1 A

w
4
5
6
7
8
[;J
10
11
12
13
14
15
16
17
[m
19
20
.
-~
22
23
24
25
26
c:]
28
29
30
31
32
33
34
35
36
37
38
39
40
Item: 1 of 44
1111 PMark
Q.ld: 3880 (
<J
C>-
Previous
Ne><t
jf
Lab Values
A new screening test was devised to detect pancreatic cancer at early stages using a serum marker
(CA19-9) of the disease. A study of this new test showed that its use prolongs the survival of patients with
pancreatic cancer by several months. The researchers concluded that use of the test improves the
prognosis of patients with pancreatic cancer. Which of the following is a potential problem with this
conclusion?
r
r
.., r
r
r
A. Observer's bias [4%]

B. Measurement bias [4%]
C. Lead-time bias [81 %]
D. Confounding [5%]
E. Ascertainment bias [5%]
User ld:
Explanation:
Lead-time bias should always be considered while evaluating any screening test. This bias occurs when
there is an incorrect assumption or conclusion of prolonged apparent survival and better prognosis due to a
screening test. What actually happens is that detection of the disease was made at an earlier point in time.
but the disease course itself or the prognosis did not change. so the screened patients appeared to live longer
from the time of diagnosis to the time of death. (USMLE tip: think of lead-time bias when you see "a new
screening test" for poor prognosis diseases like lung or pancreatic cancer.)
(Choices A, B and E) Observer's bias. measurement bias and ascertainment bias refer to misclassification
of an outcome and/or exposure (e.g .. labeling diseased subjects as non-diseased and vice versa) and are
related to the design of the study. The scenario described does not mention how the study was designed.
(Choice D) Although the results of the study could be potentially confounded. there is no information on how
potential confounders were treated during the design or analysis stage of this study.
Educational Objective:
Understand the concept of lead-time bias in screening tests. The typical example of lead-time bias is
prolongation of apparent survival in patients to whom a test is applied. without changing the prognosis of the
disease.
Time Spent 11 seconds
Copyright USMLEWorld ,LLC.
Last updated [7/26/20 13]
9-
Notes
1.~
Calculator
Item: 2 of 44
11 PMark
Q.ld: 3998 (
6
7
8
:[lJ
11
12
13
14
15
16
17
liBl
~
20
1211
. LBJ
23
24
25
26
1271
~
<:J
1>-
Previous
Ne><t
jf
Lab Values
A study was conducted to establish the average level of total serum cholesterol in a group of patients with
acute coronary syndrome. The study results were reported separately for males and females. The mean
total cholesterol level was 230 mg/dl for males and 220 mg/dl for females. Which of the following is the best
statistical method to compare the average cholesterol level for males and females?
., r
r
r
r
r
A. Two-sample t test [58%]

B. Two-sample z test [7%]
C. Analysis of variance (ANOVA) [8%]
D. Chi-square test [23%]
E. Meta-analysis [3%]
Explanation:
User ld:
The two-sample t test is commonly employed to compare two means. Several statistical approaches can be
used. butthe basic requirements needed to perform this test are the two mean values. the sample variances.
and the sample size. The t statistic is then obtained to calculate the p value. If the 'p' value is less than 0.05 .
the null hypothesis (that there is no difference between two groups) is rejected. and the two means are
assumed to be statistically different. If the 'p' value is large. the null hypothesis is retained.
(Choice B) The two-sample z test can also be used to compare two means. but population (not sample)
variances are employed in the calculations. Because population variances are not usually known. the test has
limited applicability.
29
(Choice C) The AN OVA (i.e .. analysis of variance) is used to compare three or more means.
30
31
(Choice D) The chi-square test is appropriate for categorical data and proportions.
32
33
34
35
36
37
38
39
40
(Choice E) Meta-analysis is an epidemiologic method of pooling the data from several studies to do an
analysis having a relatively big statistical power.
The two-sample t test is a statistical method that is commonly employed to compare the means of two
groups of subjects.
9-
Notes
1.":~
Calculator
Item: 3 of 44
5
6
7
8
:[lJ
11
12
13
14
15
16
17
Coronary-Artery Bypass Surgery in Patients with Left

Ventricular Dysfunction
Is coronary-artery bypass grafting (CABG) superior to optimal

medical therapy alone in the treatment of patients with coronary
artery disease and heart failure?
Methods:
Design: Randomized clinical trial
20
Median follow-up: 56 months
23
24
25
26
1271
~
29
30
31
32
33
34
35
36
37
38
39
40
jf
Lab Values
A 67-year-old male with recently diagnosed coronary artery

disease comes for follow-up. He was hospitalized for
congestive heart failure 4 weeks ago and underwent coronary
angiography showing significant narrowing of left anterior
descending and right coronary arteries. His left ventricular
ejection fraction is 30%. Currently, he can walk 3 blocks on a
flat surface limited by chest discomfort and shortness of
breath. The physician tries to optimize his medical regimen.
He asks whether surgery is 'the right option' for him. Based on
the study results. which of the following is the best statement?
Hypothesis:
Blinding: none
1211
. LBJ
C>Ne><t
Item 1 of2
liBl
~
-<:1
Previous
11 f>Mark
Q.ld: 7690 (
r A. Medical therapy comes with a risk of exacerbating

heart failure [3%]
r B. Once you have surgery you might need more

long-term procedures [3%]
Setting: Multicenter (99 medical centers in 22 countries)
., r C. Surgery carries a higher initial risk of dying [41%]

r D. Surgery would not affect your overall long-term
Patients: Patients with a left ventricular ejection fraction of 35% or

less and coronary artery disease amenable to CABG. Patients with
a stenotic lesion leading to loss of 50% or more of the diameter of
the left main coronary artery and/or Canadian Cardiovascular
Society class Ill or IV angina while receiving medical therapy were
excluded from the study. (The Canadian Cardiovascular Society
angina classification ranges from class 0. which indicates no
symptoms, to class IV, which indicates angina at any level of
physical exertion.)
well-being [48%]
r E. The benefits of surgery are limited to preventing a

heart attack [5%]
Explanation:
The study above is a randomized clinical trial that looks at the

benefit of CABG with medical therapy versus medical therapy
alone for the management of significant coronary artery
disease and heart failure (defined as less than 35% left
ventricular ejection fraction). The study's primary end point
was overall mortality.
Intervention: CABG plus optimal medical therapy vs optimal medical

therapy alone
Outcome measures: The primary outcome was the rate of death
from any cause. Major secondary outcomes included the rates of
death from cardiovascular causes and of death from any cause or
hospitalization for cardiovascular causes.
User ld:
The results are reported as a hazard ratio. which is the ratio of

an event rate occurring. ~n the treatment group co';lpare~ to an
Notes
1.~
Calculator
5
6
7
8
:[lJ
11
12
13
14
15
16
17
f1sl
-~
20
:I;~ I
23
24
25
26
:I ;~I
29
30
31
32
33
34
35
36
37
38
39
40
Item: 3 of 44
Q. ld: 7690
\' Mark
<j
t:>
G
::::l'
Previous
NeKt
Lab Values
Results:
A total of 1212 patients were randomly assigned to medical therapy
alone (602 patients) or medical therapy plus CABG (610 patients).
By the end of the follow-up period. 100 patients in the
medical-therapy group ( 17%) underwent CABG, and 555 patients in
the CABG group (91%) underwent CABG.
Primary ou1come (CABG vs medical therapy alone)
Rate of death from any cause Hazards Ratio (HR) 0.86 (0.72-1.04)
P= 0.12
Secondary outcomes (CABG vs medical therapy alone)
Death from any cause within 30 days after randomization HR 3.1 2
(1.33-7.31) P=0.006
Death from cardiovascular causes HR 0.81 (0.66- 1.00) P=0.05
Death from any cause or hospitalization for heart failure HR 0.84
(0.71-0.98) P=0.03
Death from any cause or hospitalization for cardiovascular causes
HR 0.74 (0.64-0.85) P<0.001
Death from any cause or hospitalization for any cause HR 0.81
(0.71-0.93) P=0.003
Death from any cause or revascularization HR 0.60 (0.51-0.71)
P<0.001
Conclusion:
In this randomized trial. there was no significant difference between
medical therapy alone and medical therapy plus CABG with respect
to the primary end point of death from any cause.
Funding Source: Supported by grants from the National Heart. Lung.

and Blood Institute and by Abbott Laboratories.
Structured abstract is based orr. NEJM 2011 ;364(17) 1607-16.
The results are reported as a hazard ratiO , wh1ch 1s the ratiO of

an event rate occurring in the treatment group compared to an
event rate occurring in the non-treatment group. A ratio less
than 1 indicates that the treatment group had a significantly
lower event rate while values greater than 1 indicate that the
treatment group had a much higher event rate .
In this study, the hazard ratio for all cause mortality was 0.86,
bu1 the p value was not statistically significant (not less than
om). This means that surgery overall did not improve
all-cause mortality during the average follow-up period of 56
months. CABG was also associated with a higher mortality in
the first 30 days after randomization with a hazard ratio of 3.12,
with a statistically significant p value less than O.D1 . This is due
to the inherent immediate risk of surgery and should be
explicitly conveyed to the patient All the other secondary
outcomes were favorable for the CABG group.
(Choice A) The hazard ratio for death from any cause or
hospitalization for heart failure was 0.84, which means that the
surgery (in addition to medical therapy) likely improves heart
failure symptoms. This should not be interpreted as medical
therapy worsening heart failure since medical therapy was
used in both groups, and it was the addition of surgery that
improved it.
(Choice B) The hazard ratio for death from any cause or
revasculari zation was 0.60, which indicates that the surgery
added to medical therapy likely leads to fewer revascularization
procedures than medical therapy alone.
(Choice D) The ha zard ratio for death from any cause or
hospitalization for any cause was 0.81. which indicates that the
surgery added to medical therapy likely leads to fewer
hospitalizations with better long-term well being than medical
therapy alone.
.V
~
Notes
... j
,...
Calculator
Item: 3 of 44
Q.ld: 7690 (
5
6
7
8
:[lJ
11
12
13
14
15
16
17
liBl
~
20
1211
. LBJ
23
24
25
26
1271
~
29
30
31
32
33
34
35
36
37
38
39
40
!II PMark
<J
t>-
Previous
Ne><t
Results:
A total of 1212 patients were randomly assigned to medical therapy

alone (602 patients) or medical therapy plus CABG (61 0 patients).
By the end of the follow-up period. 1DO patients in the
medical-therapy group ( 17%) underwent CABG. and 555 patients in
the CABG group (91%) underwent CABG.
Primary outcome (CABG vs medical therapy alone)
Rate of death from any cause Hazards Ratio (HR) 0.86 (0. 72-1 .04)
P= 0.12
Secondary outcomes (CABG vs medical therapy alone)
Death from any cause within 30 days after randomization HR 3.12
(1.33-7 .31) P=0.006
Death from cardiovascular causes HR 0.81 (0 .66-1 .DO) P=O .05
Death from any cause or hospitalization for heart failure HR 0.84
(0. 71-0 .98) P=O .03
Death from any cause or hospitalization for cardiovascular causes
HR 0. 74 (0 .64-0 .85) P<O .DO 1
Death from any cause or hospitalization for any cause HR 0.81
(0. 71-0 .93) P=O .003
Death from any cause or revascularization HR 0.60 (0 .51-0.71)
P<0.001
jf
Lab Values
used in both groups. and it was the addition of surgery that

improved it.
(Choice B) The hazard ratio for death from any cause or
revascularization was 0.60. which indicates that the surgery
added to medical therapy likely leads to fewer revascularization
procedures than medical therapy alone.
(Choice D) The hazard ratio for death from any cause or
hospitalization for any cause was 0.81 . which indicates that the
surgery added to medical therapy likely leads to fewer
hospitalizations with better long-term well being than medical
therapy alone.
(Choice E) All the secondary outcomes except 30-day
mortality were favorable for surgery with medical therapy,
indicating that there are other benefits to surgery with medical
therapy other than just preventing heart attacks.
Educational objective:
Hazard ratios are the ratio of an event rate occurring in the

treatment arm versus the non-treatment arm. Ratios less than
1 indicate that the treatment arm had a lower event rate while
ratios higher than one indicate the treatment arm had a higher
rate of events.
Conclusion:
In this randomized trial, there was no significant difference between

medical therapy alone and medical therapy plus CABG with respect
to the primary end point of death from any cause.
Funding Source: Supported by grants from the National Heart, Lung.

and Blood Institute and by Abbott Laboratories.
References:
1. Biostatistics primer: what a clinician ought to

know: hazard ratios.
2. Estimation of the 2 -sample hazard ratio
function using a semi parametric model.
Structured abstract is based on NEJM 2011; 364( 17) 1607-16.

Time Spent 23
seconds
Copyright
USMLEWorld ,LLC.
Last updated:
[7/1/2013]
1.":~
Notes
Calculator
Item: 4 of 44
<:1
C>-
Previous
Ne><t
11 PMark
Q.ld: 7691 (
jf
Lab Values
ltem2 of2
6
7
8
:[lJ
11
12
13
14
15
16
17

Hypothesis:

Blinding: none
20
1211
. LBJ
23
24
25
26
1271
~
29
30
31
32
33
34
35
36
37
38
39
40
r
r
r
r
.., r
Methods:
liBl
~
As reported by investigators. by the end of the follow-up period.

17% of patients in the medical-therapy group underwent CABG.
and 91% of patients in the CABG group underwent CABG.
Analyzing outcomes for patients from both groups who actually
underwent CABG versus received only medical therapy might
introduce which of the following?
A. Lead-time bias [17%]

B. Measurement bias [34%]
C. Observer bias [13%]
D. Recall bias [4%]
E. Susceptibility bias [32%]
User ld:
Explanation:

physical exertion.)
This study was a randomized clinical trial, but many patients in

the medical therapy only arm also underwent surgery. By
combining all surgical from both groups in one group and then
comparing against medical therapy may introduce susceptibility
bias. which is a subgroup of selection bias that occurs when
the treatment regimen selected for a patient depends on the
severity of the patient's condition. This type of bias fails to take
into account other confounding variables that may be
accounting for the patient's condition. This bias also negates
the benefits of randomization. which usually avoids selection
bias and confounding variables.

therapy alone
In this study, 17% of the patients in the medical therapy group

were not adequately treated with medical therapy and were
selected to have CABG. There may have been other
confounding variables that caused this group of patients to fail
medical t~erapy that were not anal~zed in this study. ~y
""
Notes
1.":~
Calculator
Item: 4 of 44
6
7
8
:[lJ
11
12
13
14
15
16
17

Methods:
23
24
25
26
1271
~
29
30
31
32
33
34
35
36
37
38
39
40
(Choice A) Lead time bias happens when two interventions

are compared to diagnose a disease. and one intervention
diagnoses the disease earlier than the other without an effect
on the outcome (such as survival). This would make it appear
that the intervention prolonged survival when it really just
diagnosed the disease sooner.
20
(Choice B) Measurement bias occurs from poor data

collection with inaccurate results. which is not what is
described in this case.

physical exertion.)
(Choice C) Observer bias occurs when the observer may be

influenced by prior knowledge or details of the study that can
affectthe results. Blinded studies usually avoid this bias by
preventing the observers from knowing which treatment or
intervention the participants are receiving. and also measuring
objective outcomes (such as mortality).
(Choice D) Recall bias occurs when a study participant is
affected by prior knowledge to answer a question. This is more
common in case-control studies rather than randomized
clinical trials.

therapy alone
jf
Lab Values
1.":~
Notes
Calculator
In this study, 17% of the patients in the medical therapy group

were not adequately treated with medical therapy and were
selected to have CABG. There may have been other
confounding variables that caused this group of patients to fail
medical therapy that were not analyzed in this study. By
grouping them with the patients in the surgery group. this may
lead to erroneous analysis and introduce more confounding
factors. To avoid this in studies. patients are randomized and
analyzed separately to account for these possible confounding
variables.
Hypothesis:
Blinding: none
1211
. LBJ
C>Ne><t

liBl
~
<:J
Previous
!il P' Mark
Q.ld: 7691 (
Susceptibility bias is a type of selection bias where a treatment
regimen is selected for a patient based on the severity of their
Item: 4 of 44
6
7
8
:[lJ
11
12
13
14
15
16
17

(Choice B) Measurement bias occurs from poor data

collection with inaccurate results. which is not what is
described in this case.
Methods:
(Choice C) Observer bias occurs when the observer may be

influenced by prior knowledge or details of the study that can
affectthe results. Blinded studies usually avoid this bias by
preventing the observers from knowing which treatment or
intervention the participants are receiving. and also measuring
objective outcomes (such as mortality).
20
23
24
25
26
1271
~
29
30
31
32
33
34
35
36
37
38
39
40
jf
Lab Values
Hypothesis:

Blinding: none
1211
. LBJ
[:>Ne><t
(Choice A) Lead time bias happens when two interventions

are compared to diagnose a disease. and one intervention
diagnoses the disease earlier than the other without an effect
on the outcome (such as survival). This would make it appear
that the intervention prolonged survival when it really just
diagnosed the disease sooner.
liBl
~
<J
Previous
1111 f' Mark
Q.ld: 7691 (
(Choice D) Recall bias occurs when a study participant is

affected by prior knowledge to answer a question. This is more
common in case-control studies rather than randomized
clinical trials.

physical exertion.)
Susceptibility bias is a type of selection bias where a treatment
regimen is selected for a patient based on the severity of their
condition. without taking into account other possible
confounding variables.
References:

therapy alone
1. The role of susceptibility bias in epidemiologic

research.
Time Spent 7
seconds
Copyright
USMLEWorld ,LLC.
Last updated:
[7/20/2013]
Notes
1.":~
Calculator

2
3
4
6
7
8
[;J
10
11
12
13
14
15
16
17
[m
19
20
.
-~
22
23
24
25
26
c:]
28
29
30
31
32
33
34
35
36
37
38
39
40
Item: 5 of 44
11 PMark
Q.ld: 3885 (
<:J
1>-
Previous
Ne><t
jf
Lab Values
A prospective cohort study revealed a strong positive association between smoking and liver cirrhosis (relative
risk 2 .8). The researchers then divided the cohort into two groups: alcohol consumers and
non-consumers. Subsequent statistical analysis did not reveal any association between smoking and liver
cirrhosis with either group. The scenario described above is an example of which of the following?
r
r
r
r
.., r
A. Selection bias [12%]

B. Observer's bias [4%]
C. Measurement bias [6%]
D. Recall bias [2%]
E. Confounding [76%]
Explanation:
User ld:
Confounding refers to the bias that results when the exposure-disease relationship is mixed with the effect of
extraneous factors (i.e .. confounders). Confounders influence both the exposure and outcome. In the given
study, crude analysis of the data initially revealed an association between smoking and liver cirrhosis. The
most likely potential confounder is concomitant alcohol consumption in people who smoke. It is a well-known
fact that alcohol consumption is strongly associated with liver cirrhosis. Furthermore. alcohol consumption
can explain at least part of the association observed between smoking and liver cirrhosis.
Methods to deal with confounding include matching of cases and controls based on the confounding factor, or
stratification of the study population based on the confounding factor. In this case, running separate analyses
for alcohol consumers and non-consumers (this technique is called stratified analysis) can unmask
confounding and disclose the true unconfounded value of the RR.
(Choice A) Selection bias results from the manner in which people are selected for the study, or from the
selective losses from follow-up. The scenario does not mention any of these problems.
(Choices B and C) Observer's bias and measurement bias distort the measure of association by
misclassifying exposed/unexposed and/or diseased/non-diseased subjects. The scenario does not describe
this classification process.
(Choice D) Recall bias results from the inaccurate recall of past exposure by people in the study. It applies
mostly to case-control studies. not cohort studies.
Educational Ojective:
9-
Notes
1.":~
Calculator

2
3
4
6
7
8
[;J
10
11
12
13
14
15
16
17
[m
Item: 5 of 44
11 PMark
Q.ld: 3885 (
r
r
r
.., r
<:1
1>-
Previous
Ne><t
jf
Lab Values
B. Observer's bias [4%]

C. Measurement bias [6%]
D. Recall bias [2%]
E. Confounding [76%]
Explanation:
User ld:
Confounding refers to the bias that results when the exposure-disease relationship is mixed with the effect of
extraneous factors (i.e .. confounders). Confounders influence both the exposure and outcome. In the given
study, crude analysis of the data initially revealed an association between smoking and liver cirrhosis. The
most likely potential confounder is concomitant alcohol consumption in people who smoke. It is a well-known
factthat alcohol consumption is strongly associated with liver cirrhosis. Furthermore. alcohol consumption
can explain at least part of the association observed between smoking and liver cirrhosis.
19
. 20
-~
22
23
24
25
26
c:]
28
29
30
31
32
33
34
35
36
37
38
39
40
Methods to deal with confounding include matching of cases and controls based on the confounding factor. or
stratification of the study population based on the confounding factor. In this case. running separate analyses
for alcohol consumers and non-consumers (this technique is called stratified analysis) can unmask
confounding and disclose the true unconfounded value of the RR.
(Choice A) Selection bias results from the manner in which people are selected for the study, or from the
selective losses from follow-up. The scenario does not mention any of these problems.
(Choices B and C) Observer's bias and measurement bias distort the measure of association by
misclassifying exposed/unexposed and/or diseased/non-diseased subjects. The scenario does not describe
this classification process.
(Choice D) Recall bias results from the inaccurate recall of past exposure by people in the study. It applies
mostly to case-control studies. not cohort studies.
Educational Ojective:
Know the concept of confounding. Distinguish between crude and adjusted measures of association.
Confounding refers to the bias that can result when the exposure-disease relationship is mixed with the effect
of extraneous factors (i.e .. confounders).
9''
Notes
1.":~
Calculator
:CJ
Item: 6 of 44
11 PMark
Q.ld: 3960 (
<:1
1>-
Previous
Ne><t
jf
Lab Values
A prospective cohort study was conducted to assess the relationship between caffeine consumption and the
risk of colon cancer in middle-aged women. The study showed that caffeine consumption decreases the risk
of colon cancer with a relative risk of 0.83 and p value of 0.04. The factthatthe study was conducted
specifically in middle-aged women raises most concerns regarding which of the following issues?
:[lJ
11
12
13
14
15
16
17
liBl
~
20
1211
. LBJ
23
24
25
26
1271
~
29
30
31
32
r
., r
r
r
r
A. Bias [1 0%]
B. Generalizability [84%]
C. Internal validity [3%]
D. Reliability [3%]
E. Reproducibility [1 %]
Explanation:
User ld:
The generalizability, or external validity, of a study is defined as the applicability of the obtained results beyond
the cohort that was studied. External validity answers the question. "How generalizable are the results of a
study to other populations?" In this study, the external validity seems to be quite limited due to the restriction
of the study population to middle-aged women. In other words. because the cohort is restricted to
middle-aged women. the results of the study are applicable only to middle-aged women. This is an important
concept when applying the results of observational studies and clinical trials to everyday practice if the original
inclusion and exclusion criteria are not reviewed carefully. The other answer options are relevant for internal
validity answering the following question: "Are the results obtained in this specific cohort valid?"
(Choice A) In this scenario. there is no description of how the study was designed. conducted. and analyzed.
so it is not possible to determine whether the internal validity was violated (ie. if bias is present).
(Choice C) Within this cohort (middle-aged women). the study could be valid. This property is called internal
validity, or validity as generally taught
33
34
(Choices D and E) Reliability or reproducibility is the measure of random error.
35
36
37
38
39
Generalizability or external validity pertains to the applicability of study results to other populations (eg. the
results of a study in middle-aged women would not be expected to be applicable to elderly men).
40
Time Soent: 2 seconds
Coovriaht USMLEWorld .LLC.
Last uodated f7/15/20 131
9''
Notes
1.":~
Calculator
:CJ
5
6
Item: 7 of 44
HRT
NoHRT
Total
:[lJ
11
liBl
~
20
1211
. LBJ
23
24
25
26
1271
~
29
30
31
1>Ne><t
jf
Lab Values
A study was conducted to assess the association between hormone replacement therapy (HRT) in
post-menopausal women and the level of serum C-reactive protein (CRP). The data from the study are
presented below:
12
13
14
15
16
17
<:J
Previous
11 PMark
Q.ld: <W79 (
CRP high
32
28
60
CRP normal
Total
41
49
90
73
77
150
Which of the following is the best statistical method to assess the association between HRT and elevated
CRP levels?
r
r
., r
r
r
A. Two-sample z-test [6%]

B. Two-sample t-test [33%]
C. Chi-square test [53%]
D. ANOVA [5%]
Explanation:
User ld:
The chi-square test is used to compare the proportions of a categorized outcome. In this case, the outcome
(serum CRP level) is categorized as either "high" and "normal," and then presented with the exposure ("HRT"
or "no HRT") in a 2 x 2 table. In one of the commonly used chi-square tests, the observed values in each of
the cells are compared to expected (under the hypothesis of no association) values. If the difference between
the observed and expected values is large. an association between the exposure and the outcome is
assumed to be present.
32
33
34
(Choices A and B) The two-sample z-test and two-sample t-test are used to compare two means. not
proportions.
35
36
(Choice D) Analysis of variance (ANOVA) is used to compare the means of three or more variables.
37
38
39
40
9-
Notes
1.":~
Calculator
:CJ
Item: 7 of 44
5
6
:[lJ
11
12
13
14
15
16
17
<:J
1>-
Previous
Ne><t
11 PMark
Q.ld: <W79 (
nr<e1
.OL.
... ,
,.,
NoHRT
Total
28
49
60
90
77
150
jf
Lab Values
r
r
., r
r
r
B. Two-sample t-test [33%]

C. Chi-square test [53%]
D. ANOVA [5%]
Explanation:
1211
. LBJ
The chi-square test is used to compare the proportions of a categorized outcome. In this case, the outcome
(serum CRP level) is categorized as either "high" and "normal," and then presented with the exposure ("HRT"
or "no HRT") in a 2 x 2 table. In one of the commonly used chi-square tests, the observed values in each of
the cells are compared to expected (under the hypothesis of no association) values. If the difference between
the observed and expected values is large. an association between the exposure and the outcome is
assumed to be present.
1271
~
29
30
31
32
33
34
35
36
37
38
39
40
A. Two-sample z-test [6%]
User ld:
20
26
1.":~
Calculator
Which of the following is the best statistical method to assess the association between HRT and elevated
CRP levels?
liBl
23
24
25
9-
Notes
(Choices A and B) The two-sample z-test and two-sample t-test are used to compare two means. not
proportions.
(Choice D) Analysis of variance (ANOVA) is used to compare the means of three or more variables.
The chi-square test is used to compare proportions. A 2 x 2 table may be used to compare the observed
values with the expected values.
:CJ
5
6
12
13
14
15
16
17
Item: 8 of 44
11 PMark
Q.ld: 3886 (
<J
C>-
Previous
Ne><t
jf
Lab Values
A new test is devised to detect HIV. It has a sensitivity of 90% and specificity of 80% compared to the gold
standard. Consider that the test is used in two populations: a population in Africa having an HIV prevalence of
0.20 (20%). and a population in Asia having an HIV prevalence of 0.01 ( 1%). Which is the most accurate
statement concerning the new test?
r
r
., r
r
r
A. Sensitivity of the test is higher in the African population [8%]

B. Specificity of the test is higher in the Asian population [4%]
C. Positive predictive value of the test is higher in the African population [78%]
D. Negative predictive value of the test is lower in the Asian population [6%]
E. The test is not reliable in the African population [3%]
liBl
Explanation:
20
The traditional interpretation of positive predictive value (PPV) is the proportion of people with positive test
results that actually have the disease. It is easier to understand the concept in terms of probability: if a patient
has a positive test result, what is the probability that he/she has the disease? PPV depends on the
prevalence of the disease (very important concept to remember). The more common the disease (e.g .. 20%
prevalence of HIV in the African population). the greater the probability that a patient from that population and
with a positive test actually has the disease (i.e .. has a true positive result). In the Asian population. the
probability of a false-positive result for a patient with a positive test result is higher, because of the low
prevalence of HIV.
1211
. LBJ
23
24
25
26
1271
~
29
30
31
32
33
34
35
36
User ld:
(Choices A and B) The sensitivity and specificity of a test do not depend on the prevalence of the disease in
the population.
(Choice D) Like positive predictive value. negative predictive value depends on the prevalence of the disease
in the population. but has an inverse association with the prevalence. As the prevalence of the disease
increases. the negative predictive value decreases because the probability of a true negative result for a
patient who tested positive is high in a population with a low prevalence of the disease (i.e .. the Asian
population example).
37
38
39
(Choice E) The statement that the test is not reliable in the African population is not correct, because the test
has high sensitivity and specificity.
40
~,. .. ,.~tiftn~l
nhiat"tiua
9-
Notes
1.~
Calculator
:CJ
5
6
12
13
14
15
16
17
liBl
~
20
1211
. LBJ
23
24
25
26
1271
~
29
30
31
32
33
34
35
36
37
38
39
40
Item: 8 of 44
Q.ld: 3886 (
r
r
., r
r
r
11 PMark
<J
C>-
Previous
Ne><t
jf
Lab Values
A. Sensitivity of the test is higher in the African population [8%]

B. Specificity of the test is higher in the Asian population [4%]
C. Positive predictive value of the test is higher in the African population [78%]
D. Negative predictive value of the test is lower in the Asian population [6%]
E. The test is not reliable in the African population [3%]
User ld:
Explanation:
The traditional interpretation of positive predictive value (PPV) is the proportion of people with positive test
results that actually have the disease. It is easier to understand the concept in terms of probability: if a patient
has a positive test result, what is the probability that he/she has the disease? PPV depends on the
prevalence of the disease (very important concept to remember). The more common the disease (e.g .. 20%
prevalence of HIV in the African population). the greater the probability that a patient from that population and
with a positive test actually has the disease (i.e .. has a true positive result). In the Asian population. the
probability of a false-positive result for a patient with a positive test result is higher, because of the low
prevalence of HIV.
(Choices A and B) The sensitivity and specificity of a test do not depend on the prevalence of the disease in
the population.
(Choice D) Like positive predictive value. negative predictive value depends on the prevalence of the disease
in the population. but has an inverse association with the prevalence. As the prevalence of the disease
increases. the negative predictive value decreases because the probability of a true negative result for a
patient who tested positive is high in a population with a low prevalence of the disease (i.e .. the Asian
population example).
(Choice E) The statement that the test is not reliable in the African population is not correct, because the test
has high sensitivity and specificity.
Know the concept of positive predictive value (PPV). PPV depends on the prevalence of the disease of
interest in the population to which the test is applied. PPV increases with an increase in prevalence of
disease in the study population. For NPV, the inverse is true.
9-
Notes
1.~
Calculator
:CJ
5
6
7
8
11
12
13
14
15
16
17
The drug ad is focused on Efrenzia, a novel anti-platelet agent for the treatment of acute coronary syndromes.
Item 1 of2
A 58-year-old man with a history of hypertension and type 2 diabetes mellitus comes to the emergency
department because of chest pain and diaphoresis. The symptoms started two hours ago and have a
stuttering course. He has never had similar symptoms before. In the emergency department, his
electrocardiogram shows horizontal ST segment depression in leads V1 to V4. He is given the appropriate
medical therapy including low-dose aspirin, and referred to the catheterization laboratory due to persistence of
his angina. Based on the information provided in the drug ad, giving the patient Efrenzia as opposed to
clopidogrel would most likely decrease the risk of developing which of the following subsequent events?
liBl
View Drug Ad
20
1211
. LBJ
23
24
25
r
r
r
., r
A. Cardiovascular death [21%]

B. Major bleeding [6%]
C. Non-fatal stroke [14%]
D. Recurrent myocardial infarction [59%]
26
1271
~
29
30
31
32
33
34
35
36
37
38
39
40
Explanation:
User ld:
This drug ad is comparing the effect of administering Efrenzia vs. clopidogrel in combination with aspirin for
patients with acute coronary syndrome undergoing percutaneous coronary intervention, including those with
unstable angina (UA)/non-ST elevation myocardial infarction (NSTEMI) and ST-elevation Ml (STEMI). The
results are reported as the percentage of patients developing a composite endpoint of cardiovascular death,
non-fatal Ml, or nonfatal stroke over the subsequent 18 months. A subgroup analysis was also performed in
diabetics presenting with NSTEM/UA or STEML
There was a 2. 7% overall decrease of events in patients with STEMI ( 12.4% reduced to 9. 7%) and a 1.8%
decrease of events in UAINSTEMI patients ( 10. 7% reduced to 8.9%). The text under the section titled "Benefit
in STEMI and UAINSTEMI patients" indicates that the reason for the difference between the treatments was
primary due to a significant reduction in (recurrent) non-fatal Mls.
:CJ
5
6
7
8
11
12
13
14
15
16
17
liBl
~
20
1211
. LBJ
23
24
25
26
1271
~
Item: 9 of 44
11 PMark
Q.ld: 7708 (
-<:J
C>-
Previous
Ne><t
jf
Lab Values
r
r
r
., r
A. Cardiovascular death [21%]

B. Major bleeding [6%]
C. Non-fatal stroke [14%]
D. Recurrent myocardial infarction [59%]
Explanation:
User ld:
This drug ad is comparing the effect of administering Efrenzia vs. clopidogrel in combination with aspirin for
patients with acute coronary syndrome undergoing percutaneous coronary intervention. including those with
unstable angina (UA)/non-ST elevation myocardial infarction (NSTEMI) and ST-elevation Ml (STEMI). The
results are reported as the percentage of patients developing a composite endpoint of cardiovascular death.
non-fatal MI. or nonfatal stroke over the subsequent 18 months. A subgroup analysis was also performed in
diabetics presenting with NSTEM/UA or STEML
There was a 2. 7% overall decrease of events in patients with STEMI ( 12.4% reduced to 9. 7%) and a 1.8%
decrease of events in UAINSTEMI patients ( 10. 7% reduced to 8.9%). The text under the section titled "Benefit
in STEMI and UAINSTEMI patients" indicates that the reason for the difference between the treatments was
primary due to a significant reduction in (recurrent) non-fatal Mls.
30
31
that there was no significant difference between the treatments in cardiovascular death or non-fatal stroke in
both UAINSTEMI and STEMI patients.
34
35
36
37
38
39
40
View Drug Ad
(Choices A and C) The drug ad states under the section titled "Benefit in STEMI and UAINSTEMI patients"
33
1.":~
Calculator
clopldogrel would most likely decrease the nsk ot developing which ot the tollow1ng subsequent events't
29
32
""
Notes
(Choice B) Major bleeding was not part of the primary composite endpoint. Also. the safety data at the
bottom of the drug ad indicates that overall bleeding rates were significantly increased with Efrenzia.
When comparing the effects of a treatment on patient outcomes. it is important to note all causes of the
differences.
:CJ
5
6
7
Item: 10 of 44
Il l' Mark
Q.ld: 7709 [
<:1
C>-
Previous
Ne><t
if
Lab Values
ltem2 of2
Based on the drug ad data. how many diabetic patients undergoing percutaneous coronary intervention for
unstable angina (UA) or non-ST segment elevation myocardial infarction (NSTEMI) need to be treated with
Efrenzia to prevent one composite event as compared to clopidogrel?
View Drug Ad
12
13
14
15
16
17
liBl
r
r
., r
r
r
A. 5 [16%]
B. 10[11%]
C. 25 [59%]
D. 50 [1 0%]
E. 1DO [4%]
20
1211
. LBJ
23
24
25
26
Explanation:
The number needed to treat (NNT) is defined as the number of people that need to receive a treatment to
prevent one defined event. It is calculated as the inverse of the absolute risk reduction (ARR). In this study.
diabetic patients with UAINSTEMI had a 16.3% incidence of composite events with clopidogrel + aspirin and
12.3% with Efrenzia + aspirin.
1271
=0.163-0.123 =0.04
NNT =1/ARR =1/0.04 =25
ARR
~
29
30
31
32
33
User ld:
As compared to clopidogrel. Efrenzia would need to be given to at least 25 diabetic patients with UAINSTEMI
who are undergoing percutaneous coronary intervention in order to prevent one composite event (Choices A,
8, D, and E).
34
35
36
37
38
39
40
The NNT is defined as the number of people that need to receive a treatment to prevent one adverse event. It
is calculated as the inverse of the absolute risk reduction (ARR).
I References:
Notes
1.":~
Calculator
:CJ
Item: 10 of 44
!II PMark
Q.ld: 7709 [
<:1
(>-
Previous
Ne><t
liBl
~
20
1211
. LBJ
1271
~
29
30
31
32
33
34
35
36
37
38
39
40
Calculator
View Drug Ad
r
r
., r
r
r
A. 5 [16%]
B. 10[11%]
C. 25 [59%]
D. 50 [1 0%]
E. 1DO [4%]
Explanation:
User ld:
The number needed to treat (NNT) is defined as the number of people that need to receive a treatment to
prevent one defined event. It is calculated as the inverse of the absolute risk reduction (ARR). In this study.
diabetic patients with UAINSTEMI had a 16.3% incidence of composite events with clopidogrel + aspirin and
12.3% with Efrenzia + aspirin .
=0.163-0.123 =0.04
NNT =1/ARR =1/0.04 =25
ARR
23
24
25
26
9-
Notes
L...l I '-'1 IL.IU I.U J-'1 '-' Y '-'1 II. Ul 1'-' '-'UI I IJ-'U~II.'-' '-' Y '-'1 II. U~ '-'UI I IJ-'UI '-''-" I.U '-"UJ-'IUU\:::11 '-'1!
5
6
7
12
13
14
15
16
17
jf
Lab Values
As compared to clopidogrel. Efrenzia would need to be given to at least 25 diabetic patients with UAINSTEMI
who are undergoing percutaneous coronary intervention in order to prevent one composite event (Choices A,
8, D, and E).
The NNT is defined as the number of people that need to receive a treatment to prevent one adverse event. It
is calculated as the inverse of the absolute risk reduction (ARR).
References:
1. How effective is that treatment? The number needed to treat.
Copyright USMLEWorld .LLC.
.
.
Item: 11 of 44
Q.ld: 2135 [
11 f>Mark
<:::1
C>-
Previous
Ne><t
jf
Lab Values
5
6
7
8
12
13
14
15
16
17
liBl
In clinical practice. the "CAGE" questionnaire is used to screen patients for alcoholism. When a patient
replies with 2 out of 4 positive responses to the "CAGE" questions. the test is considered to be positive for
alcoholism. If this criteria is changed so that 3 out of 4 positive responses to the "CAGE" questions label the
patient as alcoholic. what is the effect on the sensitivity and specificity of this test?
r
r
., r
r
r
A. Both sensitivity and specificity of the test will decrease [3%]

B. Both sensitivity and specificity of the test will increase [7%]
C. Sensitivity will decrease but specificity will increase [73%]
D. Sensitivity will increase but specificity will decrease [14%]
E. Both sensitivity and specificity will remain the same [3%]
20
1211
. LBJ
23
24
25
User ld:
Explanation:
Raising the cut-off point of a test will increase its specificity and decrease its sensitivity. In this particular
vignette. increasing the number of questions for a patientto be labelled as an alcoholic will result in fewer
people being identified as such. Consequently. false and true positives will decrease. while false and true
negatives will increase.
26
29
Raising the cut-off point (e.g .. increasing the inclusion criteria) of a screening test results in an increase in
specificity and decrease in sensitivity.
30
31
1271
~
32
33
34
35
36
37
38
39
40
Last updated [11/1 0/2011]
9-
Notes
1.":~
Calculator
:CJ
5
6
7
8
:[lJ
11
14
15
16
17
Item: 12 of 44
11 f> Mark
Q.ld: 3947 [
-<J
t:>-
Previous
Ne><t
jf
Lab Values
A prospective cohort study was conducted to assess the effects of oral contraceptive use on the incidence of
breast cancer. Crude analysis of the study results suggests an association between the use of oral
contraceptives and breast cancer (relative risk [RR] 1.4. p 0.04). Further analysis shows that in women
with a family history of breast cancer. oral contraceptives increased the risk of breast cancer (RR 2.1 D. p
0.01 ). In women without a family history of breast cancer. no effect was observed (RR 1.05. p 0.80). The
results of this study are best explained by which of the following?
=
=
r
., r
r
r
r
A. Confounding [53%]
B. Effect modification [38%]
C. Latency period [3%]
D. Observer bias [2%]

E. Selection bias [4%]
liBl
~
20
Explanation:
1211
. LBJ
User ld:
Confounding versus effect modification
23
24
25
Confounding
Effect modification
26
Crude analysis
Crude analysis
1271
~
29
Alcohol
use
Significant association
Bladder
ca ncer
Use of oral
contraceptive
Breast
cancer
30
31
32
33
34
35
36
37
38
39
40
Stratified by smoking
Stratified by family history

Family history of breast cancer:
Smokers:
Alcohol
No association
Bladder
use
------------ ~
cancer
Use of oral
contraceptive
1-------
Strong association
Breast
cancer
Notes
~~
Calculator
:CJ
Item: 12 of 44
!il f>Mark
Q.ld: 3947 [
11
14
15
16
17
liBl
-~
20
1211
. LBJ
23
24
25
26
1271
-~
29
30
31
32
33
34
35
36
37
38
39
40
t:>Ne><t
jf
Lab Values
Notes
~]il
Calculator
Confound ing versus effect modification
5
6
7
8
:[lJ
-<J
Previous
Alcohol
use
Confounding
Effect modification
Crude analysis
Crude analysis
Bladder
ca ncer
Use of oral
contraceptive
Family history of breast cancer:
Smokers:
No as.sociation
------------ ~
Bladder
cancer
Non-smokers:
Alcohol
use
Breast
cancer
Stratified by family history
Stratified by smoking
Alcohol
use
Use of oral
contraceptive
Strong association
Breast
cancer
No family history of breast cancer:

No association
------------
Bladder
ca ncer
Use of oral
contraceptive
No asse<iatlon
------------
Breast
cancer
>VSMLEWOI"W.ll{
Effect modification results when an external variable positively or negatively impacts the effect of a risk factor
on the disease of interest. It can sometimes be confused with confounding. the bias that results when the
exposure-disease relationship is obscured by the effect of an extraneous factor that is associated with both
the exposure and disease. Effect modification can be distinguished from confounding by performing a
stratified analysis centering on the variable of interest. If the variable is a confounder. there will be no
significant difference in risk between the stratified groups as the confounding effects are now removed.
However. if the variable is instead an effect modifier. there will be a significant difference between the 2 groups.
In this case. stratification by family history shows that oral contraceptives significantly increase the risk of
breast cancer in patients with a positive family history but not in patients with a negative family history. Thus.
family history is not a confounding variable (Choice A). Rather. positive family history acts as an effect
..
:CJ
5
6
7
8
:[lJ
11
14
15
16
17
liBl
-~
20
1211
. LBJ
23
24
25
26
1271
-~
29
30
31
32
33
34
35
36
37
38
39
40
Item: 12 of 44
11 PMark
Q.ld: 3947 [
Non-smokers:
Alcohol
use
-<:J
1>-
Previous
Ne><t
jf
Lab Values
No family history of breast cancer:

No association
------------
Bladder
ca ncer
Use of oral
contraceptive
No association
------------
Breast
cancer
>VSMLEW11tHL<
on the disease of interest. It can sometimes be confused with confounding. the bias that results when the
exposure-disease relationship is obscured by the effect of an extraneous factor that is associated with both
the exposure and disease. Effect modification can be distinguished from confounding by performing a
stratified analysis centering on the variable of interest. If the variable is a confounder. there will be no
significant difference in risk between the stratified groups as the confounding effects are now removed.
However. if the variable is instead an effect modifier. there will be a significant difference between the 2 groups.
In this case. stratification by family history shows that oral contraceptives significantly increase the risk of
breast cancer in patients with a positive family history but not in patients with a negative family history. Thus.
family history is not a confounding variable (Choice A). Rather. positive family history acts as an effect
modifier to increase the risk of breast cancer in patients taking oral contraceptives. Other well-known
examples of effect modification include the effect of estrogens on the risk of venous thrombosis (augmented
by smoking) and the risk of lung cancer in people exposed to asbestos (also enhanced by smoking).
(Choice C) The latency period is the time required for a given exposure to have a measurable effect on the
outcome. This study provided no information on how long oral contraceptives must be used in order to have
an effect on breast cancer risk in susceptible patients.
(Choices D and E) Flaws in a study involving subject enrollment or data recording can result in selection
bias or observer bias. respectively. Effect modification is not a bias. but rather a natural phenomenon that is
importantto recognize.
on the disease of interest. It can be distinguished from confounding by performing a stratified analysis
centered on the variable of interest. Effect modification is not a bias. but rather is a natural phenomenon that
is important to recognize.
Notes
~~
Calculator
:CJ
5
6
7
8
:[lJ
Item: 13 of 44
ill PMark
Q.ld: 3897 [
<j
C>
Previous
Ne><t
jf
Lab Values
A new multidrug chemotherapy regimen significantly prolongs the survival in patients with lung cancer. If this
new regimen is widely implemented. what changes in the prevalence and incidence of lung cancer would you
expect?
A. Incidence will decrease. prevalence will decrease [1 %]
11
12
B. Incidence will increase. prevalence will not increase [2%]

14
15
16
17
C. Incidence will not change. prevalence will increase [91%]
liBl
~
20
1211
. LBJ
D. Incidence will not change. prevalence will not change [3%]
23
24
25
E. Incidence will decrease. prevalence will increase [2%]
26
1271
~
29
Explanation:
User ld:
30
31
32
33
34
35
36
37
38
39
40
It is important to understand the difference between prevalence and incidence. the two basic measures of
disease occurrence. Incidence is the measure of new cases. the rapidity with which they are diagnosed.
Prevalence is the measure of the fofa/ number of cases at a particular point in time. Incidence answers the
question: how many new cases of the disease developed in a population during a particular period of time?
Prevalence answers the question: how many cases of the disease exist in a population at a particular point in
time? The relationship between prevalence and incidence can be expressed as:
Prevalence
"T"I-- : .. _:.1- .. - -
=(Incidence) x (Time)
--
.1:----- ;_
-<~>
-1- ----.I 1. -- I.: ...I
- .... __ .....
--<~>
I.-----.,,_- .1:----- I.--
-I.-- .1.
Notes
Calculator
:CJ
5
6
7
8
:[lJ
Item: 13 of 44
!il f> Mark
Q.ld: 3897 [
<:J
C>-
Previous
Ne><t
jf
Lab Values
Notes
1.":~
Calculator
D. Incidence will not change. prevalence will not change [3%]
r
E. Incidence will decrease. prevalence will increase [2%]
11
12
14
15
16
17
liBl
~
20
1211
. LBJ
23
24
25
26
1271
~
29
30
31
32
33
34
35
36
37
38
39
40
Explanation:
User ld:
It is important to understand the difference between prevalence and incidence. the two basic measures of
disease occurrence. Incidence is the measure of new cases. the rapidity with which they are diagnosed.
Prevalence is the measure of the total number of cases at a particular point in time. Incidence answers the
question: how many new cases of the disease developed in a population during a particular period of time?
Prevalence answers the question: how many cases of the disease exist in a population at a particular point in
time? The relationship between prevalence and incidence can be expressed as:
Prevalence
The incidence of a disease is not changed by any kind of treatment. because the disease has already
occurred when treatment is started. On the other hand. the prevalence may be affected by treatment of the
disease. In this case. treatment of an acute and rapidly fatal disease (e.g .. lung cancer) prolonged the survival
of patients; however. such treatment did not cure the disease. This will result in more people having the said
disease at one point in time; hence. the prevalence will increase.
(Choices A, B and E) The incidence of a disease is not affected by the treatment.
(Choice D) This statement is not correct because a prolonged time of survival will increase the prevalence.
Know the difference between incidence and prevalence. Incidence is the measure of new cases. the rapidity
with which they are diagnosed. Prevalence is the measure of the total number of cases at a particular point in
time.
:CJ
5
6
7
8
:[lJ
Item: 14 of 44
Q.ld: 4157 [
!ll f> Mark
<J
C>-
Previous
Ne><t
jf
Lab Values
A cohort study was conducted to assess the relationship between high saturated fat consumption and the
occurrence of colorectal carcinoma among women. A group of women aged 40-65 was selected. The
baseline saturated fat consumption was calculated using a food questionnaire. and the cohort was followed
for seven years for the development of colon cancer. The study showed that women with high baseline
saturated fat consumption have four times the risk of colorectal cancer in a 7-year period. compared to
women with low fat consumption (RR = 4.0. 95% Cl = 1.5 - 6.5). According to the study results. what percent
of colorectal carcinoma in women with high fat consumption could be attributed to their diet?
11
16
17
liBl
r
r
.., r
r
r
A. 25% [34%]
B. 50% [8%]
c. 75% [42%]
D. 90% [8%]
E. 100% [8%]
20
1211
. LBJ
23
24
25
26
1271
~
29
30
Explanation:
User ld:
Attributable risk percent (ARP) or etiologic fraction is an important measure of the impact of a risk factor being
studied. ARP represents the excess risk in a population that can be explained by exposure to a particular risk
factor. It is calculated by subtracting the risk in the unexposed population (baseline risk) from the risk in the
exposed population. and dividing the result by the risk in the exposed population:
ARP =(risk in exposed- risk in unexposed)/risk in exposed.
An easier way to calculate the ARP is to derive it from the relative risk (RR):
31
32
33
34
35
36
37
38
39
40
ARP = (RR - 1)/RR.

In this case. ARP = (4.0- 1)/4.0 = 0.75 (75%). According to the study results. 75% of colorectal carcinoma in
the high consumption group was attributable to high saturated fat intake.
ARP represents the excess risk in the exposed population that can be attributed to the risk factor. It can be
easily derived from the relative risk using the following formula ARP = (RR - 1)/RR.
Notes
!:':~
Calculator
:CJ
5
6
7
8
:[lJ
11
16
17
Item: 14 of 44
Q.ld: 4157 [
!ll f> Mark
<J
C>-
Previous
Ne><t
jf
Lab Values
occurrence or colorectal carc1noma among women. A group or women age a 4U-tio was selectea. 1ne
baseline saturated fat consumption was calculated using a food questionnaire. and the cohort was followed
for seven years for the development of colon cancer. The study showed that women with high baseline
saturated fat consumption have four times the risk of colorectal cancer in a 7-year period. compared to
women with low fat consumption (RR = 4.0. 95% Cl = 1.5 - 6.5). According to the study results. what percent
of colorectal carcinoma in women with high fat consumption could be attributed to their diet?
r
r
.., r
r
r
A. 25% [34%]
B. 50% [8%]
c. 75% [42%]
D. 90% [8%]
E. 100% [8%]
liBl
~
20
1211
. LBJ
23
24
25
26
1271
~
Explanation:
User ld:
Attributable risk percent (ARP) or etiologic fraction is an important measure of the impact of a risk factor being
studied. ARP represents the excess risk in a population that can be explained by exposure to a particular risk
factor. It is calculated by subtracting the risk in the unexposed population (baseline risk) from the risk in the
exposed population. and dividing the result by the risk in the exposed population:
ARP =(risk in exposed- risk in unexposed)/risk in exposed.
An easier way to calculate the ARP is to derive it from the relative risk (RR):
29
30
31
32
33
34
35
36
37
38
39
40
ARP = (RR - 1)/RR.

In this case. ARP = (4.0- 1)/4.0 = 0.75 (75%). According to the study results. 75% of colorectal carcinoma in
the high consumption group was attributable to high saturated fat intake.
ARP represents the excess risk in the exposed population that can be attributed to the risk factor. It can be
easily derived from the relative risk using the following formula ARP = (RR - 1)/RR.
Notes
!:':~
Calculator
:CJ
5
6
7
8
:[lJ
11
12
13
14
16
17
liBl
~
20
1211
. LBJ
23
24
25
26
1271
~
29
30
31
32
33
34
35
36
37
38
39
40
Item: 15 of 44
11 PMark
Q.ld: ~86 [
-<J
C>-
Previous
Ne><t
jf
Lab Values
Over the last 5 years. more people in town A seem to be suffering from leukemia than in town B. Which of the
following study designs is the best for calculating the different incidences of leukemia in these two towns?
r
., r
r
r
r
A. Case-control [16%]
B. Cohort [46%]
C. Case series [5%]
D. Clinical trial [1 %]
E. Cross-sectional [30%]
Explanation:
User ld:
Incidence is defined as the frequency of new cases of a disease arising in a population at risk over a specified
time period. Thus. in order to determine incidence. subjects withoutthe disease must be followed over a
period oftime to discover how many subjects develop said disease. Among the choices listed. the best
option for determining incidence is a cohort study. A cohort study is a prospective observational study in
which groups are chosen based upon presence or absence of one or more risk factors. All subjects are then
observed over time for development of the disease of interest. allowing estimation of incidence within the total
population and comparison of incidences between subgroups. In this case. residents of town A and town B
could be followed for a period of time. and the incidence of leukemia determined for each site. Comparing the
incidences between the two towns will allow determination of the relative risk for developing leukemia in town
A versus town B.
(Choice A) In a case control study. subjects with the disease of interest are compared to an otherwise similar
group that is disease free. Information is then collected about exposure to risk factors. A case control study
is retrospective and meantto determine associations between risk factors and disease occurrence. An odds
ratio can be calculated in a case control study. but the incidence of a disease cannot.
(Choice C) A case series is a study involving only patients already diagnosed with a condition of interest. A
case series can be helpful in determining the natural history of uncommon conditions but provides no
information about disease incidence.
(Choice D) A clinical trial compares the therapeutic benefit of different interventions in patients already
diagnosed with a particular disease. Clinical trials cannot be used to determine disease incidence.
9''
Notes
1.~
Calculator
:CJ
5
6
7
8
:[lJ
11
12
13
14
16
17
liBl
~
20
1211
. LBJ
23
24
25
26
1271
~
Item: 15 of 44
Q.ld: ~86 [
!il f>Mark
<:J
C>-
Previous
Ne><t
Explanation:
jf
Lab Values
User ld:
Incidence is defined as the frequency of new cases of a disease arising in a population at risk over a specified
time period. Thus. in order to determine incidence. subjects without the disease must be followed over a
period of time to discover how many subjects develop said disease. Among the choices listed. the best
option for determining incidence is a cohort study. A cohort study is a prospective observational study in
which groups are chosen based upon presence or absence of one or more risk factors. All subjects are then
observed over time for development of the disease of interest. allowing estimation of incidence within the total
population and comparison of incidences between subgroups. In this case. residents of town A and town B
could be followed for a period of time. and the incidence of leukemia determined for each site. Comparing the
incidences between the two towns will allow determination of the relative risk for developing leukemia in town
A versus town B.
(Choice A) In a case control study. subjects with the disease of interest are compared to an otherwise similar
group that is disease free. Information is then collected about exposure to risk factors. A case control study
is retrospective and meant to determine associations between risk factors and disease occurrence. An odds
ratio can be calculated in a case control study. but the incidence of a disease cannot.
(Choice C) A case series is a study involving only patients already diagnosed with a condition of interest. A
case series can be helpful in determining the natural history of uncommon conditions but provides no
information about disease incidence.
(Choice D) A clinical trial compares the therapeutic benefit of different interventions in patients already
diagnosed with a particular disease. Clinical trials cannot be used to determine disease incidence.
29
30
31
32
(Choice E) A cross-sectional study takes a sample of individuals from a population at one point in time. It
allows determination of a diseases prevalence (the total number of cases in a population at a given
time). Since subjects are not studied longitudinally. disease incidence cannot not be determined.
33
34
35
36
37
38
39
40
A prospective cohort study design is best for determining the incidence of a disease. Comparing the
incidence of the disease in two populations (one with and one without a given risk factor) allows for calculation
of a relative risk.
Notes
1.":~
Calculator
:CJ
5
6
7
8
:[lJ
11
12
13
17
liBl
~
Item: 16 of 44
11 PMark
Q.ld: 3931 [
-<J
C>-
Previous
Ne><t
jf
Lab Values
A study was conducted to assess the association between L-tryptophan use and the development of
Eosinophilia-Myalgia Syndrome (EMS). Patients with EMS were asked aboutthe use of products containing
L-tryptophan during the last 6 months. Atthe same time. people without EMS were randomly selected from
the same population where the patients came from. and asked abouttheir experience with L-tryptophan
containing products within the last 6 months. The study showed that the use of L-tryptophan is significantly
associated with EMS. Which of the following measures of association are the investigators most likely to
report?
r
r
., r
r
r
A. Relative risk [27%]

B. Median survival [1 %]
C. Exposure odds ratio [63%]
D. Relative rate [3%]
E. Prevalence odds ratio [5%]
20
1211
. LBJ
23
24
25
26
1271
~
29
30
31
32
33
34
35
36
37
38
39
40
Explanation:
User ld:
The above case describes a typical case-control study design. Patients with the disease of interest (cases)
and people without the disease (controls) are asked about previous exposure to the variable being studied
(L-tryptophan use). The main measure of association is the exposure odds ratio. in which the exposure of
people with the disease (cases) is compared to the exposure of those without the disease (controls).
(Choices A and D) Incidence measures (e.g .. relative risk or relative rate) cannot be directly measured in
case-control studies because the people being studied are those who have already developed the disease.
Relative risk and relative rate are calculated in cohort studies. where people are followed over time for the
occurrence of the disease.
(Choice B) Median survival is calculated in cohort studies or clinical trials. and is usually used to compare the
median survival times in two or more groups of patients (e.g .. receiving a new treatment or placebo).
(Choice E) Prevalence odds ratio is calculated in cross-sectional studies to compare the prevalence of a
disease between different populations.
A case-control study is used to compare the exposure of people with the disease (cases) to the exposure of
""
Notes
1.":~
Calculator
:CJ
5
6
7
8
:[lJ
11
12
13
Item: 16 of 44
11 PMark
Q.ld: 3931 [
-<J
C>-
Previous
Ne><t
jf
Lab Values
L-rrypropnan aunng rne 1asr o monrns. AI rne same II me. people w1rnour t:IVI::> were ranaom1y se1ecrea rrom
the same population where the patients came from. and asked abouttheir experience with L-tryptophan
containing products within the last 6 months. The study showed that the use of L-tryptophan is significantly
associated with EMS. Which of the following measures of association are the investigators most likely to
report?
r
r
., r
r
r
A. Relative risk [27%]

B. Median survival [1 %]
C. Exposure odds ratio [63%]
D. Relative rate [3%]
E. Prevalence odds ratio [5%]
17
liBl
Explanation:
20
The above case describes a typical case-control study design. Patients with the disease of interest (cases)
and people without the disease (controls) are asked about previous exposure to the variable being studied
(L-tryptophan use). The main measure of association is the exposure odds ratio. in which the exposure of
people with the disease (cases) is compared to the exposure of those without the disease (controls).
1211
. LBJ
23
24
25
26
1271
~
29
30
31
32
33
34
35
36
37
38
39
40
User ld:
(Choices A and D) Incidence measures (e.g .. relative risk or relative rate) cannot be directly measured in
case-control studies because the people being studied are those who have already developed the disease.
Relative risk and relative rate are calculated in cohort studies. where people are followed over time for the
occurrence of the disease.
(Choice B) Median survival is calculated in cohort studies or clinical trials. and is usually used to compare the
median survival times in two or more groups of patients (e.g .. receiving a new treatment or placebo).
(Choice E) Prevalence odds ratio is calculated in cross-sectional studies to compare the prevalence of a
disease between different populations.
A case-control study is used to compare the exposure of people with the disease (cases) to the exposure of
the people without the disease (controls). The main measure of association is the exposure odds ratio.
""
Notes
1.":~
Calculator
:CJ
5
6
7
8
:[lJ
11
12
13
14
15
16
Item: 17 of 44
Q.ld : 3934 [
11 PMark
<J
C>-
Previous
Ne><t
jf
Lab Values
A group of investigators conducted a randomized placebo-controlled clinical trial to assess the effect of a new
aldosterone-receptor antagonist on the progression of chronic heart failure. The primary outcome in this
study was an all-cause mortality. They reported a decrease in all-cause mortality in the treatment group with
a relative risk of 0. 71 (p = 0.01). Which of the following statements is the best interpretation of the reported
association?
r
r
., r
r
r
A. There is only a 1% chance thatthe result is biased [4%]

B. The 95% confidence interval for the relative risk includes 1.0 [6%]
C. There is 1% probability that the result was observed by chance [75%]
D. There is a 71% decrease in all-cause mortality due to the drug [8%]
E. The results obtained are not statistically significant [7%]
Explanation:
20
1211
. LBJ
23
24
25
26
1271
~
29
30
31
32
33
User ld:
Know how to interpretthe 'p' value in conjunction with a measure of association (e.g .. relative risk). Relative
risk is a point estimate of association. but it does not account for random error. In other words. it is always
possible that the calculated relative risk occured by chance. The 'p' value is used to strengthen the results of
the study; it is defined as the probability of obtaining the result by chance alone. In this scenario. the
probability that the result was obtained by chance is 1% (i.e .. p=O .01). The commonly accepted upper limit
(cut-off point) of the 'p' value for the study to be considered statistically significant is 0.05 (i.e .. less then 5%).
(Choice A) The 'p' value deals with random variability. not bias. which is systematic error.
(Choice B) If the 'p' value is less than 0.05. the 95% confidence interval does not contain 1.0 (a 'null' value for
relative risk).
(Choice D) A relative risk of 0. 71 shows thatthe drug decreased the risk of mortality by 29% (again. a null
value for relative risk is 1.0).
34
35
36
37
38
39
40
The p value is the probability that the result of a study was obtained by chance alone. A study is considered
statistically significant when the 'p' value used is less than 0.05. Know how to interpretthe 'p' value and its
relationship with the confidence interval.
Time Soent: 2 seconds
Coovriaht USMLEWorld .LLC.
Last uodated f7/7/20 101
9-
Notes
1.~
Calculator
:CJ
5
6
7
8
:[lJ
11
12
13
14
Item: 18 of 44
!II PMark
Q.ld: 7686 [
26
1271
~
29
30
31
32
C>Ne><t
jf
Lab Values
The drug ad in the exhibit button applies to the next 2 items

The drug ad is focused on kalgatran (Kalaxin). a new oral anticoagulant for the treatment of non-valvular atrial
fibrillation.
Item 1 of2
Bleeding risk is the major concern when anticoagulating patients with non-valvular atrial fibrillation. The ad
promotes Kalaxin as a safe medication. In the study comparing Kalaxin and warfarin. which of the following
specific bleeding risks were most similar between the two groups?
15
24
25
<J
Previous
View Drug Ad
r A. Intracranial bleeding [1 0%]

r B. Gastrointestinal bleeding [8%]
r C. Life-threatening bleeding [23%]
., r D. Major bleeding [37%]
r E. Total bleeding [21%]
Explanation:
User ld:
This drug ad is comparing the effect of Kalaxin versus warfarin for preventing stroke in non-valvular atrial
fibrillation. It compares the hazard rates for adverse events for both drugs in the form of hazard ratios. A
hazard rate is the chance of an event occurring in either the treatment arm or control arm over a set period of
time. A hazard ratio is the chance of an event occurring in the treatment arm compared to the chance of that
event occurring in the control arm during a set period of time.
33
34
35
36
37
38
39
40
A hazard ratio < 1 indicates that an event is more likely to occur in the control arm. A hazard ratio > 1 signifies
that an event is more likely to occur in the treatment arm. A ratio close to 1 implies little difference between
the two groups. In this study. the hazard ratio for major bleeding was 0.96 (given in the bleeding statistics
under the chart). which is the closest to 1 compared to the other answer options. Additionally. the 95%
confidence interval (0 .84 - 1.1 D) contains the null value of 1. indicating that there is no significant difference in
the risk of major bleeding between the 2 groups.
Notes
1.":~
Calculator
:CJ
5
6
7
8
:[lJ
11
12
13
14
15
Item: 18 of 44
11 PMark
Q.ld: 7686 [
<:J
1>-
Previous
Ne><t
jf
Lab Values
r C. Life-threatening bleeding [23%]

.., r D. Major bleeding [37%]
r E. Total bleeding [21%]
Explanation:
User ld:
This drug ad is comparing the effect of Kalaxin versus warfarin for preventing stroke in non-valvular atrial
fibrillation. It compares the hazard rates for adverse events for both drugs in the form of hazard ratios. A
hazard rate is the chance of an event occurring in either the treatment arm or control arm over a set period of
time. A hazard ratio is the chance of an event occurring in the treatment arm compared to the chance of that
event occurring in the control arm during a set period of time.
A hazard ratio < 1 indicates that an event is more likely to occur in the control arm. A hazard ratio > 1 signifies
that an event is more likely to occur in the treatment arm. A ratio close to 1 implies little difference between
the two groups. In this study. the hazard ratio for major bleeding was 0.96 (given in the bleeding statistics
under the chart). which is the closestto 1 compared to the other answer options. Additionally. the 95%
confidence interval (0 .84 - 1.1 D) contains the null value of 1. indicating thatthere is no significant difference in
the risk of major bleeding between the 2 groups.
24
25
(Choice A) The hazard ratio for intracranial bleeding is 0.39. indicating that Kalaxin has a lower chance of
causing intracranial bleeding than warfarin.
26
1271
~
29
30
31
32
33
34
35
36
37
38
39
40
(Choice B) The hazard ratio for gastrointestinal (GI) bleeding is 1 75. indicating that Kalaxin has a higher
chance of causing gastrointestinal bleeding than warfarin. The hazard ratio for major Gl bleeding is 1.38.
(Choice C) The hazard ratio for life-threatening bleeding is 0.75. indicating that Kalaxin has a lower chance of
causing life-threatening bleeding than warfarin.

(Choice E) The hazard ratio for total bleeding is 0.91 . indicating that Kalaxin has a slightly lower chance of
causing overall bleeding than warfarin.

Hazard ratios are proportions that indicate the chance of an event occurring in the treatment arm as
compared to the chance of the event occurring in the control arm.
9-
Notes
1.":~
Calculator
:CJ
5
6
7
8
Item: 19 of 44
11 PMark
Q.ld: 7687 [
-<J
C>-
Previous
Ne><t
ltem2 of2
Proper randomization when assigning patients to treatment and control groups is important in order to avoid
the effect of extraneous variables on the study results. In the current study. which of the following forms of
additional information would help most in determining randomization success?
:[lJ
11
12
13
14
15
16
17
View Drug Ad
r
., r
r
r
r
A. Annual stroke rates [2%]

B. Baseline characteristics [67%]
C. Patient compliance chart [7%]
D. Patient follow-up rates [7%]
E. Subgroup analysis tables [17%]
20
1211
. LBJ
23
24
25
26
1271
~
29
30
31
jf
Lab Values
Explanation:
User ld:
In any randomized clinical study. the goal of successful randomization is to eliminate bias in treatment
assignments in order to minimize confounding variables. An ideal randomization process minimizes selection
bias. results in near equal patient group sizes. and achieves a low probability of confounding variables. This
allows a study to achieve adequate statistical power. A listing of the baseline characteristics of the patients in
each arm would demonstrate if the two arms had patients with similar characteristics and would ensure that
proper randomization occurred in the study.
(Choice A) Annual stroke rate is an endpoint of the study that may be influenced by patient randomization.
but it does not determine success of randomization.
32
(Choice C) Patient compliance charts indicate how well a patient tolerated the treatment and followed the
33
treatment regimen. but they would not directly indicate the success of randomization.
34
35
(Choice D) Patient follow-up rates indicate how many patients followed-up or dropped out of the study. This
36
data is not a direct marker of randomization success.
37
38
39
(Choice E) Subgroup analysis tables (stratified analysis) are used to determine if certain patient
40
characteristics (eg. age or comorbidities) influenced treatmentto any significant extent. Improper
randomization can result in inaccurate results durinq subqroup analvsis.
""
Notes
1.":~
Calculator
:CJ
5
6
7
8
:[lJ
11
12
13
14
15
16
17
Item: 19 of 44
11 PMark
Q.ld: 7687 [
<:1
(:>-
Previous
Ne><t
if
Lab Values
r D. Patient follow-up rates [7%]

r E. Subgroup analysis tables [17%]
Explanation:
User ld:
In any randomized clinical study. the goal of successful randomization is to eliminate bias in treatment
assignments in order to minimize confounding variables. An ideal randomization process minimizes selection
bias. results in near equal patient group sizes. and achieves a low probability of confounding variables. This
allows a study to achieve adequate statistical power. A listing of the baseline characteristics of the patients in
each arm would demonstrate if the two arms had patients with similar characteristics and would ensure that
proper randomization occurred in the study.
(Choice A) Annual stroke rate is an endpoint of the study that may be influenced by patient randomization.
but it does not determine success of randomization.
20
(Choice C) Patient compliance charts indicate how well a patient tolerated the treatment and followed the
1211
. LBJ
treatment regimen. but they would not directly indicate the success of randomization .
23
24
25
26
1271
~
29
30
31
32
33
34
35
36
37
38
39
40
(Choice D) Patient follow-up rates indicate how many patients followed-up or dropped out of the study. This
data is not a direct marker of randomization success.

(Choice E) Subgroup analysis tables (stratified analysis) are used to determine if certain patient
characteristics (eg. age or comorbidities) influenced treatment to any significant extent. Improper
randomization can result in inaccurate results during subgroup analysis.
Successful randomization in a clinical trial allows a study to achieve adequate statistical power. An ideal
randomization process minimizes selection bias. results in near equal patient group sizes. and achieves a
low probability of confounding variables.
References:
1. Statistical properties of randomization in clinical trials.
Last updated [4/25120 13]
"!J'
Notes
1.":~
Calculator
:CJ
5
6
7
8
Item: 20 of 44
<:1
f>-
Previous
Next
IJIII f>Mark
Q.ld: 2137 [
A gynecologic oncology research institute isolates a potential tumor marker for endometrial cancer. A large
multicenter study is then performed to evaluate serum levels of the tumor marker in women with and without
endometrial cancer. The following curves are generated using the results of the study.
:[lJ
0 Healthy
D Diseased
11
12
13
14
....<=
Ill
Ql
.::
15
..."'...a.
0
Ql
..c
E
:::l
24
25
Ant igen concentration
26
1271
($) USMLEWorld, LLC
~
29
30
31
32
Clinical researchers decide to use the tumor antigen to develop a confirmatory test for patients with
suspected endometrial cancer. During preliminary design of the test, the cutoff point for positive/negative
results is set at point A. If the cutoff point is moved from A to B, the specificity of the test will change in which
of the following ways?
33
34
35
36
37
38
39
40
r
r
.., r
r
A. Cannot be determined [0%]

B. Decreased [15%]
C. Increased [83%]
D. Unchanged [2%]
jf
Lab Values
Notes
l:'::.'ll
Calculator
:CJ
Item: 20 of 44
5
6
7
8
<:J
C>-
Previous
Ne><t
ill PMark
Q.ld: 2137 [
jf
Lab Values
Notes
1.":~
Calculator
Test cutoff
I - - Negative test ---il
:[lJ
1
l-- - Positive test - ---;
...."'c
11
41
;;;
12
13
14
..."'...c.
15
..c
41
E
::l
r
24
25
26
1271
~
29
30
31
32
33
34
35
36
37
38
39
40
U SMLEWo~d.LLC
Important parameters of diagnostic tests include the following:

True positive (TP)- A patient with a positive test result who actually has the disease.
False positive (FP) - A patient with a positive test result who actually does not have the disease.
True negative (TN)- A patient with a negative test result who actually does not have the disease.
False negative (FN) - A patient with a negative test result who actually has the disease.
Sensitivity- The number of patients correctly testing positive for a disease divided by the total number
of patients with the disease (TP I [TP + FN]). High sensitivity means that negative results are less
likely to be FNs and more likely to be TNs; thus, a negative result is better able to rule out the disease
(SnOut). Because tests with high sensitivity will correctly identify most of the patients with the
disease, sensitivity is very important for screening tests (which need to minimize FNs).
Specificity- The number of patients correctly testing negative for a disease divided by the total number
of patients without the disease (TN I [TN + FP]). High specificity means that positive results are less
likely to be FPs and more likely to be TPs; thus, a positive result is better able to rule in the disease
(Spin). Because tests with high specificity will correctly identify most of the patients without the
_,; _____ ----:.t::_: .... ;_. --. :----..&.-..
- r. ..... _ .. ____ .. _ _ .. _ / .... :_,_ ---- ... _ ... : .. : ... : __ .. P\_\
<~>-
:CJ
Item: 20 of 44
Q.ld: 2137 [
11
12
13
14
15
24
25
26
1271
~
29
30
31
32
33
34
35
36
37
38
39
40
-<:J
C>-
Previous
Ne><t
l/
5
6
7
8
:[lJ
11 PMark
Antigen concentration
jf
Lab Values
"-
U SMLEWo~d.LLC

likely to be FNs and more likely to be TNs; thus. a negative result is better able to rule out the disease
disease. sensitivity is very important for screening tests (which need to minimize FNs).
likely to be FPs and more likely to be TPs; thus. a positive result is better able to rule in the disease
disease. specificity is very important for confirmatory tests (which need to minimize FPs).
In this example. moving the cutoff point from A to B will cause more patients without the disease to test
negative (fewer FPs). increasing the specificity of the test. However. as a consequence. more patients with
the disease will also test negative (more FNs). resulting in decreased sensitivity. Conversely. moving the
cutoff point in the other direction (eg. B to A) will increase the number of FPs and decrease the number of
FNs. decreasing specificity while increasing sensitivity.
Changing the cutoff point of a quantitative diagnostic test will inversely affect its sensitivity and specificity.
Typically. raising the cutoff value will increase specificity (fewer false positives) and decrease sensitivity (more
false negatives). Screening tests need high sensitivity; confirmatory tests need high specificity.
Last updated [7122120 13]
Notes
1.~
Calculator
:CJ
5
6
7
8
:[lJ
11
12
13
14
15
16
17
liBl
~
20
Item: 21 of 44
Q.ld: 7711 [
11 f>Mark
-<:1
C>-
::::ll
G
Previous
Ne><t
Lab Values
A 60-year-old male with known coronary artery disease is seen

during routine check-up. He presents with no complaints. He
denies any abdominal pain. rectal bleeding. or weight loss. He
has no family history of colon cancer. His current medications
include metoprolol. atorvastatin. and low-dose aspirin. Based
on the study results which of the following is the best statement
concerning the use of iF OBT in this patient?
Objective:
To assess the association of low-dose aspirin use with the
performance of two quantitative immunochemical fecal occult blood
tests (iF OBTs) in a large sample of patients undergoing colorectal
cancer screening
Methods:
A. Accuracy of the test of test is not affected by use of
r aspirin [7%]
r B. Use of low-dose aspirin decreases the sensitivity of
Design: Cross-sectional study
the test [20%]
Blinding Technician performing iF OBT tests was blinded to
., r C. Use of low-dose aspirin increases the sensitivity of
colonoscopy results
the test [69%]
Setting: 20 gastroenterology practices in southern Germany
r D. Use of low-dose aspirin increases the specificity of

the test [5%]
Patients: Patients undergoing screening colonoscopy (mean age

23
24
25
26
1271
~
29
30
31
62.1 years) were asked to provide stool samples before the

procedure. Information about medication use was collected using a
standardized questionnaire. Specific exclusion criteria included the
following: visible rectal bleeding or previous positive F OBT result.
history of inflammatory bowel disease. colonoscopy in the past 5
years. incomplete colonoscopy. inadequate bowel preparation for
colonoscopy. participants with pseudopolyps or histologically
undefined polyps at screening colonoscopy. and participants who
reported regular use of high-dose analgesics.
32
Intervention: Two automated. enzyme-linked immunosorbent assay
33
(ELISA) based iF OBTs (hemoglobin test and

hemoglobin-haptoglobin test) were performed before screening
colonoscopy.
34
35
36
37
38
39
40
Explanation:
The study above is looking at the effect of aspirin on the ability

of two different assays to detect colorectal cancer. with
colonoscopy used as the gold standard to detect the
lesions. Sensitivity is a measure of the true positive rate and
indicates how well a test can detect those patients with a
disease. A higher sensitivity makes it less likely thatthere are
false negatives. and more likely that a negative test result
indicates that a patient does not have a given
condition. Specificity is a measure of the true negative rate and
indicates how well a test can rule out a given condition. The
higher the specificity. the less likely there are false positives
and the more likely that a positive test result confirms the
presence of a condition.
Outcome measures: Sensitivity. specificity. positive and negative

predictive values. and area under receiver operating characteristic
(ROC) curves in detecting advanced colorectal neoplasms with two
quantitative iF OBTs usinq colonoscopv as qold standard.
User ld:
A Receiver Operating Characteristic (ROC) curve plots the true

...:.]
nnc-iti\ 10 r-:.to ( c-onc-iti\ ,jt-, '' -:.n-:.inc-t tho f-:.lc-o nnc-iti\ o r-:.to ( 1
rs
.~
Notes
r.~
Calculator
:CJ
5
6
7
8
:[lJ
11
12
13
14
15
16
17
Item: 21 of 44
Q.ld : 7711 [
.
Prevtous
t>-
lab Values
Next
Results.
1979 patients (233 regular users of low-dose aspirin and 1746 who
never used low-dose aspirin) were studied. Advanced neoplasms
were found in 24 users ( 10.3%) and 181 nonusers ( 10 A%) of
low-dose aspirin.
Figure 3 . Receiver Operating Characteristic Curves

for Detecting Advanced Colorectal Neoplasms by
Quanti tative lmmunochemical Fecal Occult B lood
Test According to Use of Low-Dose Aspi rin
The figure above shows the ideal diagnostic test ( 1DO %

sensitive and specific). which provides the most useful
information. A diagnostic test that provides no useful
information usually produces the diagnosis by random chance
and demonstrates an inverse linear relationship between
sensitivity and specificity.
Specificity, %
80
60
40
20
80
c:]
.....................
...
#. 60
~
.~
////
(/)
cQ)
(/)
40
20
...............
...'
Hemoglobin test
23
24
25
26
!:"~
Calculator
20
100
..
Notes
positive rate (sensitivity) againstthe false positive rate ( 1 specificity. or inverse true negative rate). at different cutoff
points for a given diagnostic test. This curve usually shows
that an increase in sensitivity is offset by a decrease in
specificity.
All participants
liBl
28
29
30
31
32
33
34
35
36
37
38
39
40
<:l
ill P M ark
100
80
60
40
20
Specificity (%)
~ Accuracv is defined as the proportion of true results (true
:CJ
5
6
7
8
:[lJ
11
12
13
14
15
16
17
:I::I
Item: 21 of 44
I Mak
O.ld: 7711 (
<:::1
[::>
Previous
NeHt
F1gure 3 . Rece1ver Operating Chara ctenst1c Cu rves

for Detec ting Advance d Colore ctal Neoplasms by
Quantitative lmmun ochemical F eca l Occ ult Blood
Test Acco rding to Use of Low -Dos e Aspirin
100
SpecifiCity, %
80
80
40
60
20
1 ,.----~
....
..
..
26
1271
....
..
-~
....................
..
32
33
34
35
20
40
60
80
39
40
60
40
20
A shift in the ROC curve upwards for a given cutoff indicates
increased sensitivity. A shift of the curve to the right for a given
cutoff point indicates a decrease in specificity.
100
100-Specificity, %
36
37
38
In this study , aspirin use moves the ROC curve upwards. This
translates to an increase in sensitivity(Choice 8). However,
the curve also shifts to the right for a given cutoff point with
aspirin use . indicating a decrease in the specificity (Choice
D). The total area under the plotted curve increases with
aspirin use (despite the decrease in specificity, as the
magnitude increase in sensitivity is larger). leading to an
increase in overall accuracy (Choice A) .
...........
20
Accuracy is defined as the proportion of true results (true

positive and true negative) out of all the results that are
predicted by a test in a given population. as measured by a gold
standard or reference parameter. The closer the plotted curve
approaches the left and top borders of the ROC curve. the
more accurate the test is. Accuracy can also be measured as
the total area under the plotted curve. Precision is defined as
the proportion of true positives out of the total number of
positive results produced by a test in a given
population. Precision is equivalent to positive predictive
value . Both accuracy and precision depend upon the sensitivity
and specificity of the test. as well as the prevalence of the
condition in the population being tested .
Hemoglobin test
100
80
Specificity (%)
... .
29
~./
All partici pants
20
30
31
jf
Lab Values
Hemoglobin-haptoglobin test
Specificity, %
.:::J
Time Spent 40
seconds
Copyright
USMLEWor1d,LLC.
Last updated:
[31712013)
l~:'ll
Notes
Calculato1
:CJ
Item: 21 of 44
5
6
7
8
:[lJ
!il PMark
Q.ld: 7711 [
<:1
t>
if
Previous
Next
Lab Values
Hemoglobin-haptoglobin test
Specificity, %
80
100
60
100
20
40
80
...
...................
..
..
20
..
....
23
24
25
26
29
..
0
30
31
32
20
40
60
1oo - Specificity, %
80
100
Conclusion:
33
34
35
36
37
38
39
40
40
20
In this study, aspirin use moves the ROC curve upwards. This
translates to an increase in sensitivity(Choice B). However,
the curve also shifts to the right for a given cutoff point with
aspirin use, indicating a decrease in the specificity (Choice
D). The total area under the plotted curve increases with
aspirin use (despite the decrease in specificity, as the
magnitude increase in sensitivity is larger), leading to an
increase in overall accuracy (Choice A).
..
1271
~
..
...... , . . . - - - - - - - - - - - - ,
Cut points, 1-fg/9 stool
.........................
Users of low-dose aspirin
o Nonusers of low-dose aspirin
20
80
I
60
Accuracy is defined as the proportion of true results (true

positive and true negative) out of all the results that are
predicted by a test in a given population, as measured by a gold
standard or reference parameter. The closer the plotted curve
approaches the left and top borders of the ROC curve, the
more accurate the test is. Accuracy can also be measured as
the total area under the plotted curve. Precision is defined as
the proportion of true positives out of the total number of
positive results produced by a test in a given
population. Precision is equivalentto positive predictive
value. Both accuracy and precision depend upon the sensitivity
and specificity of the test, as well as the prevalence of the
condition in the population being tested.
..
liBl
Specificity (%)
11
12
13
14
15
16
17
~/
A shift in the ROC curve upwards for a given cutoff indicates
increased sensitivity. A shift of the curve to the right for a given
cutoff point indicates a decrease in specificity.
For two iF OBTs, low-dose aspirin use affects performance of the

test in detecting advanced colorectal neoplasm.
Funding Source: the German Research Foundation, the German

Federal Ministry of Education and Research. The test kits were
provided free of charge by the manufacturer.
~
Time Spent 40
seconds
Copyright
USMLEWorld ,LLC.
Last updated:
[3/7/2013]
Notes
Calculator
:CJ
5
6
7
8
:[lJ
11
12
13
14
15
24
25
26
1271
~
Item: 22 of 44
Q.ld: 7712 [
11 PMark
<J
C>-
Previous
Ne><t
To assess the association of low-dose aspirin use with the

performance of two quantitative immunochemical fecal occult blood
tests (iF OBTs) in a large sample of patients undergoing colorectal
cancer screening
Methods:
Design: Cross-sectional study
Blinding Technician performing iF OBT tests was blinded to
32
33
34
35
36
37
38
39
40
The cutoff point for iF OBT recommended by manufacture for

hemoglobin test is 1~gig. How would the performance of the
test be affected by setting the cutoff of the test at 4~g/g for
patients taking low-dose aspirin?
r
r
., r
r
A. Negative predictive value would increase [18%]

B. Number of false-positive would increase [17%]
C. Positive predictive value would increase [56%]
D. Sensitivity of the test would increase [1 0%]
colonoscopy results
Setting: 20 gastroenterology practices in southern Germany
Explanation:
Patients: Patients undergoing screening colonoscopy (mean age
Both sensitivity and specificity depend on the cutoff value of a

given test. In this example, raising the cutoff value makes it
more difficult to diagnose the condition (colorectal neoplasm),
as more hemoglobin must be present in the stool for the test to
be positive. By raising the cutoff value, it makes it harder to
obtain a positive test result and easier to obtain a negative
result. This will increase the specificity but decrease the
sensitivity. Lowering the cutoff value would make it easier to
obtain a positive result and harder to obtain a negative result,
and doing so will increase the sensitivity but decrease the
specificity.
62.1 years) were asked to provide stool samples before the

procedure. Information about medication use was collected using a
standardized questionnaire. Specific exclusion criteria included the
following: visible rectal bleeding or previous positive F OBT result,
history of inflammatory bowel disease, colonoscopy in the past 5
years, incomplete colonoscopy, inadequate bowel preparation for
colonoscopy, participants with pseudopolyps or histologically
undefined polyps at screening colonoscopy, and participants who
reported regular use of high-dose analgesics.
29
30
31
jf
Lab Values
Intervention: Two automated, enzyme-linked immunosorbent assay

(ELISA) based iF OBTs (hemoglobin test and
hemoglobin-haptoglobin test) were performed before screening
colonoscopy.
Outcome measures: Sensitivity, specificity, positive and negative

predictive values, and area under receiver operating characteristic
(ROC) curves in detecting advanced colorectal neoplasms with two
quantitative iF OBTs using colonoscopy as gold standard.
Results:
User ld:
The positive predictive value (PPV) and negative predictive

values (NPV) are related to sensitivity and specificity, but are
defined differently. PPV can be thought of as the clinical value
of a positive test result in a given population. NPV can likewise
be considered as the clinical value of a negative test result in a
given population. The higher the PPV or NPV, the greater the
likelihood that a given positive or negative test result,
respectively, is actually a correct diagnosis.
Notes
1.~
Calculator
:CJ
5
6
7
8
:[lJ
Item: 22 of 44
!II PMark
Q.ld: 7712 [
<J
t>
Previous
Next
Results.
low-dose aspirin.
11
12
13
14
15
Figure 3. Receiver Operating Characteristic Cu rves

Test According to Use of Low-Dose Aspi rin
24
25
26
60
40
20
80
c:]
....................
.
#. 60
..
.~
,///
(/)
~ 40
20
'
..
....
..
..
True positive
(TP)
False positive
(FP)
Negative Test
result
False negative
(FN)
True negative
(TN)
=TP I (TP+FP)
NPV =TNI(TN+FN)
PPV
The Receiver Operating Characteristic (ROC) curve in the

studies above show that an increase in the cutoff value from 1
~gig to 4 ~gig decreased the sensitivity of the hemoglobin test
from 70% to 60% and increased the specificity from 80% to
90%. Similarly, the hemoglobin-haptoglobin test also had a
decrease in sensitivity from 60% to 40% and an increase in
specificity from 80% to 90% with the same increase in cutoff
values. This increase in specificity correlates with an increase
in the PPV.
""'
c
Positive Test
result
The equations for PPV and NPV are defined above, along with
the ones for sensitivity and specificity. As the sensitivity
increases, the NPV increases (due to fewer FN). As the
specificity increases, the PPV increases (due to lower FP). It
is important to note that while sensitivity and specificity depend
only upon the characteristics of a given test, PPV and NPV also
depend upon the prevalence of the condition in the population
being tested. PPV varies directly with prevalence (higher
prevalence correlates with higher PPV), while NPV varies
inversely with prevalence (higher prevalence corresponds with
lower NPV).
Specificity, %
80
Negative
condition
=TP I (TP+FN)
Specificity =TN I (TN+FP)
Hemoglobin test
100
Positive
condition
Sensitivity
All parti cipants
28
29
30
31
32
33
34
35
36
37
38
39
40
jf
Lab Values
..
(Choice A) NPV would increase following an increase in
Notes
!:':~
Calculator
:CJ
5
6
7
8
:[lJ
Item: 22 of 44
!II PMark
Q.ld: 7712 [
<:1
t>
jf
Previous
Next
Lab Values
y p
Results.
low-dose aspirin.
Figure 3 . Receiver Operating Characteristic Cu rves

Test According to Use of Low-Dose Aspirin
All participants
Hemoglobin test
Specificity, %
80
100
24
25
26
40
20
(Choice A) NPV would increase following an increase in

sensitivity. This example demonstrated a decrease in
sensitivity, which would lead to a corresponding decrease in
the NPV.
(Choices B and D) The number of false positives would

decrease if the cutoff rate were increased, since it would be
harder to obtain a positive test result. This would also lead to
decreased sensitivity since the number of false negatives
increases .
..
80
c:]
28
29
30
31
32
33
34
35
36
37
38
39
40
60
........
.
#. 60
..
~
.~
""
c
rJ3
Changing the cutoff value of a test for a given condition alters
the PPV, NPV, sensitivity, and specificity. Increased specificity
correlates with increased PPV and increased sensitivity
corresponds to increased NPV. PPV and NPV also depend
upon the condition's prevalence .
,///
(/)
40
20
..
'
....
..
..
'
The Receiver Operating Characteristic (ROC) curve in the

studies above show that an increase in the cutoff value from 1
~gig to 4 ~gig decreased the sensitivity of the hemoglobin test
from 70% to 60% and increased the specificity from 80% to
90%. Similarly, the hemoglobin-haptoglobin test also had a
decrease in sensitivity from 60% to 40% and an increase in
specificity from 80% to 90% with the same increase in cutoff
values. This increase in specificity correlates with an increase
in the PPV.
11
12
13
14
15
depend upon the prevalence of the condition in the population

being tested. PPV varies directly with prevalence (higher
prevalence correlates with higher PPV), while NPV varies
inversely with prevalence (higher prevalence corresponds with
lower NPV).
'
Time Spent 3
seconds
Copyright
USMLEWorld ,LLC.
Last updated:
[2111120 13]
Notes
1.":~
Calculator
:CJ
5
6
7
8
:[lJ
11
12
13
14
15
16
17
liBl
~
Item: 23 of 44
-<:J
t>
Previous
Ne><t
11 PMark
Q.ld: <W01 [
jf
Lab Values
A clinical study is conducted to assess the role of a vasopressin antagonist on overall survival in patients with
advanced heart failure and hyponatremia. After providing informed consent. eligible hospitalized patients are
administered either the new drug or a matching placebo along with the appropriate standard care.
Assignment to the 2 treatment arms is done randomly using numbers generated by a computer. This
assignment strategy is most helpful for controlling which of the following?
r
.., r
r
r
r
A. Ascertainment bias [30%]

B. Confounding [36%]
C. Effect modification [11 %]
D. Recall bias [2%]
E. Selective survival [22%]
Explanation:
User ld:
20
Methods to control confounding

24
25
Design stage
Matching
26
1271
Restriction
Rand om ization
29
30
31
Analysis stage
St ratified analysis
32
Statistical modelin g
33
34
1 USMLEWOfld, UC
35
36
37
38
39
40
The effects of confounding bias can be reduced through good study design and proper analysis of the results.
Methods used to control confounding during data analysis include stratified analysis and statistical modeling
(eg. multivariate analyses). Methods to control confounders during study design include the following:
1
t.ll ......... t.-.: .........
r .......................... ,,...,......J
; ....................................,..., ....... ,...J;,...,...
..~.-.; .... __... ......t.-. .... ...J ; ...., ,,...,, ,,...,... __... .......... t.-.: ........ 1, ......... ,. ................ ....................... ...J
9-
Notes
1.~
Calculator
:CJ
5
6
7
8
:[lJ
11
12
13
14
15
16
17
liBl
~
20
24
25
26
1271
~
29
30
31
32
33
34
35
Item: 23 of 44
!il f>Mark
Q.ld: <W01 [
<:J
C>-
Previous
Ne><t
The effects of confounding bias can be reduced through good study design and proper analysis of the results.
Methods used to control confounding during data analysis include stratified analysis and statistical modeling
(eg. multivariate analyses). Methods to control confounders during study design include the following:
1. Matching: Frequently used in case-control studies. this method involves matching known or suspected
confounding variables between the cases and controls. For example. if smoking status is a known
confounder. then controls will be selected on a case by case basis so that there are similar numbers
of smokers and nonsmokers in both the case and control groups.
2. Restriction: This method involves limiting study participation to individuals with specific
characteristics. For example. if sex is thoughtto be a confounder. then a restricted study may enroll
only men. However. this method is limited because the study results may not be applicable to the
excluded population (ie. women).
3. Randomization: Commonly employed in clinical trials. this technique helps to balance the distribution
of confounding variables between treatment and placebo groups so that the unconfounded effect of
the exposure of interest can be isolated. An important advantage of randomization. compared to other
methods. is the possibility of controlling known risk factors (eg. age & severity of heart failure) as well
as the unknown and difficult-to-measure confounders (eg. level of stress & socioeconomic status).
Randomization also eliminates partiality in treatment assignments (minimizing selection bias) and blinds
investigators to the identity of patients who receive the treatment arm (reducing observer bias).
(Choice A) Ascertainment (sampling) bias occurs due to nonrandom sampling of a target population. This
causes the characteristics of the study population to differ from those of the target population. leading to
results that may not be generalizable.
(Choice C) Effect modification results when an external variable positively or negatively impacts the effect of
a risk factor on the disease of interest. For instance. the risk of venous thrombosis is increased with estrogen
therapy. and this effect is augmented by smoking.
(Choice D) Recall bias results from inaccurate recall of past exposure and applies primarily to retrospective
studies. People who have suffered an adverse event are more likely to recall risk factors than people without
such experiences.
36
37
38
39
(Choice E) Selective survival bias occurs in case-control studies when cases are selected from the entire
disease population instead of just those that are newly diagnosed. For instance. a study on cancer survival
that is not limited to newly diagnosed patients will contain a higher proportion of relatively benign malignancies
40
:::.~ thP~P n:::.tiPnt~ nPnPr:::.llv livP lnnnPr
jf
Lab Values
Notes
1.":~
Calculator
:CJ
5
6
7
8
:[lJ
11
12
13
14
15
16
17
liBl
~
20
24
25
26
1271
~
29
30
31
32
33
34
35
36
37
38
39
40
Item: 23 of 44
Q.ld: <W01 [
!il f>Mark
<:J
C>-
Previous
Ne><t
jf
Lab Values
Notes
1.":~
Calculator
contounaer. men controls will be selectea on a case by case basis so mat mere are s1m11ar numbers
of smokers and nonsmokers in both the case and control groups.
2. Restriction: This method involves limiting study participation to individuals with specific
characteristics. For example. if sex is thought to be a confounder. then a restricted study may enroll
only men. However. this method is limited because the study results may not be applicable to the
excluded population (ie. women).
3. Randomization: Commonly employed in clinical trials. this technique helps to balance the distribution
of confounding variables between treatment and placebo groups so that the unconfounded effect of
the exposure of interest can be isolated. An important advantage of randomization. compared to other
methods. is the possibility of controlling known risk factors (eg. age & severity of heart failure) as well
as the unknown and difficult-to-measure confounders (eg. level of stress & socioeconomic status).
Randomization also eliminates partiality in treatment assignments (minimizing selection bias) and blinds
investigators to the identity of patients who receive the treatment arm (reducing observer bias).
(Choice A) Ascertainment (sampling) bias occurs due to nonrandom sampling of a target population. This
causes the characteristics of the study population to differ from those of the target population. leading to
results that may not be generalizable.
(Choice C) Effect modification results when an external variable positively or negatively impacts the effect of
a risk factor on the disease of interest. For instance. the risk of venous thrombosis is increased with estrogen
therapy. and this effect is augmented by smoking.
(Choice D) Recall bias results from inaccurate recall of past exposure and applies primarily to retrospective
studies. People who have suffered an adverse event are more likely to recall risk factors than people without
such experiences.
as these patients generally live longer.
Randomization is used to control confounders during the design stage of a study. It helps to control known
confounders as well as unknown and difficult-to-measure confounders.
:CJ
5
6
7
8
:[lJ
11
12
13
14
15
16
17
liBl
~
20
1211
. LBJ
Item: 24 of 44
Q.ld: 4189 [
!II f> Mark
-<J
!>-
Previous
Ne><t
jf
Lab Values
A 40-year-old woman presents to the office with several months history of atypical chest pain. Her family
history is positive for hypertension and type II diabetes mellitus. Her past medical history is insignificant. An
ECG stress test is performed to evaluate the possibility of coronary heart disease. This test has 85%
sensitivity and 80% specificity. In the low-risk group. its positive predictive value is 25%. and negative
predictive value is 96%. If the test is negative. what is the probability that the patient has coronary heart
disease?
r
r
r
.., r
r
A. 25% [4%]
B. 20% [8%]
c. 15% [10%]
D. 4% [75%]
E. 1% [2%]
Explanation:
User ld:
Pre- and post-test probabilities. as well as positive and negative predictive values. are very important
concepts in clinical medicine. as they dramatically change the approach to the diagnosis of each disease .
Take the time to learn and fully understand these terms.
23
26
1271
~
29
30
31
The pre-test probability sets your expectations before performing the diagnostic test. It describes the existing
probability of a patient to have the disease in question even before using a particular diagnostic test. (Although
there is a mathematical approach to calculate the pre-test probability. we shall simplify this term and use the
most basic explanations.) Using the example given above. the pre-test probability of the patient having
coronary artery disease can be assumed to be low since she has no risk factors aside from her significant
family history. and since she has atypical chest pain. Statistically. the prevalence is directly related to the
pre-test probability. In this example. the patient belongs to the low-risk group. which consistently has a
relatively low prevalence of coronary artery disease.
32
33
34
35
36
37
38
39
40
The positive predictive value (PPV) describes the probability of having the disease if the test result is positive.
The post-test probability of having the disease is directly related to the PPV. Using the example above. the
PPV (25%) is low. Consequently. if the test result is positive. then the post-test probability of having the
disease is low. The post-test probability is also dependent on the sensitivity. specificity. and pre-test
probability of having the disease.
The negative predictive value (NPV) describes the probability of not having the disease if the test result is
neqative. In this example. the NPV is 96%. This means that if the test result is neqative. the chances of the
Notes
1.~~
Calculator
:CJ
5
6
7
8
:[lJ
11
12
13
14
15
16
17
liBl
~
20
1211
. LBJ
23
26
1271
~
29
30
31
32
33
34
35
36
37
38
39
40
Item: 24 of 44
!il f> Mark
Q.ld: 4189 [
<:J
C>-
Previous
Ne><t
jf
Lab Values
Pre- and post-test probabilities. as well as positive and negative predictive values. are very important
concepts in clinical medicine. as they dramatically change the approach to the diagnosis of each disease.
Take the time to learn and fully understand these terms.
The pre-test probability sets your expectations before performing the diagnostic test. It describes the existing
probability of a patient to have the disease in question even before using a particular diagnostic test. (Although
there is a mathematical approach to calculate the pre-test probability. we shall simplify this term and use the
most basic explanations.) Using the example given above. the pre-test probability of the patient having
coronary artery disease can be assumed to be low since she has no risk factors aside from her significant
family history. and since she has atypical chest pain. Statistically. the prevalence is directly related to the
pre-test probability. In this example. the patient belongs to the low-risk group. which consistently has a
relatively low prevalence of coronary artery disease.
The positive predictive value (PPV) describes the probability of having the disease if the test result is positive.
The post-test probability of having the disease is directly related to the PPV. Using the example above. the
PPV (25%) is low. Consequently. if the test result is positive. then the post-test probability of having the
disease is low. The post-test probability is also dependent on the sensitivity. specificity. and pre-test
probability of having the disease .
The negative predictive value (NPV) describes the probability of not having the disease if the test result is
negative. In this example. the NPV is 96%. This means that if the test result is negative. the chances of the
patient to not have the disease is high (96%). Conversely. if the test result is negative. the chances of the
patientto have the disease is low ( 100% - 96%= 4%).
Other cases and diagnostic tests which are high-yield for the USMLE are listed below:
1. Pulmonary embolism and perfusion-ventilation scanning
2. Prostate cancer and serum PSA levels
(Choice B) False positive ratio= 1- specificity.

(Choice C) False negative ratio= 1 -sensitivity.
If a test result is negative. the probability of having the disease is 1 - negative predictive value.
Notes
1.":~
Calculator
:CJ
5
6
7
8
:[lJ
Item: 25 of 44
1111 PMark
Q.ld: 3653 [
<J
C>-
Previous
Ne><t
jf
Lab Values
A study was conducted to evaluate the efficacy of a new antiviral drug for the treatment of the common cold in
young children. The study population consisted of 1DO children between the age of 2 to 8 years. These
children were diagnosed with rhinovirus infection and subsequently given the particular antiviral drug. One
week later. it was observed that 92 of the 1DO patients were asymptomatic. Which of the following is the true
conclusion of this study?
11
12
13
14
15
16
17
liBl
~
r A. The drug is highly effective as the effectiveness is 90% [2%]

r B. The drug is moderately effective as the efficacy is 90% [1 %]
., r C. An exact conclusion cannot be drawn from the study. as common cold is often a self-limiting
disease of less than 1 week [93%]
r D. The drug is not effective as the sample size is very small [2%]
r E. No conclusion can be made. as compliance is generally very low in small children [1 %]
20
1211
. LBJ
23
24
26
1271
~
29
30
31
32
33
34
Explanation:
User ld:
It is very importantto consider the natural history of a disease when evaluating any drug trial. The common
cold is a self-limiting disease. which generally resolves within a week. For this reason. conclusions are
difficult to draw regarding the effectiveness of the drug. The resolution of symptoms in the 90% of the study
population may be due to natural resolution rather than the drug. For such drug trials. a control group is useful
since it helps exclude the bias due to natural resolution of the disease.
(Choices A, 8 and D) It is difficult to comment on the effectiveness of the drug. unless a comparison is made
with the control group and statistical significance is calculated to know the power of the study.
(Choice E) The reason why no conclusion can be made is the absence of a control group in the study design.
rather than improper patient compliance. Nevertheless. it is importantto consider the level of compliance in a
drug study done on small children.
35
36
37
38
39
40
Consider the natural history of a disease when evaluating the effectiveness of a drug in a trial.
9-
Notes
1.~
Calculator
:CJ
5
6
7
8
:[lJ
11
12
13
14
15
16
17
Item: 26 of 44
Q.ld: 3989 [
1111 PMark
<J
C>-
Previous
Ne><t
jf
Lab Values
A study was conducted to assess the relationship between serum HDL2 level (a subtraction of HDL) and
carotid intima-media thickness as a marker of atherosclerosis. Interpretation of the results revealed a linear
relationship between these two variables and the correlation coefficient (r -0.35. p 0.005). Which of the
following statements about the results of the study is the most correct?
r
r
., r
r
r
A. The association is not statistically significant [5%]

B. There is a strong correlation between HDL2 level and carotid intima-media thickness [17%]
C. As the level of HDL2 increases the carotid intima-media thickness decreases [65%]
D. There is positive correlation between HDL2 level and carotid intima-media thickness [9%]
E. Decreased HDL2 level is the cause of carotid intima-media thickening [4%]
liBl
Explanation:
20
The correlation coefficient assesses a linear relationship between two variables. The null value for the
correlation coefficient is 0 (no association). and the range of plausible values is from -1 to 1. The sign of the
correlation coefficient indicates a positive or negative association. The closer the value is to its margins (-1 or
1). the stronger the association.
1211
. LBJ
23
User ld:
(Choice A) The association is actually statistically significant because the p value is low.
(Choice B) The strength of association is weak because 0.35 is close to 0.
(Choice D) In this scenario. 'r' is negative. This means that as the level of HDL2 increases. the carotid
intima-media thickness decreases.
30
31
32
(Choice E) The correlation coefficient shows the strength of association. but does NOT necessarily imply
causality.
33
34
35
36
37
38
39
40
The correlation coefficient shows the strength and direction (positive. negative) of linear association between
two variables. It does not necessarily imply causality.
9-
Notes
1.~
Calculator
:CJ
5
6
7
8
:[lJ
11
12
13
14
15
16
17
liBl
Item: 27 of 44
11 f>Mark
Q.ld: 7688 [
-<:1
C>-
Previous
Ne><t
Risk of Hyperkalemia in Non-diabetic Patients with Chronic

Kidney Disease Receiving Antihypertensive Therapy
A 43-year-old African American male with a strong family

history of hypertension presents to the office with generalized
fatigue. His past medical history is noncontributory. He is an
active smoker. He is currently taking metoprolol. His blood
pressure is 162/1 DO mmHg and his pulse is 76/min. The apical
impulse is laterally displaced. His physical examination is
otherwise unremarkable. EKG shows high voltage consistent
with left ventricular hypertrophy with secondary repolarization
changes. His serum potassium is 3.9 mEq/L and estimated
GFR is 36 mUmin/1.73 m2 . His baseline UP/Cr is 0.20. The
patient is started on an ACE inhibitor and he shows compliance
with follow-up visits. Based on study results. which of the
following factors on follow-up will most significantly increase his
chance of hyperkalemia event?
Objective:
Explore the incidence and factors associated with hyperkalemia in

patients with chronic kidney disease (CKD) treated with
antihypertensive drugs
Methods:

Blinding Double-blinded
Follow-up: 3 to 6.4 years
1211
. LBJ
Setting Multicenter (21 medical centers)
jf
Lab Values
20
23
24
25
26
29
30
31
32
33
34
35
36
37
38
39
40
r
r
r
., r
Patients: African American patients. aged 18 to 70 years. with

hypertensive CKD as defined by a diastolic blood pressure (BP)
higher than 95 mm Hg and a glomerular filtration rate (GFR)
between 20 and 65 mUminl1.73 m2 . Specific exclusion criteria
included diabetes mellitus; urinary protein to urinary creatinine ratio
(UP/Cr) higher than 2.5; accelerated or malignant hypertension;
secondary hypertension; serious systemic disease; congestive
heart failure; and initial potassium level higher than 5.5 mEq/L.
D. Follow-up serum potassium level of 4.7 meq/1 as

compared to baseline [70%]
r E. Fallow-up UP/Cr of 1.1 as compared to baseline

[10%]
Explanation:
Intervention: Patients were randomized to initial treatment with

either a beta-blocker (metoprolol). an ACE inhibitor (ramipril). or a
calcium channel blocker (amlodipine) and to 1 of 2 mean arterial BP
goals ( 102-1 07 mmHg or <92 mmHg).
Outcome measures: hyperkalemic event (potassium level higher
than 5.5 mEq/L).
Results:
A. Addition of a calcium channel blocker [7%]

B. Addition of a thiazide diuretic [8%]
C. Discontinuation of metoprolol [5%]
'
:..:..J
User ld:
This study is a double-blinded randomized clinical trial

comparing the incidence of hyperkalemia in African-American
(non-diabetic) patients with CKD and hypertension (diastolic
greater than 95 mm Hg). The patients were randomized to
either metoprolol. amlodipine. or ramipril and followed up to 6
years. The primary endpoint was the incidence of
hyperkalemia. defined as greater than 5.5 mEq/L. The results
\AtPrP rPnnrtPrl :=I~ P\/Pnt~ nPr 1nn n:::.tiPnt \/P:::Ir~ :::.nrl :::.l~n
Notes
1.~
Calculator
:CJ
5
6
7
8
:[lJ
11
12
13
14
15
16
17
liBl
~
Item: 27 of 44
A total of 1094 non-diabetic patients were randomized. A total of

6497 potassium measurements were obtained. and 80
hyperkalemic events in 51 subjects were identified.
Hazard ratio is the ratio of an event rate occurring in the

treatment group compared to an event rate occurring in the nontreatment group. A ratio less than 1 indicates that the
treatment group had a significantly lower event rate while
values greater than 1 indicate that the treatment group had a
much higher event rate. In this study. the risk factors for
developing hyperkalemia were reported as hazard ratios. The
higher the ratio. the more likely that factor is associated with the
incidence of hyperkalemia. The factor that had the highest ratio
(of the choices listed) is a follow-up serum potassium level
between 4-5 mEq/L as compared to less than 4 mEq/L. with a
hazard ratio of 7.25.
~ 15
~-
RaMomlzed Oruo Groop

0 Rarnipril
0 Metroprol succinata
AmiOPidine besylale
12
"'&!
s;!
"
.!l
1211
. LBJ
(Choice A) As shown in the results figure. patients with a GFR

between 30-40 mUmin/1 .73 m2 given amlodipine actually had a
lower incidence of hyperkalemia per 1DO patient years than
ramipril.
!!!
0
Baseline GFR
(Choice B) The hazard ratio for follow-up use of a diuretic was

0 A 1 (with a statistically significant p value <0 .01). indicating
that adding a diuretic is associated with a lower incidence of
hyperkalemia.
figure 1. Hyperkalemia event rate per 100 patient-years by randomized drug

groups and baseline glomerular filtration rate (GFR). Error bars indicated
95% confidence intervals.
29
30
31
32
33
36
37
38
39
40
(Choice C) This patient is already on metoprolol and has a

normal potassium. The addition of metoprolol in the study did
not significantly increase the incidence of hyperkalemia in
patients with GFR > 30 and ~40 according to the figure above.
so discontinuation of the drug likely would not have a significant
effect on the potassium.
Table 4. Association of Risk of Hyperkalemia With

Time-Dependent Factors In Mulllvariable Analysls 3
34
35
jf
Lab Values
reported as hazard ratios.
a:
26
t:>
Ne><t
Results:
20
23
24
25
<:J
Previous
!il PMark
Q.ld: 7688 [
Variable
Follow-up diuretic use
follow-up GFR s 30 vs
> 50 mUmin/1.73 m2
Follow-uo GFR > 30 to_;;s40 vs
Hazard Ratio
(95% Conlidence
Interval)
p
Value
0.41 (0.22-0.78)
9.07 (3.1 8-25.88)
.006
<.001
3.6711 .21-11.15\
.02_
(Choice E) Fallow-up UP/Cr of 1.1 as compared to baseline

has a hazard ratio of 1.84. indicating that there is an increased
incidence of hyperkalemia. However. this ratio is lower than the
hazard ratio for follow-up potassium level between 4-5 mEq/L
;:~nrl it i~ nnt ~t;:~ti~tir.;:~llv ~innifir.;:~nt (n=n 1R\
Notes
Calculator
:CJ
5
6
7
8
:[lJ
11
12
13
14
15
16
17
liBl
~
20
1211
. LBJ
23
24
25
26
Item: 27 of 44
Q.ld : 7688 [
-<:1
1>-
Prevrous
Ne><t
11 f> Mark
Time-Dependent Factors In Multlvariable Analysls 3
Variable
Follow-up GFR :!i30 vs
> 50 mUmin/1.73 m2
Follow-up GFR > 30 to :s40 vs
> 50 mUmln/1.73 mz
Follow-up GFR > 40 to :s50 vs
> 50 mUmin/1.73 mz
Follow-up UP/Cr > 0.08 to :s0.22
vs !50.08
Follow-up UP/Cr >0.22 to :s0.66
vs !50.08
Followup UP/Cr > 0.66 vs :s0.08
followup potassium level4-5
vs < 4 mEq/L
Follow-up potassium level > 5
vs < 4 mEq/L
Hazard Ratio
(95% Confidence
Interval)
p
Value
0.41 (0.22-0.78)
9.07 (3.1 8-25.88)
.006
<.001
3.67 (1 .21-1 1.15)
.02
1.98 (0.59-6.61)
.27
2.01 (0.92-4.39)
.08
1.50 (0.62-3.63)
.37
1.84 (0.78-4.30)
7.25 (1.7230.58)
.16
.007
30.83 (6.89-138.0)
< .001
il
Lab Values
(Choice B) The hazard ratio for follow-up use of a diuretic was

0 A 1 (with a statistically significant p value <0 .01). indicating
that adding a diuretic is associated with a lower incidence of
hyperkalemia.
(Choice C) This patient is already on metoprolol and has a
normal potassium. The addition of metoprolol in the study did
not significantly increase the incidence of hyperkalemia in
patients with GFR >30 and ~40 according to the figure above.
so discontinuation of the drug likely would not have a significant
effect on the potassium.
(Choice E) Fallow-up UP/Cr of 1.1 as compared to baseline
has a hazard ratio of 1.84. indicating that there is an increased
incidence of hyperkalemia. However. this ratio is lower than the
hazard ratio for follow-up potassium level between 4-5 mEq/L
and it is not statistically significant (p=O .16).
Hazard ratios are the ratio of an event rate occurring in the
treatment arm versus the non-treatment arm. Ratios less than
1 indicate that the treatment arm had a lower event rate while
ratios higher than one indicate the treatment arm had a higher
rate of events.
Conclusion:
29
30
31
In non-diabetic patients with hypertensive CKD treated with ACEis.

the risk of hyperkalemia is small.
References:
1. Biostatistics primer: what a clinician ought to
know: hazard ratios.
32
33
Funding Source: The National Institute of Diabetes and Digestive
34
and Kidney Diseases (NIDDK) grant; additional financial support

from the Office of Research in Minority Health and drug donations
from Pfizer Inc. AstraZeneca Pharmaceuticals. and King
Pharmaceuticals.
35
36
37
38
39
40
Structured abstract is based on: Arch Intern Med.
2009; 169( 17) 1587-94
2. Estimation of the 2 -sample hazard ratio

function using a semi parametric model.
Time Spent 18
seconds
Copyright
USMLEWorld .LLC.
Last updated:
[7/21/2013]
I'Ol:J""
Notes
It~
Calculator
:CJ
5
6
7
8
:[lJ
11
12
13
14
15
16
17
:I::I
:I ~~I
20
23
29
30
31
32
33
34
35
36
37
38
39
40
Item: 28 of 44
I ' Mak
O.ld: 7689 (
<:::]
[:>
Prevous
jf
NeHt
Lab Values
For the study, patients were randomized to initial treatment with

either a beta-blocker (metoprolol), an ACE inhibitor (ramipril), or
a calcium channel blocker (amlodipine) and to 1 of 2 mean
arterial BP goals ( 102-1 07 mmHg or <92 mmHg). The study is
best classified as:
Risk of Hyperkalemia In Non-diabetic Patients with Chronic

Kidney Disease Receiving Antihypertensive Therapy
Objective:
Explore the incidence and factors associated with hyperkalemia in
patients with chronic kidney disease (CKD) treated with
antihypertensive drugs
>~
Methods:
r
r
r
r
A. Cluster (12%)
B . Cross-over (12%)
C . Factorial des1gn [28%)
D . Parallel-group [48%)

User ld:
Blinding: Double-blinded
Explanation :
Follow-up: 3 to 6.4 years
Factorial Design Study Exhibit
Selling: Multicenter (21 medical centers)
This study utili zes 3 different interventions (metoprolol, ramipril.

amlodipine) with 2 different variable blood pressure endpoints
(1 02-107 mm Hg or <92 mm Hg), which is consistent with a
factorial design study. A factorial design (as shown above)
involves 2 or more experimental interventions. each with 2 or
more variables that are studied independently. In this study, the
patients w ere being investigated to see which one of these 3
drugs caused hyperkalemia with the greatest frequency. In
addition. the drugs also had 2 different blood pressure end
points rather than one common end point to see what effect
they also had on blood pressure.
Patients: African American patients . aged 18 to 70 years. with

hypertensive CKD as defined by a diastolic blood pressure (BP)
higher than 95 mm Hg and a glomerular filtration rate (GFR)
between 20 and 65 mUmin/1.73 m2 . Specific exclusion criteria
included diabetes mellitus: urinary protein to urinary creatinine ratio
(UP/Cr) higher than 2.5: accelerated or malignant hypertension:
secondary hypertension: serious systemic disease: congestive
heart failure: and initial potassium level higher than 5.5 mEq/L
Intervention: Patients were randomized to initial treatment with

(Choice A) Cluster analysis is the grouping of different data
point into similar categories . which is not employed in this
study. Cluster analysis usually involves randomization at the
level of groups rather than at the level of individuals.
either a beta-blocker (metoprolol), an ACE inhibitor (ramipril), or a

calcium channel blocker (amlodipine) and to 1 of 2 mean arterial BP
goals (1 02- 107 mmHg or <92 mmHg).
Outcome measures: hyperkalemic event (potassium level higher

than 5.5 mEqll).
.:::J
(Choice B) A cross-over study is one in which group of

participants is randomized to one treatment for a period of time
....... ...t ...................................... : ........
........ .... ,.......................................................... f ........... ....
Notes
l~:'ll
Calculato1
:CJ
5
6
7
8
:[lJ
11
12
13
14
15
16
17
liBl
~
20
addition, the drugs also had 2 different blood pressure end

A total of 1094 non-diabetic patients were randomized. A total of

6497 potassium measurements were obtained, and 80
hyperkalemic events in 51 subjects were identified.
~ 15
RaMomlzed Oruo Groop
l!i
0 Rarniprll
0 Metroprol succinate
AmiOPidine besylale
12
"'&'!

point into similar categories, which is not employed in this
and the other group is given an alternate treatment for the
same period of time. At the end of the time period, the two
groups then switch treatments for another set period of time .
This study gave the same treatment to the patients for the
duration of the trial.
s;!
"'
.!l
a:
(Choice D) A parallel study randomizes one treatment to one

group and a different treatment to the other group, such as
treatment drug to one group versus placebo to the other group.
There are usually no other variables measured, such as the
two blood pressure goals in this study.
1211
. LBJ
23
24
25
Figure 1. Hyperkalemia event rate per 100 patientyears by randomized drug

groups and baseline glomerular filtration rate (GFR). Error bars indicated
95% confidence intervals.
Factorial design studies involve randomization to different
interventions with additional study of 2 or more variables.
Table 4. Association of Risk of Hyperkalemia With

Time-Dependent Factors In Mulllvariable Analysls 3
30
31
32
33
34
35
36
37
38
39
40
Variable
Followup GFR s 30 vs
> 50 mUmin/1.73 m2
Followup GFR > 30 to :s40 vs
> 50 mUmln/1.73 mz
Follow-up GFR > 40 to s so vs
References:
Hazard Ratio
(95% Confidence
Interval)
p
Value
0.41 (0.22-0.78)
9.07 (3.1 8-25.88)
.006
<.001
3.67 (1.21-1 1.15)
.02
1.98 (0.596.61)
.27
1. Design of experiments with multiple

independent variables: a resource
management perspective on complete and
reduced factorial designs.
Time Spent 8
seconds
Copyright
USMLEWorld ,LLC.
Last updated:
[7/20/2013]
:CJ
5
6
7
8
:[lJ
11
12
13
14
15
16
17
liBl
-~
20
1211
. LBJ
23
24
25
Item: 28 of 44
11 !" Mark
Q.ld : 7689 [
. I .
'
Variable
follow-up GFR s 30 vs
>50 mUmin/1.73 m2
Follow-up GFR >30 to :s40 vs
> 50 mUmln/1.73 mz
Follow-up GFR > 40 to :sSO vs
>50 mUmin/1.73 mz
Follow-up UP/Cr > 0.08 to :s0.22
vs :50.08
Follow-up UP/Cr > 0.22 to s 0.66
vs :50.08
Followup UP/Cr > 0.66 vs :s0.08
followup potassium level45
vs < 4 mEqll
Follow-up potassium level > 5
vs < 4 mEqll
T.
-<:1
t:>
Prevrous
Ne><t
I .
Hazard Ratio
(95% Confidence
Interval)
p
Value
0.41 (0.22-0.78)
9.07 (3.1 8-25.88)
.006
<.001
3.67 (1.21 -11.15)
.02
1.98 (0.59-6.61)
.27
2.01 (0.924.39)
.08
1.50 (0.623.63)
.37
1.84 (0.78-4.30)
7.25 (1.7230.58)
.16
.007
30.83 (6.89-138.0)
<.001
32
33
34
35
36
37
38
39
40
rot&
"
Notes
~~~
~
Calculator
n. the drugs also had 2 different blood pressure end

point into similar categories. which is not employed in this

and the other group is given an alternate treatment for the
same period of time. At the end of the time period. the two
groups then switch treatments for another set period of time.
This study gave the same treatment to the patients for the
duration of the trial.
(Choice D) A parallel study randomizes one treatment to one
group and a different treatment to the other group. such as

treatment drug to one group versus placebo to the other group.
There are usually no other variables measured. such as the
two blood pressure goals in this study.
Conclusion:
30
31
M.f
Lab -Values
Factorial design studies involve randomization to different

interventions with additional study of 2 or more variables.
In non-diabetic patients with hypertensive CKD treated with ACEis.

the risk of hyperkalemia is small.
References:
Funding Source: The National Institute of Diabetes and Digestive

and Kidney Diseases (NIDDK) grant; additional financial support
from the Office of Research in Minority Health and drug donations
from Pfizer Inc. AstraZeneca Pharmaceuticals. and King
Pharmaceuticals.
1. Design of experiments with multiple

independent variables: a resource
management perspective on complete and
reduced factorial designs.
Structured abstract is based on: Arch Intern Med.
2009; 169( 17) 1587-94
Time Spent 8
seconds
Copyright
USMLEWorld .LLC.
Last updated:
[7/20/2013]
:CJ
5
6
7
8
:[lJ
11
12
13
14
15
16
17
liBl
Item: 29 of 44
11 PMark
Q.ld: 364S [
<:J
1>-
Previous
Ne><t
jf
Lab Values
A clinical trial is performed to study the effect of a new antihypertensive drug. If the subjects of the study
change their behavior because they are aware that they are under observation. what type of bias will take
place?
r
r
r
., r
r
A. Sample distortion bias [7%]

B. Information bias [14%]
C. Confounding bias [6%]
D. Hawthorne effect [69%]
E. Not a source of bias [3%]
User ld:
Explanation:
1211
. LBJ
Bias poses a threat to the validity of any study. A sample which is not representative of the population is said
to be a biased sample. In this vignette. a biased sample with Hawthorne effect is described .
20
23
24
25
26
30
31
32
33
34
Hawthorne effect can be defined as the tendency of a study population to affect the outcome because these
people are aware that they are being studied. This awareness leads to a consequent change in behavior
while under observation. thereby seriously affecting the validity of the study. Hawthorne effect is commonly
seen in studies that concern behavioral outcomes or outcomes that can be influenced by behavioral
changes. In order to minimize the potential of the Hawthorne effect. studied subjects can be kept unaware
that they are being studied; however. this may pose ethical problems. Randomized control trials have a
sense of uncertainty and risk due to randomization. which may be more potent behavior modifiers than mere
observation.
(Choice A) Sample distortion bias is seen when the estimate of exposure and outcome association is biased
because the study sample is not representative of the target population with respect to the joint distribution of
exposure and outcome.
35
36
37
38
39
40
(Choice B) Information bias occurs due to the imperfect assessment of association between the exposure
and outcome as a result of errors in the measurement of exposure and outcome status. It can be minimized
by using standardized techniques for surveillance and measurement of outcomes. as well as trained
observers to measure the exposure and outcome.
9-
Notes
1.":~
Calculator
:CJ
5
6
7
8
:[lJ
11
12
13
14
15
16
17
liBl
~
20
1211
. LBJ
23
24
25
26
30
31
32
33
34
35
36
37
38
39
40
Item: 29 of 44
11 PMark
Q.ld: 364S [
<:1
1>-
Previous
Ne><t
jf
Lab Values
., r D. Hawthorne effect [69%]

r E. Not a source of bias [3%]
Explanation:
User ld:
Bias poses a threat to the validity of any study. A sarnple which is not representative of the population is said
to be a biased sarnple. In this vignette. a biased sarnple with Hawthorne effect is described.
Hawthorne effect can be defined as the tendency of a study population to affect the outcome because these
people are aware that they are being studied. This awareness leads to a consequent change in behavior
while under observation. thereby seriously affecting the validity of the study. Hawthorne effect is cornrnonly
seen in studies that concern behavioral outcomes or outcomes that can be influenced by behavioral
changes. In order to rninirnize the potential of the Hawthorne effect. studied subjects can be kept unaware
that they are being studied; however. this rnay pose ethical problems. Randomized control trials have a
sense of uncertainty and risk due to randomization. which rnay be rnore potent behavior modifiers than rnere
observation.
(Choice A) Sarnple distortion bias is seen when the estimate of exposure and outcome association is biased
because the study sarnple is not representative of the target population with respect to the joint distribution of
exposure and outcome.
(Choice B) Information bias occurs due to the imperfect assessment of association between the exposure
and outcome as a result of errors in the rneasurernent of exposure and outcome status. It can be rninirnized
by using standardized techniques for surveillance and rneasurernent of outcomes. as well as trained
observers to measure the exposure and outcome.
(Choice C) Confounding bias occurs due to the presence of one or rnore variables associated independently
with both the exposure and the outcome. For example. cigarette smoking can be a confounding factor in
studying the association between maternal alcohol drinking and low birth weight babies. as cigarette smoking
is independently associated with both alcohol consumption and low birth weight babies.
Know the different kinds of bias. which can decrease the validity of study results. Hawthorne effect is the
tendency of the study population to affect the outcome since they are aware that they are being studied.
Tirne Spent 2 seconds
9''
Notes
1.":~
Calculator
:CJ
5
6
7
8
:[lJ
11
12
13
14
15
16
17
liBl
~
20
1211
. LBJ
23
24
25
Item: 30 of 44
11 PMark
Q.ld: 3932 [
<:J
1>-
Previous
Ne><t
jf
Lab Values
A new test is devised to measure serum cholesterol level. A sample of blood is taken from a patient. and the
test is performed three times. The results are 200. 190 and 184 mg/dl. Based on these results. one can
conclude that the new test is
., r
r
r
r
r
A. Not reliable [62%]

B. Not valid [4%]
C. Not accurate [28%]
D. Not sensitive [3%]
E. Not specific [3%]
User ld:
Explanation:
This example deals with test-retest reliability. A reliable test gives similar or very close results on repeat
measurements. In this example. repeat measurements of the same sample yielded different results;
therefore. the new test is not reliable .
(Choices B and C) Validity or accuracy is defined as the test's ability to measure what it is supposed to
measure. In order to determine the validity of a test. the results are compared to those obtained from the gold
standard test. In this case. since there were no test results obtained using the gold standard. the validity or
accuracy of the test cannot be determined.
(Choices D and E) The sensitivity and specificity of a test compare results to those obtained using the gold
standard. These parameters can show how accurate the results are. but do not measure reliability.
32
33
34
35
36
37
38
39
40
A reliable test gives similar results on repeat measurements. Reliability is maximal when random error is
minimal.
9-
Notes
1.":~
Calculator
:CJ
5
6
7
8
:[lJ
11
12
13
14
15
16
17
liBl
~
Item: 31 of 44
11 PMark
Q.ld: 4121 [
-<J
C>-
Previous
Ne><t
jf
Lab Values
A large-scale randomized. double-blinded clinical trial was conducted to evaluate the effect of beta-blocker
therapy on the survival of patients with chronic heart failure. class IV. The patients with severe heart failure
were randomly assigned to receive either carvedilol (a beta-blocker) or a placebo. In their report of the results
of the study. the investigators included the table with baseline characteristics (i.e .. age. race. prevalence of
hypertension. etc.) of the patients in the treatment and placebo groups. According to the table. both groups
had a similar distribution of these characteristics. With the information given. which of the following is most
probable?
r
r
r
., r
r
A. The sample size is adequate [5%]

B. The study is negative [4%]
C. The power of the study is big [8%]
D. Randomization is successful [74%]
E. Observer's bias might be an issue [9%]
20
1211
. LBJ
23
24
25
26
1271
~
29
30
32
33
34
35
Explanation:
User ld:
The purpose of randomization is to make the distribution of all potential confounders even (i.e .. between the
treatment and placebo groups). Unlike all the other methods of controlling confounding (e.g .. matching.
stratified analysis). randomization potentially controls known. as well as unknown confounders. One of the
methods to assess the adequacy of randomization is to look at the distribution of baseline characteristics in
both groups. If they are similar. one can assume that the randomization evenly distributed the confounders
between the groups. and that randomization was successful.
(Choices A and C) From the information given. we cannot judge the adequacy of the sample size to detect
the difference in the survival (if it is present) between the treatment and placebo groups. We also cannot
judge the power of the study based on the above information.
(Choice B) The results of the study were not mentioned.
(Choice E) Observer's bias can be controlled with blinding. not randomization.
36
37
38
39
40
In clinical trials. randomization is said to be successful when a similarity of baseline characteristics of the
patients in the treatment and placebo groups is seen.
""
Notes
1.":~
Calculator
:CJ
5
6
7
8
:[lJ
11
12
13
14
15
16
17
liBl
~
20
1211
. LBJ
23
24
25
26
1271
~
29
30
Item: 32 of 44
11 PMark
Q.ld : 4184 [
-<J
C>-
Previous
Ne><t
jf
Lab Values
A 35-year-old Caucasian female presents to the office due to a self-palpated breast mass. After the
appropriate work-up. fine-needle aspiration (FNA) of the mass is performed. The results of the FNA return as
negative. As you are explaining the test result. the patient asks. "What are the chances that I really do not
have breast cancer?" Which of the following values best addresses this patient's question?
r
r
r
., r
r
A. Sensitivity [11 %]
B. Specificity [13%]
C. Positive predictive value [7%]
D. Negative predictive value [63%]
E. Validity [5%]
Explanation:
User ld:
The negative predictive value (NPV) is defined as the probability of being free of a disease if the test result is
negative. One very important thing to remember is that the NPV will varv with the pretest probability of a
disease. A patient with a high probability of having a disease will have a low NPV. and a patient with a low
probability of having a disease will have a high NPV. Specific examples are given below:
1) Breast cancer and FNA test results
A patient with a high pre-test probability for having breast cancer (e.g .. female. has first degree relatives
with breast cancer. greater than 40 years old) has a low NPV. A patient with a low pre-test probability of
having breast cancer (e.g .. less than 40 years old. as in this case) has a high NPV.
2) HIV and ELISA test results
36
A patient who belongs to a high-risk group (e.g .. multiple sexual partners. admits to not using condoms.
IV drug user) has a high pre-test probability; consequently. this patient will have a low NPV. On the other
hand. a patient who belongs to a low-risk group (e.g .. one sexual partner who has no other sexual
partners. uses condoms correctly all the time. no history of IV drug use) has a low pre-test probability;
consequently. this patient will have a high NPV.
37
38
39
'Note: The prevalence of a disease is directly related to the pre-test probability of having the disease. and also
affects the NPV.
34
35
40
""
Notes
1.":~
Calculator
:CJ
5
6
7
8
:[lJ
11
12
13
14
15
16
17
liBl
~
20
1211
. LBJ
23
24
25
26
1271
~
29
30
Item: 32 of 44
!il f> Mark
Q.ld : 4184 [
<:J
C>-
Previous
Ne><t
jf
Lab Values
negar1ve. une very 1mporranr rmng ro rememoer 1s mar rne 1\ft-'V wm varv w1rn rne preresr prooaomrv or a
disease. A patient with a high probability of having a disease will have a low NPV. and a patient with a low
probability of having a disease will have a high NPV. Specific examples are given below:
1) Breast cancer and FNA test results
A patient with a high pre-test probability for having breast cancer (e.g .. female. has first degree relatives
with breast cancer. greater than 40 years old) has a low NPV. A patient with a low pre-test probability of
having breast cancer (e.g .. less than 40 years old. as in this case) has a high NPV.
2) HIV and ELISA test results
A patient who belongs to a high-risk group (e.g .. multiple sexual partners. admits to not using condoms.
IV drug user) has a high pre-test probability; consequently. this patient will have a low NPV. On the other
hand. a patient who belongs to a low-risk group (e.g .. one sexual partner who has no other sexual
partners. uses condoms correctly all the time. no history of IV drug use) has a low pre-test probability;
consequently. this patient will have a high NPV.
'Note: The prevalence of a disease is directly related to the pre-test probability of having the disease. and also
affects the NPV.
(Choices A and B) The sensitivity and specificity of a test are fixed values which do not vary with the pretest
probability of a disease. Most researchers agree that the ideal diagnostic test should have high sensitivity and
specificity. In this case. FNA also has a high sensitivity and specificity; however. the statistical parameter
being described by the patient and physician was the high negative predictive value.
(Choice C) The positive predictive value follows the same concept. but applies if the test result is positive.
(Choice E) Validity represents the appropriateness of the test (i.e .. the test measures what it is supposed to
measure). It does not depend on the pretest probability of the disease.

34
35
36
37
38
39
40
NPV is the probability of being free of a disease if the test result is negative. Remember: the NPV will varv
with the pretest probability of a disease. A patient with a high probability of having a disease will have a low
NPV. and a patient with a low probability of having a disease will have a high NPV.
Last updated [11/1 0/2011]
Notes
1.":~
Calculator
:CJ
5
6
7
8
:[lJ
11
12
13
14
15
16
17
liBl
~
20
1211
. LBJ
23
24
25
26
1271
Item: 33 of 44
<:::1
C>-
Previous
Ne><t
11 f>Mark
Q.ld: <W19 [
if
Lab Values
Two studies were conducted in the same population to assess the relationship between oral contraceptive
use and the risk of deep venous thrombosis (DVT). Study A demonstrated an increased risk of DVT among
oral contraceptive users. with a relative risk of 2.0 and 95% confidence interval of 1.2-2 .8. Study B showed a
relative risk of 2.01 and 95% confidence interval of 0.8-3.1 . Which of the following statements is most likely
true about these 2 studies?
r A. The result in study A is not statistically significant [6%]
r
r
., r
r
B. The result in study B is biased [9%]

C. The result in study A is not accurate [3%]
D. The sample size in study B is smaller than in study A [65%]

E. The P value in study B is less than 0.05 [16%]
Explanation:
User ld:
Statistical significance can be expressed with either P values or confidence intervals. but both are interrelated:
1. The P value is inversely related to the confidence interval. A P value of 0.05 corresponds to a 95%
confidence interval. while a P value of 0.01 is equivalent to a 99% confidence interval.
2. If the null value (ie. 1.0) lies outside of a given confidence interval. then the P value of the study is less
than the equivalent confidence interval value. Conversely. if the null value is within a given confidence
interval. then the P value of the study is greater than or equal to the equivalent confidence interval
value.
~
29
30
31
95% confidence
interval
99% confidence
interval
32
Null value outside

confidence interval
P value< 0.05
P value< 0.01
34
Null value inside

confidence interval
35
36
37
38
39
40
value~
0.05
value~
0.01
In a statistically significant study. the P value should be less than 0.05. This corresponds to a 95% confidence
interval that does not include the null value. Study A is statistically significant as the 95% confidence interval
rlnP.~ nnt inr.h !riP. thP. mill v;:~h IP. IChoice A\ ThP. Rl;% r.nnfir!P.nr.P. intP.rv;:~l fnr ~til rill R inr.l11r!P.~ thP. mill v;:~IIIP.
Notes
1.":~
Calculator
:CJ
5
6
7
8
:[lJ
11
12
13
14
15
16
17
liBl
~
20
1211
. LBJ
23
24
25
26
1271
~
29
30
31
32
34
35
36
37
38
39
40
Item: 33 of 44
!il PMark
Q.ld: <W19 [
-<J
C>-
Previous
Ne><t
jf
Lab Values
1. The P value is inversely related to the confidence interval. A P value of 0.05 corresponds to a 95%
confidence interval. while a P value of 0.01 is equivalent to a 99% confidence interval.
2. If the null value (ie. 1.0) lies outside of a given confidence interval. then the P value of the study is less
than the equivalent confidence interval value. Conversely. if the null value is within a given confidence
interval. then the P value of the study is greater than or equal to the equivalent confidence interval
value.
95% confidence
interval
99% confidence
interval
Null value outside

confidence interval
P value< 0.05
P value< 0.01
Null value inside

confidence interval
value~
0.05
value~
0.01
In a statistically significant study. the P value should be less than 0.05. This corresponds to a 95% confidence
interval that does not include the null value. Study A is statistically significant as the 95% confidence interval
does not include the null value (Choice A). The 95% confidence interval for study B includes the null value .
and thus the study has a P value greater than 0.05 (Choice E).
The confidence interval in study B is wider than in study A. most likely due to a smaller sample size. A small
sample size can lead to lack of statistical significance from insufficient power to detect the difference between
exposed and unexposed subjects. Increasing the sample size would increase the studies power and make
the confidence interval tighter. By doing so. the null value can be excluded. thus increasing statistical
significance.
(Choices B & C) The accuracy or validity of the studies cannot be judged because there is no information on
how they were designed and conducted.
The power of a study represents its ability to detect a difference between 2 groups (eg. exposed versus
nonexposed) when there truly is a difference. Increasing the sample size increases the power of a study and
consequently narrows the confidence interval surrounding the point estimate (best guess). Confidence
intervals express statistical significance and can be converted into P values.
Notes
1.":~
Calculator
:CJ
5
6
7
8
:[lJ
11
12
13
14
15
16
17
liBl
Item: 34 of 44
Q.ld: 4107 [
<J
C>-
Previous
Ne><t
1111 PMark
jf
Lab Values
Two cross-sectional studies were conducted using different questionnaires to determine the prevalence of
over-the-counter analgesics use in a population. The first study showed a prevalence of 7.5% (95%
confidence interval 6.0 - 9 .0). and the second study demonstrated a prevalence of 7.3% (95% confidence
interval 6.9 - 7 .6). If the true prevalence of over-the-counter analgesics use in the population is 7.4%. which of
the following statements about the results of the study is the most accurate?
r
r
r
r
., r
A. The first study results are more specific [3%]

B. The second study results are more sensitive [9%]
C. The first study results are more valid [5%]
D. The first study results are more accurate [5%]
E. The second study results are more precise [78%]
Explanation:
1211
. LBJ
Precision is the measure of random error in the study. The study is precise if the results are not scattered
widely; this is reflected by a tight confidence interval. The first study has a wider confidence interval
compared to the second study; therefore. the second study is more precise.
User ld:
20
23
24
25
26
1271
~
29
30
31
32
36
37
38
39
40
(Choices A and B) Specificity and sensitivity are measures of validity. The sensitivity and specificity of the
questionnaires used in these studies cannot be determined from the given information.
(Choices C and D) Validity and accuracy are measures of systematic error (bias). Accuracy is reduced if
the result does not reflect the true value of the parameter measured. Increasing the sample size increases
the precision of the study. but does not affect accuracy. In our case. the results of both studies are pretty
close to the true value. and are thus seemingly accurate.
Precision is the measure of random error. The tighter the confidence interval. the more precise the result.
Increasing the sample size increases precision.
9-
Notes
1.~
Calculator
:CJ
5
6
7
8
:[lJ
11
12
13
14
15
16
17
liBl
~
20
1211
. LBJ
23
24
25
26
1271
~
29
30
31
32
Item: 35 of 44
11 PMark
Q.ld: 4109 [
-<J
C>-
Previous
Ne><t
jf
Lab Values
A case-control study was conducted to assess the relationship between tampon use and toxic shock
syndrome (TSS). The odds ratio for tampon use comparing the patients with TSS and the healthy subjects
(controls) was calculated and turned outto be 5.0 (p < 0 .DO 1). The authors concluded thatthe risk of TSS is
5 times higher in tampon users. The conclusion is valid if which of the following assumptions is satisfied?
r
r
r
., r
r
A. The sample size is big [33%]

B. The odds ratio is highly statistically significant [16%]
C. The confidence interval for the odds ratio is tight [23%]
D. TSS is a rare disease [19%]
E. The exposure is not associated with the disease [9%]
Explanation:
User ld:
Case-control studies are very popular in exploring an exposure-disease association. because they are
relatively cheap and less time-consuming than cohort studies. One of the major drawbacks of case-control
studies is the fact that the risk cannot be derived directly from their results .
Exposure-odds ratio is the measure of association in case-control studies. It compares the odds of exposure
in cases to the odds of exposure in controls. It is NOT the same as relative risk. Relative risk can be
calculated in follow-up studies by comparing the risk in exposed individuals to the risk in unexposed
individuals. Direct calculation of the relative risk is not possible in case-control studies. because the study
design does not include following people over time. Nevertheless. the relative risk can sometimes be
approximately equal to the odds ratio. If the prevalence of the disease is low. the exposure odds ratio
approximates the relative risk. This statement is called 'the rare disease assumption' and represents one of
the fundamental epidemiologic concepts.
34
(Choices A, B and C) Increasing the sample size would decrease the 'p' value for the odds ratio and make
the confidence interval tighter. The precision of the estimate would increase; but. that fact does not affect the
approximation of the odds ratio to the relative risk.
36
(Choice E) The results of the study suggest that the exposure is associated with the disease.
33
37
38
39
40
If the outcome of a case-control study is not common in the population. the odds ratio is close to the relative
risk.
""
Notes
1.":~
Calculator
:CJ
5
6
7
8
Item: 36 of 44
<::::1
t>
Previous
Next
!il PMark
Q.ld: 2138 [
jf
Lab Values
Notes
!:':~
Calculator
Researchers at a large pharmaceutical company discover a tumor-specific antigen present in high quantities
in the serum of patients with pancreatic cancer. A study is then performed to evaluate serum levels of the
tumor marker in patients with and withoutthe disease. The following curves are generated using the results
of the study.
:[lJ
0 Healthy
D Diseased
11
12
13
14
15
16
17
....<=
Ill
Ql
.::
liBl
..."'a.....
20
..c
E
1211
. LBJ
Ql
:::l
23
24
25
26
1271
($) USMLEWorld, LLC
29
30
31
32
33
38
39
40
The drug company decides to utilize the newly discovered tumor antigen by creating a screening test for
pancreatic cancer. During preliminary design of the test, the cutoff point for positive/negative results is set at
point B. If the cutoff point is moved from B to A, the sensitivity of the screening test will change in which of the
following ways?
r
r
., r
r
A. Cannot be determined [0%]

B. Decreased [15%]
C. Increased [84%]
D. Unchanqed fO%1
.::1
:CJ
Item: 36 of 44
<:J
C>-
Previous
Ne><t
ill PMark
Q.ld: 2138 [
Test cutoff
5
6
7
8
I - - Negative test ---il
:[lJ
1
l-- - Positive test - - - - ;
...."'c
41
;;;
11
..."'...c.
12
13
14
15
16
17
41
..c
E
::l
liBl
~
20
1211
. LBJ
23
24
25
26
1271
~
29
30
31
32
33
38
39
40

U SMLEWo~d.LLC

jf
Lab Values
Notes
1.":~
Calculator
:CJ
Item: 36 of 44
Q.ld: 2138 [
11
12
13
14
15
16
17
liBl
~
20
1211
. LBJ
23
24
25
26
1271
~
29
30
31
32
33
-<:J
C>-
Previous
Ne><t
l/
5
6
7
8
:[lJ
11 PMark
Antigen concentration
jf
Lab Values
40
1.~
Calculator
"-
U SMLEWo~d.LLC

In this example. moving the cutoff point from B to A will cause more patients with the disease to test positive
(fewer FNs). increasing the sensitivity of the test. However. as a consequence. more patients without the
disease will also test positive (more FPs). resulting in decreased specificity. Conversely. moving the cutoff
point in the other direction (eg. A to B) will increase the number of FNs and decrease the number of FPs.
decreasing sensitivity while increasing specificity.
Changing the cutoff point of a quantitative diagnostic test will inversely affect its sensitivity and specificity.
Typically. lowering the cutoff value will increase sensitivity (fewer false negatives) and decrease specificity
(more false positives). Screening tests need high sensitivity. and confirmatory tests need high specificity.
38
39
Notes
Last updated [6111120 13]
:CJ
5
6
7
8
:[lJ
Item: 37 of 44
t>
Ne><t
Lab Values
It is hypothesized that high glucose levels during an episode of acute myocardial infarction may be associated
with higher short-term and long-term cardiovascular outcomes in non-diabetic individuals. A group of
investigators studied the relationship between the blood glucose level on admission and the number of
episodes of significant ventricular arrhythmias during the first 24 hours after an acute myocardial infarction in
patients with no previous history of diabetes. The following plot was constructed.
Blood sugar level
11
12
13
14
15
16
17
150
140
130
120
liBl
110
20
1211
. LBJ
23
24
25
<j
Previous
ill PMark
Q.ld: 4315 [
10
20
30
40
50
Incidence ot errhyltwnie episodes
Which of the following is the best statement about the study results?
26
1271
~
29
30
31
32
33
r
r
., r
r
r
A. The correlation coefficient is negative [2%]

B. The correlation coefficient is close to 0 [3%]
C. A linear association is demonstrated [88%]
D. Confounders are effectively controlled [1 %]
E. High blood sugar levels cause arrhythmia [4%]
34
35
36
38
39
40
Explanation:
User ld:
Scatter plots are useful for crude analysis of data. These can demonstrate the type of association (i.e .. linear.
non-linear). if any is present. If a linear association is present. the correlation coefficient can be calculated to
provide a numerical description of the linear association. In this case. the scatter plot shows an almost
.
.
.
r.:J-
Notes
1.":~
Calculator
:CJ
5
6
7
8
:[lJ
11
12
13
14
15
16
17
Item: 37 of 44
10
<:J
C>-
Previous
Ne><t
!II PMark
Q.ld: 4315 [
20
30
40
jf
Lab Values
Which of the following is the best statement about the study results?
r
r
., r
r
r
A. The correlation coefficient is negative [2%]

B. The correlation coefficient is close to 0 [3%]
C. A linear association is demonstrated [88%]
D. Confounders are effectively controlled [1 %]
E. High blood sugar levels cause arrhythmia [4%]
20
non-linear). if any is present. If a linear association is present. the correlation coefficient can be calculated to
provide a numerical description of the linear association. In this case. the scatter plot shows an almost
perfect linear association between the blood glucose level on admission and the number of episodes of
significant ventricular arrhythmias.
23
24
25
26
1271
~
29
30
31
32
33
34
35
36
38
39
40
Incidence ot errhyltwnie episodes
Explanation:
1211
. LBJ
1.":~
Calculator
50
liBl
~
Notes
User ld:
(Choice A) The scatter plot demonstrates a positive association (i.e .. an increase in the blood glucose level is
associated with an increase in the number of episodes of significant ventricular arrhythmias); therefore. the
correlation coefficient is positive.
(Choice B) The correlation coefficient of an almost perfect linear association is close to 1.
(Choices D and E) Crude analysis of the association using scatter plots does not account for possible
confounders (e.g .. severity of the disease. degree of sympathetic activation. etc.). and does not necessarily
imply causal relationships between variables.
non-linear). if any is present.
:CJ
5
6
7
8
:[lJ
11
12
13
14
15
16
17
Item: 38 of 44
11 PMark
Q.ld: 3915 [
<:1
1>-
Previous
Ne><t
jf
Lab Values
A study was conducted to assess the relationship between ethnicity and end-stage renal disease. Two
groups of pathologists independently studied specimens from 1.DOD kidney biopsies. The first group of
pathologists was aware of the race of the patient from whom the biopsy came. while the second group was
blinded from the patient's race. The first group reported 'hypertensive nephropathy' much more frequently for
black patients than the second group. Which of the following types of bias is most likely present in this study?
r
r
r
r
., r
A. Confounding [3%]
B. Respondent bias [4%]
C. Recall bias [2%]
D. Selection bias [5%]
E. Observer bias [84%]
liBl
Explanation:
20
Observer bias occurs when the investigator's decision is adversely affected by knowledge of the exposure
status. In this case. some pathologists' decisions were influenced by the fact that hypertensive nephropathy
is a common cause of ESRD in the black population. The pathologists who were blinded from the patients'
race were not under this influence. so their interpretation was more unbiased.
1211
. LBJ
23
24
25
26
1271
~
29
30
31
32
User ld:
(Choice A) Confounding is present when at least part of the exposure-disease relationship can be explained
by another variable (confounder). No information on possible confounders is given in this scenario.
(Choice B) Respondent bias is present when the outcome is obtained by the patient's response. and not by
objective diagnostic methods (e.g .. migraine headache). In this case. the diagnosis was ascertained via
kidney biopsy.
(Choice C) Recall bias results from inaccurate recall of past exposure by patients. It is not applicable to this
case.
33
34
35
(Choice D) Selection bias results from the manner in which the subjects are selected for the study. or from
the selective losses from the follow-up.
36
Observer bias occurs when the investigator's decision is adversely affected by knowledge of the exposure
status.
9''
Notes
1.":~
Calculator
Item: 39 of 44
:CJ
Q.ld: <W02 [
:[lJ
., r
r
r
r
r
5
6
7
8
11
12
13
14
15
16
17
liBl
~
20
1211
. LBJ
23
24
25
11 PMark
<J
C>-
Previous
Ne><t
jf
Lab Values
Ten measurements of systolic blood pressure were obtained from a patient in the ICU using an intra-arterial
cannula over several hours. The maximal value recorded is 120 mmHg. and the minimal value is 1DO mmHg.
If the next measurement shows the value of 240 mmHg. which of the following is most likely to happen?
A. The mean would increase significantly [80%]

B. The median would increase significantly [1 0%]
C. The mode would increase significantly [4%]
D. The standard deviation would not change [4%]
E. The range would not change [1 %]
Explanation:
User ld:
An outlier is defined as an extreme and unusual value observed in a dataset. It may be the result of a
recording error. a measurement error. or a natural phenomenon. An outlier can affectthe measures of central
tendency. as well as the measures of dispersion. For example. the mean is extremely sensitive to outliers
and easily shifts toward them. In this case. the value of 240 mmHg is the outlier.
(Choice B) The median is much more resistant to outliers. because it is located in the middle of the dataset
where the observations usually do not differ much from each other.
26
1271
~
29
30
31
32
33
34
35
36
37
38
(Choice C) The mode is not affected by outliers. because they do not change the most frequent value
observed.
(Choice D) The standard deviation is sensitive to outliers. because it is the measure of dispersion within the
dataset. and outliers significantly increase the dispersion.

(Choice E) The range is equal to the maximal value minus the minimal value; therefore. it would definitely
change.
An outlier is defined as an extreme and unusual observed in a dataset. The mean is very sensitive to outliers
and easily shifts toward them. The median and mode are more resistant to outliers.
9-
Notes
1.~
Calculator
:CJ
5
6
7
8
:[lJ
11
12
13
14
15
16
17
liBl
~
20
1211
. LBJ
23
24
25
26
1271
~
29
30
31
32
33
34
Item: 40 of 44
11 PMark
Q.ld: 4172 [
-<J
C>-
Previous
Ne><t
jf
Lab Values
In a large population with little migration. the incidence of diabetes mellitus. type II is equal to 3 cases per
1.000 per year. and has been stable for the last 30 years. The prevalence of this disease increased
progressively over the same period. Which of the following could be the most likely explanation of this trend
overtime?
r
r
., r
r
r
A. High mortality in diabetics [2%]

B. Selective survival [4%]
C. Improved quality of care [84%]
D. Decreased hospitalization rate [2%]
E. Increased diagnostic accuracy [8%]
User ld:
Explanation:
Incidence and prevalence are two important concepts in epidemiology. It is thus very important to know the
difference between these two. Incidence is the measure of the appearance of new cases. Prevalence is the
measure of those with the disease in the population at a particular point in time. The relationship between
these two categories can be demonstrated by the following approximation in a stable population (little
migration)
Prevalence
The above vignette described a disease with a rising prevalence but stable incidence. More people are being
documented to have the disease. while the number of people being diagnosed remains the same. Such trend
can be attributed to factors which prolong the duration of the disease (e.g .. improved quality of care).
(Choice A) A high mortality rate in diabetics would result in a decreased prevalence.
(Choices B and D) Selective survival and decreased hospitalization rate are not helpful in explaining the
increased prevalence described in this scenario.
35
36
37
38
(Choice E) Increased diagnostic accuracy affects both the prevalence and incidence of a disease.
An increasing prevalence and stable incidence can be attributed to factors which prolong the duration of a
disease ( e .a .. imoroved aualitv of care - this scenario is tvoical for the USMLEl.
""
Notes
1.":~
Calculator
.
.
1
2
4
5
6
7
8
[;J
10
11
12
13
14
15
16
17
Item: 41 of 44
11 PMark
Q.ld: 3992 [
<:J
C>-
Previous
Ne><t
jf
Lab Values
A group of investigators are planning a study to evaluate the relationship between serum fibrinogen level and
the incidence of acute coronary syndrome. They assume that serum fibrinogen level is a normally distributed
variable in the population of interest. Which of the following statements is most consistent with this
assumption?
r
r
., r
r
r
A. Mean is greater than median [5%]

B. Mean is greater than mode [3%]
C. Mean is equal to median [85%]
D. Median is greater than mean [3%]
E. Mode is greater than mean [3%]
[m
19
. 20
-~
22
23
24
25
26
c:]
28
29
30
31
32
33
34
35
36
37
38
39
40
Explanation:
User ld:
The normal distribution is one of the most popular statistical distributions. Interestingly. many variables in real
life (e.g .. laboratory values) are normally distributed or close to normal. The normal distribution has some nice
statistical properties. and is easy to work with. It is bell-shaped and symmetric. Consequently. all its
measures of central tendency are equal: mean = median = mode (In real life. the values are very close to
each other). Skewed distributions do not have this property.
(Choices A and B) In a positively skewed distribution (tail on the right). the mean is greater than the median
and greater than the mode.

(Chocies D and E) In a negatively skewed distribution (tail on the left). the mean is less than the median and
less than the mode.

A normal distribution is symmetric and bell-shaped. All its measures of central tendency are equal: mean=
median= mode.
Notes
1.~
Calculator
.
.
1
2
4
5
6
7
8
[;J
10
11
12
13
14
15
16
17
[m
19
. 20
-~
22
23
24
25
26
c:]
28
29
30
31
32
33
34
35
36
37
38
39
40
Item: 42 of 44
<J
C>
Previous
Ne><t
!il PMark
Q.ld: 3909 [
II
Lab Values
A prospective cohort study was conducted to evaluate the long-term effects of a high-fat diet on the incidence
of colon cancer. The study participants were randomly selected from the population of interest. Dietary
patterns were assessed through the use of periodic self-completed questionnaires. The investigators
reported a 5-year relative risk of 1.60 for people who consumed a high-fat diet compared to individuals who
consumed a low-fat diet. The 95% confidence interval was 1.02 to 2.15. This study is most likely to have
which of the following p values?
.., r
r
r
r
r
A. 0.04 [81 %]
B. 0.06 [7%]
c. 0.09 [5%]
D. 0.11 [3%]
E. 0.20 [3%]
Explanation:
User ld:
Relative risk (RR) is the probability of the outcome of interest occurring in the exposed group compared to the
probability of it occurring in the non-exposed group. The null value of the RR is 1.0; a RR of 1.0 means that
the outcome occurs with equal frequency in both groups and that there is no association between the
exposure and the outcome. A RR > 1.0 means that the outcome occurs more frequently in the exposed
group (positive association). The RR says nothing about the statistical significance of a study.
Statistical significance can be expressed with either p values or confidence intervals. but both are
interrelated. For instance. p < 0.05 corresponds to a 95% confidence interval that does not contain the null
value. Likewise p < 0.01 is equivalent to a 99% confidence interval that does not contain the null
value. Conversely. if the null value is within a given confidence interval. then the p value is~ the equivalent
confidence interval.
Null value outside

confidence intervals
95% confidence
99% confidence
interval
interval
p<O.OS
p < 0.01
Notes
1.":~
Calculator
.
.
1
2
4
5
6
7
8
[;J
10
11
12
13
14
15
16
17
[m
19
20
.
- ~
22
23
24
25
26
c:]
28
29
30
31
32
33
34
35
36
37
38
39
40
Item: 42 of 44
<:J
C>-
Previous
Ne><t
!il PMark
Q.ld: 3909 [
'
'
"
'-'
jf
Lab Values
'
exposure and the outcome. A RR > 1.0 means thatthe outcome occurs more frequently in the exposed
group (positive association). The RR says nothing about the statistical significance of a study.
Statistical significance can be expressed with either p values or confidence intervals. but both are
interrelated. For instance. p < 0.05 corresponds to a 95% confidence interval that does not contain the null
value. Likewise p < 0.01 is equivalent to a 99% confidence interval that does not contain the null
value. Conversely. if the null value is within a given confidence interval. then the p value is~ the equivalent
confidence interval.
Null value outside

confidence intervals
Null value inside
confidence interval
95o/o confidence
99% confidence
interval
interval
p<O.OS
p < 0.01
p~O.O l
0.05
@ USMIWorld. l l C
In general. for study results to be statistically significant. the 95% confidence interval must not contain the null
value. This is equivalent to p < 0.05. which means that there is less than a 5% chance that the results are
due to chance alone. In this case. the 95% confidence interval is 1.02 to 2.15 and does not contain the null
value of 1.0. This corresponds to p < 0.05. thus 0.04 is the only correct answer choice.
(Choices 8, C, D, and E) These p values are all> 0.05. meaning the associated 95% confidence interval
would contain the null value.
Confidence intervals and p values are interrelated and express the statistical significance of a study. In a
statistically significant study. p should be< 0.05. This corresponds to a 95% confidence interval that does not
include the null value.
Last updated [7121/20 13]
Notes
1.":~
Calculator
.
.
1
2
4
5
6
7
8
[;J
10
11
12
13
14
15
16
17
[m
19
20
.
-~
22
23
24
25
26
c:]
28
29
30
31
32
33
34
35
36
37
38
39
40
Item: 43 of 44
Q.ld: 3982 [
11 PMark
-<:1
C>-
Previous
Ne><t
jf
Lab Values
Consecutive readings of pulmonary capillary wedge pressure (PCWP) were obtained from a patient in the
Intensive Care Unit (ICU) using a Swan-Ganz catheter. The readings are 20 mmHg. 22 mmHg. 21 mmHg. 22
mmHg. and 18 mmHg. Which of the following is the median of the values given above?
r
r
.., r
r
r
A. 18 mmHg [2%]
B. 20 mmHg [8%]
C. 21 mmHg [79%]
D. 22 mmHg [5%]
E. 20.6 mmHg [6%]
Explanation:
User ld:
It is important to know the difference between the measures of central tendency. The median of a dataset is
the number that divides the right half of the data from the left half. In this case. 21 mmHg is in the middle of
the dataset; therefore. it is the median. If the number of observations is even. finding the median becomes
tricky. You should find the middle two values. add them together. and divide by two.
(Choices A and B) are not measures of the center in this dataset.
(Choice E) To find the mean of a dataset. you should add all the observations and divide that sum by the
number of observations. In this case. the mean is equal to 20.6 mmHg.
(Choice D) Another measure of the center of a dataset is the mode. Finding the mode is the easiest. The
mode is the most frequent value of a dataset. In the scenario described. the mode is 22 mmHg.
The median is the value that is located in the middle of a dataset. It divides the right half of the data from the
left half.
Notes
1.":~
Calculator
w
2
4
5
6
7
8
[;J
10
11
12
13
14
15
16
17
[m
19
. 20
-~
22
23
24
25
26
c:]
28
29
30
31
32
33
34
35
36
37
38
39
40
Item: 44 of 44
11 PMark
Q.ld: 3941 [
<:::1
1>-
Previous
Ne><t
jf
Lab Values
A large clinical trial is performed to evaluate the treatment response to statins in statin-naive patients admitted
for anginal chest pain. Patients are randomly assigned to moderate- or high-dose statin therapy upon
discharge. At 3 months. the adverse cardiovascular event rate in the high-dose group compared to the
low-dose group was 0.98 with a p value of 0.80. At 1 year. the relative risk ratio was 0.67 with a p value of
0.01. Treatment and follow-up compliance was high in both groups throughout the study interval. The
difference between the 2 risk estimates is best explained by which of the following?
r
., r
r
r
r
A. Confounding bias [5%]

B. Latency period [65%]
C. Observer bias [3%]
D. Selection bias [4%]
E. Selective survival bias [23%]
Explanation:
User ld:
Latency is an important natural phenomenon of disease epidemiology. Most infectious diseases have
relatively short latency periods (ie. the time elapsed from initial exposure to clinically apparent disease). In
contrast. some disease processes (eg. cancer. heart disease) have a long latency period before clinical
manifestations develop.
The concept of a latency period can also be extended to risk factors and risk reducers. Sometimes. a
significant amount of time must pass before exposure to a risk modifier has a clinically evident effect on the
disease process. In addition. exposure to a risk modifier may need to occur continuously over a certain period
before the disease outcome is affected. In this case. at least 1 year of high-dose statin therapy was required
to show a significant protective advantage over moderate-dose therapy.
(Choice A) Confounding bias occurs when a perceived exposure-disease relationship is actually caused by
an extraneous factor that correlates with both the exposure and the disease. There is no information
suggesting that confounding occurred in this study.
(Choice C) Observer bias occurs when an observer misclassifies data due to individual differences in
interpretation or preconceived expectations regarding the study. It can be reduced by performing a
double-blind study with multiple observers to encode and verify recorded data.
(Choice D) Selection bias can occur with inappropriate (ie. nonrandom) assiqnment methods or throuqh
Notes
1.~~
Calculator
w
2
4
5
6
7
8
[;J
10
11
12
13
14
15
16
17
[m
19
. 20
-~
22
23
24
25
26
c:]
28
29
30
31
32
33
34
35
36
37
38
39
40
Item: 44 of 44
Q.ld: 3941 [
!il PMark
<::1
C>-
Previous
Ne><t
Explanation:
jf
Lab Values
User ld:
Latency is an important natural phenomenon of disease epidemiology. Most infectious diseases have
relatively short latency periods (ie. the time elapsed from initial exposure to clinically apparent disease). In
contrast. some disease processes (eg. cancer. heart disease) have a long latency period before clinical
manifestations develop.
The concept of a latency period can also be extended to risk factors and risk reducers. Sometimes. a
significant amount of time must pass before exposure to a risk modifier has a clinically evident effect on the
disease process. In addition. exposure to a risk modifier may need to occur continuously over a certain period
before the disease outcome is affected. In this case. at least 1 year of high-dose statin therapy was required
to show a significant protective advantage over moderate-dose therapy.
(Choice A) Confounding bias occurs when a perceived exposure-disease relationship is actually caused by
an extraneous factor that correlates with both the exposure and the disease. There is no information
suggesting that confounding occurred in this study.
(Choice C) Observer bias occurs when an observer misclassifies data due to individual differences in
interpretation or preconceived expectations regarding the study. It can be reduced by performing a
double-blind study with multiple observers to encode and verify recorded data.
(Choice D) Selection bias can occur with inappropriate (ie. nonrandom) assignment methods or through
selective attrition of the study participants. It results in a study population that does not accurately represent
the actual population. leading to erroneous conclusions regarding the exposure-disease relationship.
as these patients generally live longer.
The concept of a latency period can be applied to both disease pathogenesis and exposure to risk modifiers.
Exposure to risk factors and the initial steps in disease pathogenesis sometimes occur years before clinical
manifestations are evident. In addition. exposure to risk modifiers may need to be continuous over a certain
period before influencing the outcome.
Notes
1.":~
Calculator

1 A

Uploaded by

Document Informationclick to expand document informationAnswers to Section 1

Document Informationclick to expand document information

Copyright:

Available Formats

1 A

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

1 A

Uploaded by

Copyright:

Available Formats

w

A. Observer's bias [4%]

Copyright USMLEWorld ,LLC.

Last updated [7/26/20 13]

A. Two-sample t test [58%]

Copyright USMLEWorld ,LLC.

Last updated [7/7/20 10]

Coronary-Artery Bypass Surgery in Patients with Left

Is coronary-artery bypass grafting (CABG) superior to optimal

Median follow-up: 56 months

A 67-year-old male with recently diagnosed coronary artery

r A. Medical therapy comes with a risk of exacerbating

r B. Once you have surgery you might need more

Setting: Multicenter (99 medical centers in 22 countries)

., r C. Surgery carries a higher initial risk of dying [41%]

Patients: Patients with a left ventricular ejection fraction of 35% or

r E. The benefits of surgery are limited to preventing a

The study above is a randomized clinical trial that looks at the

Intervention: CABG plus optimal medical therapy vs optimal medical

The results are reported as a hazard ratio. which is the ratio of

Funding Source: Supported by grants from the National Heart. Lung.

The results are reported as a hazard ratiO , wh1ch 1s the ratiO of

A total of 1212 patients were randomly assigned to medical therapy

used in both groups. and it was the addition of surgery that

Hazard ratios are the ratio of an event rate occurring in the

In this randomized trial, there was no significant difference between

Funding Source: Supported by grants from the National Heart, Lung.

1. Biostatistics primer: what a clinician ought to

Structured abstract is based on NEJM 2011; 364( 17) 1607-16.

Coronary-Artery Bypass Surgery in Patients with Left

Design: Randomized clinical trial

Median follow-up: 56 months

As reported by investigators. by the end of the follow-up period.

A. Lead-time bias [17%]

Setting: Multicenter (99 medical centers in 22 countries)

Patients: Patients with a left ventricular ejection fraction of 35% or

This study was a randomized clinical trial, but many patients in

Intervention: CABG plus optimal medical therapy vs optimal medical

In this study, 17% of the patients in the medical therapy group

Is coronary-artery bypass grafting (CABG) superior to optimal

Median follow-up: 56 months

(Choice A) Lead time bias happens when two interventions

Design: Randomized clinical trial

Setting: Multicenter (99 medical centers in 22 countries)

(Choice B) Measurement bias occurs from poor data

Patients: Patients with a left ventricular ejection fraction of 35% or

(Choice C) Observer bias occurs when the observer may be

Intervention: CABG plus optimal medical therapy vs optimal medical

In this study, 17% of the patients in the medical therapy group

Coronary-Artery Bypass Surgery in Patients with Left

!il P' Mark

Is coronary-artery bypass grafting (CABG) superior to optimal

(Choice B) Measurement bias occurs from poor data

(Choice C) Observer bias occurs when the observer may be

Design: Randomized clinical trial

Median follow-up: 56 months

Coronary-Artery Bypass Surgery in Patients with Left

(Choice A) Lead time bias happens when two interventions