EPIDEMIOLOGY-BIOSTATISTICS EXAM

Midterm 2004
PRINT YOUR LEGAL NAME: __________________________________
Instructions: This exam is 20% of your course grade. The maximum number of points
for the course is 1,000; hence, this exam is worth 200 points. There are 20 questions on
this exam. Each question is worth 10 points to yield the maximum of 200 points for this
exam. For questions 1-10, record the best answer in pencil on the answer sheet provided.
For questions 11-20, write your answers in the spaces provided. Submit your exam and
your answer sheet as directed after you have completed the exam.
1. Investigators want to know if a new diet can reduce the risk of stroke in adult
women. They conduct a RCT assigning one arm to the new diet and the other
arm to their routine diet. It is known that smoking and hypertension are strong
risk factors for stroke. Table 1 appears below:
New Diet Routine Diet
(n=500) (n=500)
% Smokers 25 40
% Hypertension 10 10
% family history of stroke 3 3
Select the best answer:
a.) The lack of adult men in the trial affects its internal validity.
b.) It is possible that the crude relative risk for new diet vs. routine diet is 1.0.
c.) The authors should adjust for hypertension because it is a potential confounder in
this trial.
d.) The authors should make the two arms comparable for smoking by matching.
e.) The smoking-adjusted relative risk for new diet vs. routine diet must be closer to
1.0 than the crude relative risk.
CONTINUE TO THE NEXT PAGE
PRINT YOUR LEGAL NAME: ______________________________
2. Researchers want to study Paroxysmal Nocturnal Hemoglobinuria (PNH), a
disease with an incidence of about 1 per million per year, which manifests itself
during the early adult years. They wonder if there is an association between low
birth weight and the later development of PNH. The researchers should conduct
a: (Select the best answer.)
a.) RCT with low birth weight and normal birth weight as the two arms.
b.) Retrospective cohort study with PNH as one arm and no PNH as the other
arm.
c.) Retrospective cohort study with low birth weight as one arm and normal
birth weight as the other arm.
d.) Case control study with PNH as cases and subjects without PNH as
controls
e.) Large case series
3. Public health officials believe that they could prevent 10 cases of Hepatitis C per
5,000 interns over a 10-year period by preventing all of their needle stick injuries.
Select the best answer:
a.) This is the attributable risk of Hepatitis C due to needle stick injuries in
interns.
b.) This is the incidence of Hepatitis C in interns with needle stick injuries
divided by the incidence of Hepatitis C in interns without needle stick
injuries.
c.) This is the cumulative incidence of Hepatitis C in interns with needle stick
injuries.
d.) This is the incidence rate of Hepatitis C in interns with needle stick
injuries.
e.) This is the prevalence of Hepatitis C in interns with needle stick injuries.
CONTINUE TO THE NEXT PAGE
PRINT YOUR LEGAL NAME: ____________________________
4. A RCT of 10,000 adults was conducted to determine if a new anti-platelet
medication taken with aspirin is better than aspirin taken alone in subjects with
heart disease. The outcome of interest was MI (heart attack). Select the best
answer:
a.) It is possible that confounding by the new anti-platelet medication could
occur.
b.) Comparing one arm taking two medications to another arm taking only
one medication will likely lead to interaction by the arm taking the two
medications.
c.) After randomization, the arms will likely be comparable for unknown
potential confounders.
d.) Table 1 with the baseline characteristics of both study arms should show
both the known and unknown potential confounders.
e.) Confounding is eliminated by the choice of a RCT.
5. Which statement best describes adjustment? Select the best answer.
a.) Its a method to assure there is an equal distribution of potential
confounders in study arms.
b.) Its used when there is a concern that confounding may lead to an over
estimate or an under estimate of the appropriate measure of association.
c.) It can offset the effects of subjects not adhering to their assigned
protocols.
d.) Its a study design method to help prevent confounding.
e.) Its useful in calculating interaction.
6. Select the best statement concerning recall bias:
a.) Its a bias unique to retrospective cohort studies.
b.) It occurs in RCTs when subjects do not recall their assigned protocols.
c.) It occurs if the outcome assessment committee is not blinded such that
they can recall to which arm the subject was assigned.
d.) It occurs when cases and controls know their outcome status and
differentially recall their exposure history in a case control study.
e.) It occurs when subjects in a case control study differentially recall their
disease outcome.
CONTINUE TO THE NEXT PAGE
PRINT YOUR LEGAL NAME: __________________________
7. The Central Limit Theorem (CLT) says that if many random samples of sufficient
size are taken from a population, and the means from each sample are calculated
and plotted, that the plot will have certain characteristics. Select the best
statement.
a.) The CLT predicts that the mean of the distribution of sample means will
be the population mean.
b.) The CLT predicts that the plot of the sample means will be skewed if the
population is skewed.
c.) The standard deviation of the sampling distribution is only dependent on
the standard deviation of the population distribution.
d.) The standard deviation of the population distribution is called the standard
error of the mean.
e.) The standard deviation of the population distribution is interchangeable
with the term standard error.
8. Select the best statement:
a.) The p value can be used to help predict the probability that the null
hypothesis is false.
b.) A trials power equals the probability of a Type II error.
c.) A study result with a p value less than alpha must be statistically
significant.
d.) A study result with a p value less than alpha must be clinically significant.
e.) Performing multiple comparisons without adjusting alpha increases the
probability of a Type II error.
CONTINUE TO THE NEXT PAGE
PRINT YOUR LEGAL NAME: _________________________
9. A researcher is conducting a prospective cohort study to determine if there is an
association between high fiber diets and gallbladder cancer. She is the only
investigator and she is not blinded. There is, however, an independent outcomes
committee responsible for determining whether a subject develops gallbladder
cancer during the study. The researcher is very concerned that subjects not on a
high fiber diet might develop colon cancer, a secondary outcome of the study, so
she follows them more closely for colon cancer vs. the subjects on the high fiber
diet. Select the best statement:
a.) By following the subjects not on a high fiber diet more closely, the
researcher will eliminate the chance of non-random misclassification of
gallbladder cancer.
b.) The study is subject to reporting bias if she erroneously estimates the
relative risk of colon cancer in subjects on a high fiber diet compared to
those that are not.
c.) This study is prone to interviewer bias if the researcher misclassifies the
amount of fiber in the subjects diets.
d.) Selection bias could have occurred if the researcher only selected subjects
from a narrow geographical area.
e.) Surveillance bias could lead to non-random misclassification of colon
cancer.
10. Select the correct statement regarding societal determinants of health.
a.) USA citizens have the longest life expectancy of any nation because the
USA spends the most money per capita on health care of any nation.
b.) Discrepancies of health status between industrialized nations can partly be
explained by variations in the social cohesiveness of its citizens.
c.) Once the vast majority of a countrys population is lifted out of poverty,
all of its citizens experience roughly the same health status.
d.) Physicians spend most of their time managing the societal determinants of
their patients health.
e.) The stressor is a more powerful health determinant than the stress
response.
CONTINUE TO THE NEXT PAGE
PRINT YOUR LEGAL NAME: _______________________________
11. What is meant by an intention-to-treat analysis?
12. How does non-random misclassification bias of outcomes affect study results?
13. A physician initially wants 80% power to detect a true difference of at least 20%
between two study arms with alpha set at 0.05. After further thought, she wants
to achieve 90% power to detect a true difference of at least 20%, still keeping
alpha at 0.05. What does she need to do?
14. An author reports that the relative risk of adults having a heart attack while on
aspirin compared to placebo is 0.8. This result is statistically significant. The
author further reports that the relative risk of skiers having a heart attack while on
aspirin compared to skiers on placebo is 0.7. The relative risk of non-skiers
having a heart attack while on aspirin compared to non-skiers on placebo is 0.5.
The difference between 0.7 and 0.5 is not statistically significant. The author
reports that skiing is an effect modifier in this study. Is the author correct?
Answer yes or know and then state your reasons to support your answer.
CONTINUE TO THE NEXT PAGE
PRINT YOUR LEGAL NAME: _____________________________
15. What is the essential study design difference between a case control study and a
retrospective cohort study?
16. The test scores from an exam are normally distributed with a mean of 80%. One
standard deviation was 5%. What percent of those taking the exam scored
between 75% and 90%?
17. List the two necessary conditions for confounding to occur.
18. What does it mean when an investigator reports that her study result is statistically
significant?
19. The 95% confidence interval on a sample mean extends from 5.0 to 10.0 Is it
possible that the population mean from which the sample was drawn, the parameter,
does not lie somewhere between 5.0 and 10.0? Explain your answer.
20. List two advantages of a prospective cohort study compared to a case control
study.
THE END