Transcript Teaching Diagnostic Reasoning - Centre for Evidence

Teaching Diagnostic
Reasoning
Goutham Rao, MD
Associate Professor, University of
Pittsburgh
Key Concepts Taught
Introduction to Medical Decision
Making

Mandatory course
for first year medical
students
 Replaced “Clinical
epidemiology and
biostatistics” in 2001.
Key Concepts Taught
Core Principles

Traditional journal clubs overemphasize critical appraisal.

Understanding the quantitative aspects of research design and
data analysis promotes an in depth understanding of original
research.

As they read a research article, instead of asking themselves
whether the article is useful or not, students ought to ask, “What
is useful about this article?”

Physician numeracy can form the basis of an EBM course.

The content of an EBM course should be based upon what
numeracy concepts and skills students require to accurately
interpret evidence and apply it to individual patients.
Key Concepts Taught
Historical perspective (e.g. diagnosis in
ancient times).
 Differentiating diagnostic strategies

Induction
 Deduction
 Pattern recognition

Instance based
 Prototype based

Key Concepts Taught
Deductive and Inductive
Reasoning

Arthur is a 24-year-old third year
medical student beginning his
first clinical rotation on an
internal medicine ward in a large
teaching hospital. One morning
he is given two tasks by his
supervising physician. For his
first task, he is told, “Arthur, I
think the patient in room 330 has
rheumatic fever. Go find out if he
has it or not.” For his second
task Arthur is told, “The patient in
room 405 is feeling unwell and
passed out before coming to the
hospital this morning. Let me
know what you believe she may
have.” Arthur has very little
clinical experience. How do you
think he will go about each task?
Which task will prove more
difficult for him?
Key Concepts Taught
Test Characteristics

Scenario:
Your friend Jack has just invented a new test for colon
cancer that involves just taking a small blood sample
and combining it with a reagent that reacts to cancer
cells. Jack developed his test after years of studying
the blood of patients with and without colon cancer and
identifying certain markers in the blood of patients with
cancer that could be identified with a reagent. He asks
you to determine the quality of his new diagnostic test.
How will you go about this task?
Key Concepts Taught
Four Simple Questions






You know your patient’s disease status:
1. “I know my patient has the disease. What is the chance that the
test will show that my patient has it?” = sensitivity
2. “I know my patient does not have the disease. What is the
chance that the test will show that my patient doesn’t have it?” =
specificity
You know your patient’s test status:
3. “I just got a positive test result back on my patient. What is the
chance that my patient actually has the disease?” = PPV
4. “I just got a negative test result back on my patient. What is the
chance that my patient actually doesn’t have the disease?” = NPV
Key Concepts Taught
The Likelihood Ratio
Ratio of two proportions:
The proportion who have a particular test
result (e.g. positive, negative, high
probability) among those with a disease
divided by
The proportion who have the same test
result among those without the disease
LR = TEST RESULT/DISEASE +
TEST RESULT/DISEASE –

Key Concepts Taught
Thomas Bayes (1702 – 1761)




English mathematician
and Minister
Essay towards solving a
problem in the doctrine of
chances
Method for evaluating
new information in
conjunction with prior
information.
Example (Economist)
Key Concepts Taught
Three Patients



Patient One: Mr. D. is an obese sixty-year old smoker with a
long-standing history of diabetes, high blood pressure and high
cholesterol who complains of two hours of “crushing” retrosternal
chest pain.
Patient Two: Ms. A is a twenty-one year old college student who
complains of episodic “twinges” in the left side of her chest for the
past two days. She does not smoke, take any medications, or
have any medical history of significance. She admits to being
under a great deal of stress recently, as final exams are about to
begin.
Patient Three: Mr. Y. is a 47-year old smoker who complains of
an “unusual pressure” in his chest for the past five hours. He has
not had symptoms of this type before. He has no history of
hypertension, diabetes, dyslipidemia or family history of heart
disease. He exercises regularly without any chest discomfort.
Key Concepts Taught
Influence of Pre-Test Probability

Assume that the test
is positive in all
patients:
PrePretest
test
proba odds
bility
Posttest
odds
Mr.
D
75% 3/1
15/1 0.94
Ms.
A
1%
Mr.
Y
25% 1/3
Key Concepts Taught
Posttest
proba
bility
1/99 5/99 .048
1.67/1
.62
Important Principle

Diagnostic tests are most useful in
patients with intermediate pre-test
probabilities.
Key Concepts Taught
Errors in Estimation of Pre-Test
Probabilities
Availability
 Representativeness

Key Concepts Taught
Conjunction Fallacy

134 beginning medical students completed the following problem:
 Amelia is a twenty-three-year-old medical student who comes to your
office for help. You suspect she has the common cold. In the blank
spaces below, based on your knowledge and experience with the
common cold, estimate the probability that Amelia would experience
each of the following symptoms or symptom combinations. For example,
if you believe Amelia has a 100% chance of experiencing “b” and a 90%
chance of experiencing “c” put 100% and 90% in the respective blanks.






runny nose and diarrhea
fatigue
diarrhea
ear pain and shortness of breath
sore throat
headache
Key Concepts Taught
__
__
__
__
__
__
Diagnostic Research: Phase II
example
A new “questionnaire” test has been
developed to determine if a patient is
suffering from a migraine headache. The
patient simply answers the questions
shown below. One point is assigned to
each “yes.”
 Questionnaire administered to 200
sufferers of frequent headaches.
 All patients have gold standard as well.

Key Concepts Taught
ROC Curves
Key Concepts Taught