Chat with us, powered by LiveChat Measure of Association for Syphilis Infection Case Study Paper | All Paper
+1(978)310-4246 credencewriters@gmail.com
  

Analytic epidemiology is all about looking at risk factors for
specific diseases. We have measures such as relative risk, odds ratio,
population attributable risk etc. Select a disease that interests you
and evaluate 3 to 4 primary peer-reviewed articles or meta-analyses on
the disease. Report back on what the measures of association are for
this disease. Do articles report different values for the same exposure?
Why might this be?
The post must be between 400-500 words.This week we are exploring study designs. The textbook/ attached powerpoints has a list of
limitations of various study designs. Your goal for this week is to look at the article links listed below or another relevant topic from the site. It must be a primary
peer-reviewed article, no meta-analysis or review articles are
acceptable. In the discussion section of the article, there should be a
discussion of the limitations of this article. How do these limitations
compare to the limitations presented in the textbook? In a post that is at least 350 words, please summarize the article,
the limitations, and reflect on where the authors may have missed some
limitations based on the textbook readings. The initial post must be between 350 – 500 wordshttps://www.ncbi.nlm.nih.gov/pmc/articles/PMC46041…https://www.ncbi.nlm.nih.gov/pmc/articles/PMC56373…
lecture_slides_ch5_gordis.ppt.pptx

lecture_slides_ch8_gordis.pptx

lecture_slides_ch10_gordis.ppt.pptx

Unformatted Attachment Preview

ASSESSING THE VALIDITY
AND RELIABILITY OF
DIAGNOSTIC AND
SCREENING TESTS
Chapter 5
Screenings





The use of quick and simple testing procedures to
identify and separate persons who are apparently
well, but who may be at risk of a disease, from
those who probably do not have the disease
Given to large groups
Have cutoff points
Examples: vision tests, blood pressure, pap smears,
blood tests, chest x-rays
Used to refer for diagnosis
Diagnosis


In addition to using the
results of tests, involves the
evaluation of sings and
symptoms and may involve
subjective judgment.
Applied on a one-on-one
basis
Validity



The ability of a test to distinguish
between who has a disease and
who does not
Sensitivity = the ability of the test
to identify correctly those who
have the disease
Specificity = the ability of the test
to identify correctly those who do
not have the disease
Tests with Dichotomous Results




True positive = person who has the disease and
tests positive
False positive = person who does not have the
disease yet tests positive
False negative = person who has the disease yet
tests negative
True negative = person who does not have the
disease and tests negative
Sensitivity

True Positives divided by sum of true positives and
false negatives
Sensitivity =
TP
________ x 100
TP + FN
The proportion of those who have the disease that are
correctly identified as positive
Specificity
True negatives divided by the sum of true negatives
and false positives
Specificity =
TN
________ x 100
TN + FP
The proportion of non-diseased people who are
correctly identified as negative by the test

Percent False Negatives


Proportion of people with the disease who were not
detected by the test
% False Negative =
FN
___________ x 100
FN + TP
Percent False Positives


Proportion of people without the disease who were
incorrectly labeled by the test as diseased
% False Positives =
FP
_________ x 100
FP + TN
Tests of Continuous Variables


Cutoff levels above which a
test in considered positive
Examples: blood glucose
level, blood pressure
Use of Multiple Tests



Sequential (two-stage) Testing
Start with least invasive, least uncomfortable, least
expensive test
Net sensitivity and net specificity, using both tests in
sequence
Predictive Value of a Test


What proportion of the people who have the
disease will be correctly identified?
Predictive Value =
TP
__________ x 100
TP + FP
Relationship between Predictive Value
and Disease Prevalence



For a disease with low prevalence, positive
predictive value of a test is low
For a disease with higher prevalence, the positive
predictive value of the same test increases
Screening is most effective and productive if
directed at a high-risk population
Relationship between Positive
Predictive Value and Specificity

For a disease with low prevalence Positive
predictive Value is increased more by increasing
specificity than sensitivity
Reliability

Can the results obtained be replicated if the test is
repeated?
 Intrasubject
Variation – human characteristics vary over
time
 Intraobserver Variation – variation between two or
more readings by the same observer
 Interobserver Variation – variation between observers
1





Which of the following is a difference between
screening and diagnostic testing?
A. Screening is performed in a clinical setting and
diagnostic testing is done in the community
B. Screening is done on a one-on-one basis, while
diagnostic testing is done in groups
C. Screening is done on a seemingly well population
D. Screening is done to determine a treatment plan
2





The ability to distinguish between who has a
disease and who does not is called . . .
A. Validity
B. Reliability
C. Sensitivity
D. Specificity
3





Someone who has a disease and tests positive for
the disease is a . . .
A. True Negative
B. False Negative
C. False Positive
D. True Positive
4





Someone who tests negative for the disease yet
really does have the disease is a . . .
A. True Negative
B. False Negative
C. False Positive
D. True Positive
5





Which of the following is NOT a negative
consequence of a false positive test?
A.
B.
C.
D.
A burden on the health care system
People are incorrectly ‘labeled’ as positive
People may delay treatment
Anxiety and worry
6





When applying a two-stage screening, which test is
performed first?
A. The least expensive
B. The most invasive
C. The most uncomfortable
D. The longest
7





The sensitivity of a screening test is calculated . . .
A.
B.
C.
D.
TP/ TP + FN x 100
FN/ FN + TP x 100
TN/ TN + FP x 100
FP/ FP + TN x 100
8





The specificity of a screening test is calculated . . .
A.
B.
C.
D.
TP/ TP + FN x 100
FN/ FN + TP x 100
TN/ TN + FP x 100
FP/ FP + TN x 100
9





The percent false negatives of a screening test is
calculated . . .
A.
B.
C.
D.
TP/ TP + FN x 100
FN/ FN + TP x 100
TN/ TN + FP x 100
FP/ FP + TN x 100
10





The percent false positives of a screening test is
calculated . . .
A.
B.
C.
D.
TP/ TP + FN x 100
FN/ FN + TP x 100
TN/ TN + FP x 100
FP/ FP + TN x 100
11





What happens to the positive predictive value of a
screening test when it is applied to a population
with greater prevalence of the disease ?
A. it increases
B. it decreases
C. it remains the same
D. it is undefined
12





Which effects a greater increase in the positive
predictive value of a screening test ?
A. Sensitivity
B. Specificity
C. Reliability
D. Validity
13





If two scientists are determining the results of a test,
the proportion of the occurrence when they
disagree is called . . .
A. intrasubject variation
B. intraobserver variation
C. interobserver variation
D. Validity
14





Another word for accuracy is . . .
A. reliability
B. precision
C. variance
D. validity
RANDOMIZED TRIALS:
SOME FURTHER ISSUES
Chapter 8
Sample Size

How many subjects do we need to study?
Four Possibilities in Testing Whether the
Treatments Differ




1. The treatments do not differ, and we correctly
conclude that they do not differ. Correct decision
2. The treatment do not differ, but we conclude that
they do differ. Type I error = alpha
3. The treatments differ, but we conclude that they
do not differ. Type II error = beta
4. The treatments do differ, and we correctly
conclude that they do differ. Correct decision
Power of the Study




Beta = probability of type 2 error
Alpha = probability of type 1 error
1- beta = power of the study
Power is the probability of correctly concluding that
the treatments differ
Estimate the Sample Size Needed in a
Random Trial

Need to specify
 The
difference in response rates to be detected
 An estimate of the response rate in one of the groups
 Level of statistical significance (alpha)
 The value of the power desired (1 minus beta)
 Whether the test should be one-sided or two-sided
The Hypertension Detection and
Follow-up Program







Veterans Administration
Is antihypertensive therapy of benefit for those with
only slightly elevated blood pressure?
22,994 subjects with elevated diastolic blood pressure
Treatment group: stepped care
Comparison group: referred care
Stepped care resulted in lower mortality
Encouraged MDs to treat mild to moderate
hypertension
M.R.F.I.T.





Multiple Risk Factor Intervention Trial
Designed to determine whether mortality from
myocardial infarction could be reduced by changes
in lifestyle and other measures
SI = special intervention (stepped care for
hypertension, and education about lifestyle
changes)
UC = comparison group (usual care)
Differences in risk factors however no significant
difference in CHD mortality
M.R.F.I.T. (continued)


control group may have been ‘contaminated’ by
widespread changes in population
Intermediate measures (reduction in smoking,
changes in blood pressure and cholesterol levels) do
not show effect if CHD mortality did not change
Study of Breast Cancer Prevention





1992-1997
13,388 women 35 years and older
Randomly assigned to placebo or tamoxifen
Tamoxifen group had lower rates of breast cancer
yet higher rates of endometrial cancer
Benefit vs. Harm?
Phases in Testing New Drugs in the US

Phase I: clinical Pharmacologic studies
20-80 patients
 Toxic and pharmacologic effects


Phase II: Clinical Investigations
100 to 200 patients
 Efficacy and relative safety


Phase III: Large-Scale Randomized Control Trials
Multicentered
 Effectiveness and safety
 If passes can be marketed


Phase IV: Post Marketing Surveillance
Ethical Considerations




Randomization is ethical only if we do not have
adequate evidence to support the conclusion that
drug A is better than drug B
Placebo, withholding treatment?
Informed Consent while distressed by new
diagnosis?
Unforeseen risks?
Randomized Trials


To evaluate new forms of interventions
To evaluate interventions that are highly
controversial or that have been widely used or
recommended without having been adequately
evaluated
Arthroscopic Knee Surgery for Arthritis

Arthroscopic surgery





Lavage or debridement of the
knee joint
Compared to no treatment,
improved knee pain
Yet, compared to saline injection,
same improvement
Perceived benefits?
Results of randomized trial
showed that surgery is no better
than placebo
Group Psychosocial Support for
Patients with Breast Cancer





1989
Women with metastatic breast cancer
Randomly assigned to supportive-expressive
therapy or control group
Survival was not different
Mood and pain perception improved
Registration of Clinical Trials


Publication bias – only
positive results are published
Registration in a registry
accessible to the public at no
charge is required before
any clinical trial will be
considered for publication by
any major journal
1





If a study concludes that treatments are different
when in reality they are not different, there is a . . .
A. type I error
B. type II error
C. correct decision
D. strong bias
2





When a study finds that treatment are not different
yet in reality they are different there is a
A. type I error
B. type II error
C. correct decision
D. bias
3





The ‘power’ of a study is calculated
A. as the probability of making a type I error
B. as the probability of making a type II error
C. as 1 minus the probability of making a type I
error
D. as I minus the probability of making a type II
error
4





The number of subjects needed in a study is
determined by
A. The projected difference in cure rates
B. The risk of making a type I error
C. The risk of making a type II error
D. all of the above
5





The Hypertension Detection and Follow-up Program
concluded that
A. Only those with severe high blood pressure should be
treated
B. Even those with moderate high blood pressure
benefit from treatment
C. Blood pressure is not a risk factor for heart disease
D. Only veterans need stepped care
6





The M.R.F.I.T. found that
A. no statistical differences were found between groups
for CHD mortality
B. no statistical differences were found between groups
for risk factors for CHD
C. statistical differences were found between groups for
all-cause mortality rates
D. statistical differences were found between groups for
gender
7





Tamoxifen
A. reduces the rate of endometrial cancer
B. reduces the rate of ovarian cancer
C. reduces the rate of breast cancer
D. reduces the rate of cervical cancer
8





A new drug can be marketed in the US after is
passes
A. Phase I drug testing
B. Phase II drug testing
C. Phase III drug testing
D. Phase IV drug testing
9





Which of the following is true about arthroscopic
knee surgery?
A. it results in more pain than no treatment
B. it results in less pain than placebo intervention
C. it results in less pain because the patient
perceives a benefits
D. it results in greater physical functioning than the
placebo
10





Group psychosocial support for patients with
metastatic breast cancer
A. prolongs survival
B. improves mood
C. increases pain
D. is always recommended for patients
CASE-CONTROL STUDIES AND
OTHER STUDY DESIGNS
Chapter 10
Design of a Case-Control Study




Identify a group of persons with a certain disease
(CASES)
Identify a similar group of persons without the
disease (CONTROLS)
Determine the proportion that were EXPOSED to a
particular factor and those who were NOT
EXPOSED in each group
Compare the exposure rates.
Case-Control Study
Selection of Cases






Hospital patients
Patients in physician’s practices
Clinic patients
Community registries
Cases from a single hospital may have unique risk
factors
Hospital patients may have more acute illness than
those in the community
Incident or Prevalent Cases






Incident cases are better for assessing etiology
Need to wait for diagnosis
Excludes those who die before diagnosis
Prevalent cases are better for assessing survival
Already diagnosed
If those diagnosed die soon, prevalence may
underrepresent some cases
Selection of Controls



Rate of exposure in control group
should mirror the rate in the
general population
Hospital patients may
overrepresent exposure
Controls: similar to cases in all
other respects other than having
the disease? OR representative
of the population from which
cases are selected?
Sources of Controls

Nonhospitalized persons in community
 Neighborhood
 Best

Friends
Patients admitted for some other disease
 Sample
of all admissions
 Sample with particular reason for admission, however,
admission should not affect exposure
Matching







Economic status
Age
Race
Gender
Occupation
Group matching
Individual Matching
Group Matching


Also called frequency matching
By proportion of certain characteristics
Individual Matching



Also called matched pairs
Difficult to find a match if too many characteristics
are chosen
Cannot determine difference in exposure if
exposure variable is used as a control
Limitations in Recall

Human beings are
limited is their
ability to recall
information
Recall Bias



Cases may remember exposure more often than
controls
Rumination bias
Differential recall
Use of Multiple Controls


Same Type
Different Types
 Example
different cancer and no cancer
When is a case-control study
warranted?



First step in searching for a cause
Followed by a cohort study
Disease is rare
Case-Control Studies

Strengths
 Good
for rare diseases
 Requires relatively little time to conduct
 Possibility of exploring multiple exposures
 Inexpensive

Weaknesses
 Reliance
on recall and/or historical data on exposure
 Temporality can be difficult to establish
 Comparability of cases and controls
Case-Control Study based on a
Defined Cohort


Population is identified and followed over time
A Case-Control Study is carried out using cases and
control from the original cohort
 Nested
case-control studies
◼ Controls
selected at the time cases are diagnosed
 Case-Cohort
◼ Controls
began
studies
randomly selected from cohort which the study
Case-Crossover Design



Each subject serves as own control
Different periods of time with high and low
exposure are compared
Used if suspected exposure is transient and occurs
over a short time
Cross Sectional Studies



Also called a prevalence study
Disease and exposure are measured at the same
time
Cannot establish temporal sequence
1





If a group of people are first identified by
exposure then followed forward to see if they
develop a disease, they are part of a . . .
A. Case-control study
B. Randomized Trial
C. Cohort Study
D. Cross Sectional Study
2





If a group of people are assigned to treatment
groups then examined they are part of a . . .
A. Case-control study
B. Randomized Trial
C. Cohort Study
D. Cross Sectional Study
3





If a group of people are first identified by disease
status then asked questions to determine their past
exposures, they are part of a . . .
A. Case-control study
B. Randomized Trial
C. Cohort Study
D. Cross Sectional Study
4





If a group of people are given a survey to
determine both disease status and exposure status
at the same time, they are part of a . . .
A. Case-control study
B. Randomized Trial
C. Cohort Study
D. Cross Sectional Study
5





If a group of people are evaluated during time of
high and low stress for migraine headaches, they
are part of a . . .
A. Cross-Sectional Study
B. Case-Cohort Study
C. Case-Crossover Study
D. Embedded Case-Cohort Study
6





If the cases in a case-control study are matched by
proportion of characteristics to the controls, the
researchers have done . . .
A. Group matching
B. Individual matching
C. Excessive matching
D. No matching
7





A recall bias is due to . . .
A. hospital patients have more serious illness than
community members
B. faulty memory
C. deception
D. blinding
8





Rumination bias refers to . . .
A. cows chewing their cud cannot remember details
B. selective memory
C. cases recalling exposures with more frequency
than controls
D. controls remembering more exposure than cases

Purchase answer to see full
attachment