USMLE STEP 3 EXAMINATION
TEST COMPLETE QUESTIONS
AND SOLUTIONS
◉The curve that graphically represents the family of cutoff points
for a positive vs negative test
The area under this curve is a quantitative measure of the
information content of the test, what are its axes? Answer: ROC
curve
plots sensitivity on the y-axis and (1-spec) on the x-axis
each point on the curve represents a cutoff point of sensitivity and
specificity
the area under the curve can be used as a quantitative measure of
the information content of the test
values range from 0.5 (45 degree line) representing no diagnostic
information to 1.0 for a perfect test
,◉35 YO F with no PMHx presents to clinic with chest pain. She
reports occasional intermittent CP unrelated to exercise, but is
related to spicy food. Pre-test prob for CAD is low; however, the
patient is referred to exercise treadmill stress test, which shows ST
depression after moderate exercise. Using Bayes Theorem, how do
you interpret these results? Answer: The pretest prob is low, and the
sensitivity and specificity of exercise treadmill testing in females is
poor; therefore, the treadmill stress test is not helpful in clinical
decision making in this case
Bayes Theorem is used in an attempt to quantify uncertainty by
employing an equation that combines pretest prob with the testing
characteristics of specificity and sensitivity. The pretest prob
quantitatively describes the clinician's certainty of a diagnosis after
H&P. This equation is:
Posttest prob = pretest prob x test sens // [pretest prob x test sens +
(1-disease prev) x test FP rate]
◉An effective way to measure the accuracy of a diagnostic test is a
positive likelihood ratio [sensitivity/(1-specificity)], which is also
defined as the ratio of the prob of a positive test result in a patient
with the disease to the probability of a positive test result in a
patient without the disease. What other piece of information is
needed along with the PLR to estimate the possibility of a given
disease in a certain patient with a positive test? Answer: pretest
prob of disease
,a PLR can only be interpreted in the context of a pretest prob of
disease. Disease prevalence in certain regions contributes to the
patient's pretest probability. However, other RF like age, clinical hx,
etc are also important in pretest prob
◉Drug X is investigated in a meta-analysis for its effect on mortality
after a MI. It is found that mortality drops from 10% to 2% when the
drug is given.
What is the ARR of drug x?
How many patients will need to be treated with drug X to prevent
one death? Answer: ARR is the difference in mortality (or other
endpoint) between the treatment and the placebo arms. In this case,
the ARR is 10% - 2% or 8%.
From this number, we can calculate the NNT, which is 1/ARR. The
NNT here is 1/.08 or 12.5 patients.
◉23 YO F referred to clinic after being seen in the ED for
intermittent, severe CP. During her visit, she is ruled out for ischemia
with negative biomarkers and EKG. An exercise single photon
emission CT (SPECT) myocardial perfusion scan was performed, and
a reversible exercise-induced perfusion defect was noted. The test
, was read as positive. The patient was placed on ASA. She is
concerned that she continues to have CP on a daily basis without any
consistency in regards to time or antecedent activity. She is
otherwise well. She smokes socially on weekends. She has no famhx
of early coronary artery disease. What should you do next?
A: cardiac cath
B: CT coronary arteries
C: dobutamine stress echo
D: evaluate her for non-cardiac source of CP
E: repeat exercise SPECT test Answer: Based on her age and hx, the
patients pretest prob of CAD is extremely low. Even though the
SPECT is a test with good performance characteristics, a positive test
is only meaningful in a patient with a medium pretest prob of CAD.
This patient's posttest prob is still low to medium. The test should
not have been ordered in the first place and is an example of
defensive medicine. Any further testing could expose the patient to
undue invasive testing and further anxiety. Her ASA should be
stopped, she should be reassured, and other causes of her CP should
be evaluated
◉Which of the following statements regarding gender health is true?
A: Alzheimer's disease affects men and women at equal rates
B: Alzheimer's disease affects men 2x more than women
TEST COMPLETE QUESTIONS
AND SOLUTIONS
◉The curve that graphically represents the family of cutoff points
for a positive vs negative test
The area under this curve is a quantitative measure of the
information content of the test, what are its axes? Answer: ROC
curve
plots sensitivity on the y-axis and (1-spec) on the x-axis
each point on the curve represents a cutoff point of sensitivity and
specificity
the area under the curve can be used as a quantitative measure of
the information content of the test
values range from 0.5 (45 degree line) representing no diagnostic
information to 1.0 for a perfect test
,◉35 YO F with no PMHx presents to clinic with chest pain. She
reports occasional intermittent CP unrelated to exercise, but is
related to spicy food. Pre-test prob for CAD is low; however, the
patient is referred to exercise treadmill stress test, which shows ST
depression after moderate exercise. Using Bayes Theorem, how do
you interpret these results? Answer: The pretest prob is low, and the
sensitivity and specificity of exercise treadmill testing in females is
poor; therefore, the treadmill stress test is not helpful in clinical
decision making in this case
Bayes Theorem is used in an attempt to quantify uncertainty by
employing an equation that combines pretest prob with the testing
characteristics of specificity and sensitivity. The pretest prob
quantitatively describes the clinician's certainty of a diagnosis after
H&P. This equation is:
Posttest prob = pretest prob x test sens // [pretest prob x test sens +
(1-disease prev) x test FP rate]
◉An effective way to measure the accuracy of a diagnostic test is a
positive likelihood ratio [sensitivity/(1-specificity)], which is also
defined as the ratio of the prob of a positive test result in a patient
with the disease to the probability of a positive test result in a
patient without the disease. What other piece of information is
needed along with the PLR to estimate the possibility of a given
disease in a certain patient with a positive test? Answer: pretest
prob of disease
,a PLR can only be interpreted in the context of a pretest prob of
disease. Disease prevalence in certain regions contributes to the
patient's pretest probability. However, other RF like age, clinical hx,
etc are also important in pretest prob
◉Drug X is investigated in a meta-analysis for its effect on mortality
after a MI. It is found that mortality drops from 10% to 2% when the
drug is given.
What is the ARR of drug x?
How many patients will need to be treated with drug X to prevent
one death? Answer: ARR is the difference in mortality (or other
endpoint) between the treatment and the placebo arms. In this case,
the ARR is 10% - 2% or 8%.
From this number, we can calculate the NNT, which is 1/ARR. The
NNT here is 1/.08 or 12.5 patients.
◉23 YO F referred to clinic after being seen in the ED for
intermittent, severe CP. During her visit, she is ruled out for ischemia
with negative biomarkers and EKG. An exercise single photon
emission CT (SPECT) myocardial perfusion scan was performed, and
a reversible exercise-induced perfusion defect was noted. The test
, was read as positive. The patient was placed on ASA. She is
concerned that she continues to have CP on a daily basis without any
consistency in regards to time or antecedent activity. She is
otherwise well. She smokes socially on weekends. She has no famhx
of early coronary artery disease. What should you do next?
A: cardiac cath
B: CT coronary arteries
C: dobutamine stress echo
D: evaluate her for non-cardiac source of CP
E: repeat exercise SPECT test Answer: Based on her age and hx, the
patients pretest prob of CAD is extremely low. Even though the
SPECT is a test with good performance characteristics, a positive test
is only meaningful in a patient with a medium pretest prob of CAD.
This patient's posttest prob is still low to medium. The test should
not have been ordered in the first place and is an example of
defensive medicine. Any further testing could expose the patient to
undue invasive testing and further anxiety. Her ASA should be
stopped, she should be reassured, and other causes of her CP should
be evaluated
◉Which of the following statements regarding gender health is true?
A: Alzheimer's disease affects men and women at equal rates
B: Alzheimer's disease affects men 2x more than women
