

a) The average weight gain in the infants fed the experimental formula is EXACTLY THE SAME as the
average weight gain among infants fed the standard formula.
b) A clinically important difference in average weight gain between the two groups of lower-weight
infants may exist, but the power of the test may have been too low to detect it.
c) The chance of type II error may be high, i.e. low power.
d) The observed difference in average weight gain between the experimental and control groups is likely
to be due to random chance.
8. To compare the efficacy of two drugs in lowering serum cholesterol levels, a physician measures serum
cholesterol levels in a sample of men receiving drug A and in a sample of men receiving drug B and learns
that the average serum cholesterol is lower in those receiving drug B. Plausible explanations for the
observed difference in the response to the two drugs include all of the following EXCEPT:

a) Drug B is superior to drug A in lowering cholesterol levels among the population of potential
recipients.
b) The observed difference occurred by random chance.
c) Extraneous, uncontrolled differences (other than drug therapy) between the sample of the men
receiving drug A and the sample of men receiving drug B account for the observed difference in
efficacy.
d) The observed difference may be attributed to a small p-value for the statistical test.

Please refer the following scenario to answer questions 9 and 10.

A pediatrician wished to determine the relationship between chronic otitis media in young children and
parental history of such infections. From the records of a large pediatric practice, he identified 50 children
between one and three years of age who had experienced at least three middle ear infections during the
preceding year. Fifty different children in the same age group, treated by the same practice for other
illnesses, were also identified as control subjects. The pediatrician interviewed the parents of the subjects
in both groups to determine their history of chronic otitis media as young children. Of the children with
recurrent ear infections, 30 had a family history of chronic otitis media, compared to 20 of the children
treated for other illnesses.

9. This study is an example of a:

a) Meta-analysis
b) Prospective cohort study
c) Case-control study
d) Randomized controlled clinical trial

10. The pediatrician conducting this study reports the existence of a statistically significant association
between chronic otitis media in children between the ages of one and three and a parental history of such
infections (p<0.05). Which of the following represents the MOST appropriate interpretation of this
finding?

a) A strong clinically significant association exists between the study variables in the populations from
which the samples were drawn.
b) An completely unbiased comparison was made between study groups
c) There is a cause and effect relationship between chronic otitis media in young children and a family
history of this disorder.
d) There is a less than 5% chance that the observed association occurred by random chance
Please refer the following scenario to answer questions 11 and 12.

A pharmaceutical company conducts a randomized controlled clinical trial to determine the efficacy of a
newly developed topical solution for inducing hair growth in men with male pattern baldness. The
researchers decide the average hair count must increase by 20% in the drug treated group, compared to the
placebo group, to be considered a clinically important difference. They select a sample size that will ensure
80% power for the statistical test, and carry out this test at the 5% level of significance (p = 0.05).

11. The probability that this statistical test WILL FAIL TO DETECT a clinically important difference in
average hair growth between the comparison groups, given that such a difference actually exists between
the populations from which the study samples were drawn.

a) Increases as the specified effect size decreases.


b) Increases as the sample size increases
c) Increases as the power of the test increases.
d) None of the above

12. The probability that this statistical test will detect the specified clinically important difference in
average hair growth, given that such a difference does indeed exist between the populations represented by
the study samples, is:

a) 95%
b) 80%
c) 20%
d) 5%

13. In the PREVENT trial, intravenous immune globulin for RSV was administered to premature infants
with BPD (Bronchopulmonary Dysplasia). It was a randomized, double-blind, placebo controlled trial. In
the study hospitalization was decreased as evidenced by the following data: 13.5% of patients receiving
placebo were hospitalized versus 8% of patients in the intravenous RSV immune globulin group. This
represented an absolute reduction of 5.5% and a relative reduction of 41%. What is the NNT with
intravenous RSV immune globulin to prevent one hospitalization?

a) 2.4
b) 18.2
c) 7.4
d) 12.5
e) Cannot be calculated from the data provided.
Please refer the following scenario to answer questions 14 and 15.

The Atenolol Silent Ischemia Study investigated the use of atenolol in mildly symptomatic patients with
coronary artery disease. This randomized, double-blind, placebo controlled multicenter study assessed
whether treatment with atenolol decreased adverse outcomes after 1 year. Adverse outcomes were defined
as death, ventricular tachycardia/ventricular fibrillation, aggravation of angina or revascularization.

14. Death, as an outcome, is an example of which type of data?

a) Interval Scale
b) Nominal Scale
c) Ordinal Scale
d) Ratio Scale

15. Which of the following would be the appropriate statistic to test the null hypothesis in question 37?

a) Student-t test
b) Chi squared test
c) ANOVA
d) Wilcoxon signed-rank

16. A trial of a new lipid-lowering agent, newstatin, (n=60) compared its ability to decrease LDL levels to
that of lovastatin. At the beginning of the trial, the mean baseline LDL level was 192 mg/dL in the
lovastatin group and 193 mg/dL in the newstatin group. At the end of the 4-month trial, mean LDL levels
for the lovastatin group were 152 mg/dL and for the newstatin group they were 165 mg/dL. The p value
was reported as 0.09. Which of the following is true about the interpretation of these results?

a) A further discussion of power would be necessary to draw meaningful conclusions.


b) Lovastatin and newstatin appear to be equally efficacious in lowering LDL levels.
c) Lovastatin and newstatin could be used interchangeably in patients.
d) A priori assignment of p<0.1 would have made the study more clinically useful.

17. A study compared the mean decrease in visual analog pain scales for two analgesics was reported.
Results showed that one agent (Drug X) decreased pain scale scores an average of 2 points (95% CI –5
points to –1.5 points), while the other agent (Drug Y) decreased scores an average of 4 points (95% CI –6
to +1). Which of the following statements is true?

a) Both drugs are clinically equivalent in providing analgesia.


b) Drug X would be preferred clinically
c) Drug Y would be preferred clinically.
d) Neither drug appears clinically effective.
Use the following information to complete questions 18, 19 & 20

A study was performed that evaluated the difference in platelet count and evidence of bleeding after
patients were treated with unfractionated heparin, low molecular weight heparin or warfarin. Three
hundred patients were enrolled and randomly assigned to one of the three different groups.

18. The variable documenting the evidence of bleeding (yes/no) is considered what type of data?

a) Ratio
b) Ordinal
c) Nominal
d) Interval
e) Definitive

19. What is the best statistical test to evaluate the difference in evidence of bleeding between the three
groups?

a) Student – t
b) Chi square
c) Wilcoxon signed rank
d) Paired Student – t
e) ANOVA

20. What is the best statistical test to evaluate the difference between the mean platelet counts of the three
groups?

a) Student – t
b) Chi square
c) Wilcoxon signed rank
d) Paired Student – t
e) ANOVA

Use the following information to complete questions 21 & 22

A researcher is evaluating sixty patients using a cross-over design to determine whether propranolol or
hydrochlorothiazide is more effective in managing isolated systolic hypertension.

21. Blood pressure measurements are classified as what type of statistical data?

a) Ratio
b) Ordinal
c) Nominal
d) Interval
e) Descriptive
22. Choose the appropriate statistical test to analyze whether there is a difference in mean blood pressure
when using propranolol and hydrochlorothiazide.

a) Student – t
b) Chi square
c) Wilcoxon signed rank
d) Paired Student – t
e) ANOVA

Please refer to the following scenario to answer questions 23 and 24.

A double-blind clinical study was conducted to compare the efficacy of ibuprofen 400 mg, codeine 60 mg
and a placebo in providing pain relief after minor dental surgery. Fifteen participants were randomly
assigned in groups of five to their treatment groups. Two hours after receiving the drug or placebo, patients
were asked to rate their degree of pain relief on a scale from 0 to 100, and the average score for each
treatment was calculated. The results were NOT STATISTICALLY SIGNIFICANT at the 5% level of
significance.

23. The chance of error associated with this conclusion is:

a) The type II error rate (), which is 95%


b) The type II error rate (), which is 5%
c) The Type II error rate () which is unknown
d) None of the above

24. Which of the following statements BEST defines the power of the statistical test?

a) Power is the probability that the test will detect a difference between the three drugs when, in fact, no
such difference exists.
b) Power is the probability that the test will fail to detect a difference in the level of pain relief provided
by the three treatments when they actually do differ.
c) Power is the probability that there is no difference in the level of pain relief provided by the three
treatments.
d) Power is the probability that the test will correctly detect a difference among the three treatments
which does, in fact, exist.

25. Which of the following statistical tests would be utilized for the analysis of continuous data from a four
sample, cross-over study that does NOT fit parametric assumptions?

a) Mann Whitney U test


b) Paired t test
c) Friedman’s ANOVA test
d) Cochran’s Q test
e) McNemar’s test
26. Which of the following are true concerning non-parametric statistical tests?

a) Non-parametric tests are more powerful then parametric tests


b) Many non-parametric tests deal with ranks, therefore you lose the individuality of the data.
c) A normal distribution is required prior to testing.
d) All of the above are true
e) None of the above are true

27. Which of the following is a non-parametric statistical test?

a) T-test
b) Chi-square
c) ANOVA
d) Paired-t test
e) ANCOVA

28. A nonparametric alternative to the paired-t test which can be utilized for continuous data that does not
fit the parametric assumptions, is:

a) Mann Whitney U test


b) Wilcoxon signed rank test
c) Kruskal-Wallis ANOVA
d) Cochran’s Q test
e) Friedman ANOVA

29. The NNT (Number Needed to Treat) is calculated as the inverse of the:

a) Relative Risk
b) Absolute Relative Risk
c) Relative Risk Reduction
d) Number Needed to Harm

30. As the sample size is increased in a clinical trial then the power is:

a) Increased
b) Decreased

31. In the PREVENT trial, intravenous immune globulin for RSV was administered to premature infants
with BPD (Bronchopulmonary Dysplasia). It was a randomized, double-blind, placebo controlled trial. In
the study hospitalization was decreased as evidenced by the following data: 28.7% of patients receiving
placebo were hospitalized versus 11% of patients in the intravenous RSV immune globulin group. This
represented an absolute reduction of 17.7% and a relative reduction of 68%. What is the NNT with
intravenous RSV immune globulin to prevent one hospitalization?

a) 3.5
b) 9.1
c) 5.6
d) 1.5
e) Cannot be calculated from the data provided.
32. Choose the appropriate statistical test for analyzing the data from the following clinical trial:

Thirty men between the ages of 30 and 55 participated in an experiment to study the effectiveness of a
particular diet and exercise program in reducing serum cholesterol levels. Baseline cholesterol levels were
determined for all study participants prior to initiating the program and again in three months. The research
question is “Are average serum cholesterol levels lower after the three-month program of diet and
exercise?”

a) Chi square test


b) t test
c) Paired t Test
d) Mann Whitney U test
e) One-way ANOVA

33. A clinical researcher conducts a study to compare the effect of a new histamine receptor antagonist and
placebo on peptic ulcer healing times. She reports that the difference in healing times between the two
treatments is statistically significant, p < 0.05. The correct interpretation of the level of uncertainty
associated with this conclusion is?

a) The probability of obtaining a difference greater then or equal to that observed, given that the two
treatments are actually similarly efficacious, is greater then 0.05.
b) The probability of accepting the null hypothesis, given that there is no difference between the two
treatments, is 0.05.
c) The probability that the healing time observed after treatment with the new drug truly does differ from
the healing time observed after placebo is 0.95.
d) The probability that the statistical test will detect a difference in healing times when, in fact, no such
difference exists, is less than or equal to 0.05.
e) The probability that a particular peptic ulcer patient will benefit from receiving the new drug is 0.95.

Please refer to the following scenario to answer questions 34, 35 and 36.

In a report summarizing the results of a study comparing serum sodium levels in normotensive patients and
newly diagnosed hypertensive patients prior to dietary sodium restriction, a team of clinical researchers
states that the difference in mean serum sodium levels was not statistically significant at the 5% level of
significance (p>0.05). In addition, the researchers report that, for the sample sizes employed in this study,
the power of the chosen statistical test to detect a difference in mean serum sodium levels of at least 5
mEq/L between the two patient populations is 20%.

34. Assume that serum sodium levels in the two populations do, in fact, differ by at least 5 mEq/L. The
probability that the statistical test will fail to detect this difference could then be described by all of the
following statements EXCEPT:

a) This probability is equal to 0.8 or 80%.


b) This probability increases as the level of significance is decreased.
c) This probability represents the type II error rate.
d) This probability is smaller than the probability that the test will fail to detect a difference in serum
sodium levels of at least 10 mEq/L.
35. All of the following statements about the power of the chosen statistical test are true EXCEPT:

a) The reported value for power suggests that the research team should not conclude that “no difference
exists between the mean serum sodium levels of the two populations.”
b) The power of the test will decrease if the probability of a type II error increases.
c) This value represents a 20% chance that the statistical test will detect a difference in serum sodium
levels between the two populations of at least 5 mEq/L, given that such a difference actually exists.
d) This value represents the probability of detecting a difference of at least 5 mEq/L in mean serum
sodium levels when the null hypothesis is true.
e) Given the reported value of power, a statistically significant difference in the mean serum sodium
levels of the two patient populations may have been observed if the investigators had used a larger
sample size.

36. If the investigators repeat the study with a larger sample size, they would expect which of the
following changes to occur?

a) The probability that they will detect a difference in mean serum sodium levels between normotensive
and hypertensive patients, given that no such difference actually exists, increases.
b) The probability that they will fail to detect a difference in mean serum sodium levels between the two
populations, given that such a difference does exist, decreases.
c) The power of the statistical test will decrease.
d) The level of significance, alpha, will increase.
e) The probability that the statistical test will have a false negative outcome will increase.

37. Which of the following is a result of carrying out a given statistical test of a hypothesis at a level of
significance alpha error = 0.01?

a) The probability of accepting the null hypothesis when it is false is 0.01.


b) The null hypothesis is more likely to be rejected that if the investigator had chosen a level of
significance of alpha = 0.05.
c) The probability of observing a statistically significant result is 0.01.
d) The probability of rejecting the null hypothesis when it is true is 0.01.
e) The null hypothesis is true at least 1% of the time.

38. A statistical test fails to reject the hypothesis that the correlation between blood pressure and serum
cholesterol level is equal to zero at the 5% level of significance. What is the correct interpretation of this
result?

a) The correlation between blood pressure and serum cholesterol levels in the population from which the
sample was drawn is equal to zero.
b) The correlation between blood pressure and serum cholesterol level is statistically significant (p <
0.05).
c) The results are not statistically significant; that is, there is not enough evidence to substantiate the
claim of an association between blood pressure and serum cholesterol.
d) Higher blood pressure values tend to be associated with higher serum cholesterol levels.
e) The observed correlation is unlikely to have occurred as a result of random chance.
39. A survey reports that 65 of 100 patients with multiple sclerosis have cats as household pets. The
conclusion that exposure to cats causes multiple sclerosis is

a) Valid for the study population.


b) Incorrect, because of the failure to carry out an appropriate statistical analysis of the data.
c) Incorrect, because the comparison is not based on exposure rates.
d) Incorrect, because the sample size is too small to permit generalization to the entire population of
multiple sclerosis patients.
e) Incorrect, because of the lack of an appropriate control or comparison group.

Please refer to the following scenario to answer questions 40, 41, 42 and 43.

An experiment was conducted to compare the effect of four drugs on serum cholesterol levels in patients
with type III familial hypercholesterolemia. One hundred and twenty patients were randomly assigned to
receive drug A, drug B, drug C or drug D. Serum cholesterol levels were recorded after six months of drug
therapy.

40. What statistical test should be utilized to analyze the data from this study?

a) Wilcoxon signed rank test


b) 1-way ANOVA
c) Kruskal-Wallis ANOVA
d) Chi square test
e) Repeated measures ANOVA

41. The investigators report that the p values for their comparison of the study groups are all greater than
0.1 (p >0.1). All of the following statements about this conclusion are correct EXCEPT

a) It is likely that any observed differences in average serum cholesterol levels in response to the four
drug treatments occurred by random chance.
b) Average serum cholesterol levels are exactly the same among the four treatment groups.
c) The probability of a type II error may be high.
d) A clinically important difference in average serum cholesterol levels after the four drug treatments may
exist, but the power of the statistical test may have been to low to detect it.

42. The use of multiple t-tests (with a 5% level of significance for each test), rather than an analysis of
variance, to analyze the data obtained in this study would accomplish which of the following?

a) A decrease in the chance of type I error.


b) A decrease in the likelihood of erroneously declaring a difference in means statistically significant.
c) An increase in the overall type I error rate for the entire set of comparisons.
d) A decrease in the probability of a false positive result.

43. Which of the following statements about the Bonferroni multiple comparison procedure are true?

a) It is used to determine which of a set of means are different while simultaneously controlling for the
multiple comparison problem (i.e. increased alpha error).
b) It is used to control the overall alpha error rate per experiment for the entire set of comparisons.
c) It represents a modification of the t test.
d) All of the above are true.
Use the following information to complete questions 44, 45, 46 & 47

A study was conducted to assess the prevention of deep vein thrombosis in patients undergoing knee
replacement surgery. All patients did not have a history of deep vein thrombosis. A total of 8,152 patients
were randomized to receive either Enoxaparin 30 mg subcutaneously every 12 hours (n = 4,096) or heparin
5000 units subcutaneously every 12 hours (n = 4,056). All patients were followed for a total of 2 months
after surgery. The primary endpoint of the study was the occurrence of a deep vein thrombosis during the
two month postoperative period. The number of patients with the primary endpoint are as follows:

Outcome
Yes No
Enoxaparin 614 3482
Heparin 1622 2434

44. The calculated relative risk (RR) for this study is?

a) 0.09
b) 0.16
c) 0.37
d) 0.84
e) Cannot be calculated from the data provided.

45. The calculated absolute relative risk (ARR) for this study is?

a) 10%
b) 25%
c) 35%
d) 58%
e) Cannot be calculated from the data provided

46. The calculated relative risk reduction (RRR) for this study is?

a) 0.18
b) 0.27
c) 0.48
d) 0.63
e) Cannot be calculated from the data provided

47. The calculated number needed to treat (NNT) for this study is?

a) 1.7
b) 2.8
c) 4
d) 10
e) Cannot be calculated from the data provided.

S-ar putea să vă placă și