Description
1. Complete the CDC E-learning activities “Create an Epi Curve” and “Using an Epi Curve to determine mode of spread” at https://www.cdc.gov/training/quicklearns/ 2. Create an epi-graph based on the data in exercise 1 of “Create an Epi Curve.” This can be made in MS Excel or another graphing utility. 3. Submit it with your answers to the exercises in the module (questions start on slide 8) in a MS RTF file or PDF.
Unformatted Attachment Preview
Assignment 2
This assignment should be done in SAS Studio using the dataset SASHELP.HEART. You will need to
export the results into a PDF or Word document in order to copy/screenshot your results into this file.
Please save your assignment as a PDF before submitting. Hint: You will need to use the options tab to
only provide what is asked in the question. Remember classification variables are used when you want
to divide the data into certain categories/groups.
1. Create a combined histogram and boxplot for Height. Include the measures in the 5-Number
Summary as inset statistics. Hint: Use summary statistics.
1a. Insert graph (histogram with boxplot) here:
1b. Based on the boxplot in #1, are there any outliers? If so, how many?
Answer: There are 2 outliers.
The calculated IQR is found to be 5.25, and the quartile values are found to be
Q1 62.2500000 and Q3 67.5000000. Using the IQR method, we considered an outlier
any value that falls five quartile values below the lower quartile, which is 54.375,
approximately or above the upper quartile by approximately 5 quartile values, which
comes to about 75.375. Therefore, there are 2 outliers.
1c. Is the mean height greater or less than the median?
Answer: After comparing the mean (64.8131847) and the median (64.5000000), we can
conclude that the mean is slightly more significant than the median. Hence, we can define
this distribution pattern as having a slightly right-skewed distribution, meaning potential
outliers on the right side of the distribution can pull the mean slightly higher.
2.
Create a table comparing Height for males and females. Include the following measures in the table:
Number of observations, mean, standard deviation, minimum and maximum. Include comparative
histograms and boxplots. Hint: Use summary statistics.
2a. Insert table here:
2b. Insert comparative histograms here:
2c. Insert comparative boxplots here:
3. Create a table for Age at Start and Age at Death. Include n, mean, standard deviation, and CV. Hint:
Use summary statistics.
3a. Insert table here:
3b. Interpret the CV values for these two variables in 3a.
Answer: Another way to measure relative variability is to calculate the coefficient of
variation (CV), the standard deviation divided by the mean, expressed as a percentage.
For the age at start variable, for example, the Standard deviation is about 8.55 years,
which is approximately 19.48% of the mean of 44.07 years. In other words, there is a
moderate relative variability of the ages at the start as the ages at the start are
somewhat dispersed around the mean age of 44.07 years. In the case of the age at death,
the standard deviation is about 10.58 years, which is a 14.99% proportion. Consequently,
the ages at death have a slightly smaller degree of dispersion relative to their mean age
of 70.54 years than the ages at the start.
3c. Write a statement about how the relative variability for these two variables compares.
Answer: The coefficient of variation (CV) gives insights into the relative variability of both
variables. In this dataset, ‘Age at Start’ demonstrates a higher CV of approximately
19.48% compared with ‘Age at Death,’ with a CV value of 14.99%. This implies that ‘Age
at Start’ has more relative variability, which indicates that ages at the start do deviate
more from the mean age, which is approximately equal to 44.07 years when compared
with ‘Age at Death,’ where the ages do lie closer to the mean age of 70.54 years. Hence,
both the variables show moderate relative variability; this suggests a kind of dispersion in
ages at the start of alcohol use and ages at death but to different extents.
4. Create a frequency table for Smoking_Status with counts (frequency) and relative frequencies
(percent) for each class. Include a bar chart representing the distribution of Smoking Status. (Numbers
in bar chart should match numbers in the frequency table.) Hint: Use One-Way Frequencies.
4a. Insert table here:
4b. Insert bar chart here:
CHAPTER 6:
PROBLEMS AND
LIMITS OF
EPIDEMIOLOGY
CHEVONNIA L. JONES, MPH
ASSISTANT PROFESSOR-HEALTH SERVICES PROGRAM
AUGUSTA UNIVERSITY
Intervention study problem:
PROBLEMS
WITH
STUDYING
HUMANS
• Subjects may not follow prescribed behavior throughout study
period.
Cohort study problem:
• Sometimes it is hard to isolate which of many factors are
responsible for health differences. Likely to suffer a form of
bias caused by people dropping out or being untraceable
Case-control study problems:
• Control group may not be truly comparable.
• Errors may exist in reporting or recalls.
For all studies, differential drop-outs are worrisome.
SOURCES OF ERROR
Random variation: association merely due to
chance; the way a coin will land on heads or tail
if flipped the same way.
Confounding variables: associated with the
exposure & can independently affect risk of
developing disease.
SOURCES
OF ERROR
BIAS: INFLUENCE OF IRRELEVANT FACTORS OR
ASSOCIATIONS
– SELECTION BIAS: WHEN THE CONTROL GROUP IS
INSUFFICIENTLY SIMILAR TO THE TREATMENT GROUP.
– REPORTING BIAS OR RECALL BIAS: WHEN THE CASE
GROUP & CONTROL GROUP SYSTEMATICALLY REPORT DATA
DIFFERENTLY, EVEN IF THEY HAD THE SAME EXPOSURE.
FACTORS THAT LEND VALIDITY TO RESULTS
Strong
association
Dose–response
relationship
Known
biological
explanation
Consistent
results from
several studies
High relative
risk or odds
ratio
Large study
population
HORMONE REPLACEMENT THERAPY
Conflicting results exist between two major studies.
Previous positive evidence has all come from observational studies.
Clinical trial is the gold standard.
Results of cohort study were confounded by associated factors that made women taking
HRT healthier, even without the therapy.
Became popular in the 1960s
ETHICAL ISSUES
• NAZI EXPERIMENTS ON HUMANS
• TUSKEGEE SYPHILIS STUDY
• AIDS EPIDEMIC
• BONE MARROW TREATMENT FOR ADVANCED BREAST CANCER
• NEW RULES
– INFORMED CONSENT
– INSTITUTIONAL REVIEW BOARDS
• IMPORTANCE OF CLINICAL TRIALS
• POSSIBILITY OF CONFLICT OF INTEREST WITH MEDICAL PROVIDERS WHO
STAND TO PROFIT
Drug companies are required to conduct randomized controlled
trials on a new drug before it can be approved.
Harmful side effects have frequently become obvious after drugs
were approved.
There is evidence that drug companies sometimes suppress
negative findings.
All clinical trials must now be registered in advance with a public
database.
Randomized Controlled Trails are considered the best way to test
drugs because the FDA reviews results & FDA approval has
generally meant that a drug was safe and effective.
CONFLICTS
OF
INTEREST
IN DRUG
TRIALS
DISCUSSION QUESTION 1
• WHAT ARE STRENGTHS AND WEAKNESSES OF EACH OF THE MAJOR TYPES OF
EPIDEMIOLOGIC STUDY?
– RANDOMIZED CONTROLLED TRIAL
– COHORT
– CASE-CONTROL
DISCUSSION QUESTION 2
• HAVE YOU HEARD OF THE TUSKEGEE SYPHILIS STUDY?
• WHY WAS IT UNETHICAL?
• WHAT INFLUENCE HAS IT HAD ON THE CONDUCT OF CLINICAL TRIALS?
DISCUSSION QUESTION 3
• VISIT THE WEBSITE OF THE HASTINGS CENTER, WWW.THEHASTINGSCENTER.ORG.
• WHAT ISSUES IS THE HASTINGS CENTER CONCERNED WITH THIS MONTH?
1
Statistics Question.
Student’s Name:
Course Name and Number:
Institutional Affiliation:
Instructor’s Name:
Date Due:
2
Question One
Option a.
Option b. Answer: There are 2 outliers.
The calculated IQR is found to be 5.25, and the quartile values are found to be Q1
62.2500000 and Q3 67.5000000. Using the IQR method, we considered an outlier any value that
falls five quartile values below the lower quartile, which is 54.375, approximately or above the
upper quartile by approximately 5 quartile values, which comes to about 75.375. Therefore, there
are 2 outliers.
Option c.
After comparing the mean (64.8131847) and the median (64.5000000), we can conclude
that the mean is slightly more significant than the median. Hence, we can define this distribution
3
pattern as having a slightly right-skewed distribution, meaning potential outliers on the right side
of the distribution can pull the mean slightly higher.
Question two
Option a.
4
Option b
5
Option c:
6
Question Three
Option a.
Option b.
Another way to measure relative variability is to calculate the coefficient of variation
(CV), the standard deviation divided by the mean, expressed as a percentage. For the age at start
variable, for example, the Standard deviation is about 8.55 years, which is approximately 19.48%
of the mean of 44.07 years. In other words, there is a moderate relative variability of the ages at
the start as the ages at the start are somewhat dispersed around the mean age of 44.07 years. In
the case of the age at death, the standard deviation is about 10.58 years, which is a 14.99%
proportion. Consequently, the ages at death have a slightly smaller degree of dispersion relative
to their mean age of 70.54 years than the ages at the start.
Option c.
The coefficient of variation (CV) gives insights into the relative variability of both
variables. In this dataset, ‘Age at Start’ demonstrates a higher CV of approximately 19.48%
compared with ‘Age at Death,’ with a CV value of 14.99%. This implies that ‘Age at Start’ has
more relative variability, which indicates that ages at the start do deviate more from the mean
age, which is approximately equal to 44.07 years when compared with ‘Age at Death,’ where the
ages do lie closer to the mean age of 70.54 years. Hence, both the variables show moderate
7
relative variability; this suggests a kind of dispersion in ages at the start of alcohol use and ages
at death but to different extents.
Question Four
Option a.
Option b
CHAPTER 6:
PROBLEMS AND
LIMITS OF
EPIDEMIOLOGY
CHEVONNIA L. JONES, MPH
ASSISTANT PROFESSOR-HEALTH SERVICES PROGRAM
AUGUSTA UNIVERSITY
Intervention study problem:
PROBLEMS
WITH
STUDYING
HUMANS
• Subjects may not follow prescribed behavior throughout study
period.
Cohort study problem:
• Sometimes it is hard to isolate which of many factors are
responsible for health differences. Likely to suffer a form of
bias caused by people dropping out or being untraceable
Case-control study problems:
• Control group may not be truly comparable.
• Errors may exist in reporting or recalls.
For all studies, differential drop-outs are worrisome.
SOURCES OF ERROR
Random variation: association merely due to
chance; the way a coin will land on heads or tail
if flipped the same way.
Confounding variables: associated with the
exposure & can independently affect risk of
developing disease.
SOURCES
OF ERROR
BIAS: INFLUENCE OF IRRELEVANT FACTORS OR
ASSOCIATIONS
– SELECTION BIAS: WHEN THE CONTROL GROUP IS
INSUFFICIENTLY SIMILAR TO THE TREATMENT GROUP.
– REPORTING BIAS OR RECALL BIAS: WHEN THE CASE
GROUP & CONTROL GROUP SYSTEMATICALLY REPORT DATA
DIFFERENTLY, EVEN IF THEY HAD THE SAME EXPOSURE.
FACTORS THAT LEND VALIDITY TO RESULTS
Strong
association
Dose–response
relationship
Known
biological
explanation
Consistent
results from
several studies
High relative
risk or odds
ratio
Large study
population
HORMONE REPLACEMENT THERAPY
Conflicting results exist between two major studies.
Previous positive evidence has all come from observational studies.
Clinical trial is the gold standard.
Results of cohort study were confounded by associated factors that made women taking
HRT healthier, even without the therapy.
Became popular in the 1960s
ETHICAL ISSUES
• NAZI EXPERIMENTS ON HUMANS
• TUSKEGEE SYPHILIS STUDY
• AIDS EPIDEMIC
• BONE MARROW TREATMENT FOR ADVANCED BREAST CANCER
• NEW RULES
– INFORMED CONSENT
– INSTITUTIONAL REVIEW BOARDS
• IMPORTANCE OF CLINICAL TRIALS
• POSSIBILITY OF CONFLICT OF INTEREST WITH MEDICAL PROVIDERS WHO
STAND TO PROFIT
Drug companies are required to conduct randomized controlled
trials on a new drug before it can be approved.
Harmful side effects have frequently become obvious after drugs
were approved.
There is evidence that drug companies sometimes suppress
negative findings.
All clinical trials must now be registered in advance with a public
database.
Randomized Controlled Trails are considered the best way to test
drugs because the FDA reviews results & FDA approval has
generally meant that a drug was safe and effective.
CONFLICTS
OF
INTEREST
IN DRUG
TRIALS
DISCUSSION QUESTION 1
• WHAT ARE STRENGTHS AND WEAKNESSES OF EACH OF THE MAJOR TYPES OF
EPIDEMIOLOGIC STUDY?
– RANDOMIZED CONTROLLED TRIAL
– COHORT
– CASE-CONTROL
DISCUSSION QUESTION 2
• HAVE YOU HEARD OF THE TUSKEGEE SYPHILIS STUDY?
• WHY WAS IT UNETHICAL?
• WHAT INFLUENCE HAS IT HAD ON THE CONDUCT OF CLINICAL TRIALS?
DISCUSSION QUESTION 3
• VISIT THE WEBSITE OF THE HASTINGS CENTER, WWW.THEHASTINGSCENTER.ORG.
• WHAT ISSUES IS THE HASTINGS CENTER CONCERNED WITH THIS MONTH?
Purchase answer to see full
attachment
