|
Why is it important
to use probability to characterize one's uncertainty?
- Precise communication
- Comparison of probability estimates is far more precise
than exchanging verbal assessments. Words like "probably,"
mean vastly different things to different people.
- Taking account
of new information - Probability (in the form of Bayes'
theorem described below) allows for an accurate method of calculating
how much the likelihood of disease changes as new information
(e.g. a test result) becomes available.
Bayes' theorem
is a central principle of medical practice because it helps you
to know what a test result means in your patient. For example,
a radiologist may say, "this chest x-ray looks like pneumonia."
Bayes' theorem tells you the likelihood of pneumonia, given the
chest x-ray results in your patient. It takes into account your
patient's history and physical exam findings and the accuracy
of the chest x-ray in pneumonia.
How does Bayes' theorem
make this goal possible? Examine its form:
| Bayes'
Theorem
|
|---|
| Post-test
odds = pre-test odds x likehood ratio |
|
The post-test odds
of disease are the odds after getting the test result. The pre-test
odds are the odds of disease before getting the test.
Question 1.4.1
What does
Bayes' theorem say about how the characteristics of the patient
influence test interpretation?
 | The
characteristics of the patient have no impact on test interpretation. |
 | The
characteristics of the patient somewhat impact test interpretation. |
 | The
characteristics of the patient are very important determinants
of test interpretation. |
 | None
of the above. |
Question
1.4.2
| Bayes'
Theorem
|
|---|
| Post-test
odds = pre-test odds x likehood ratio |
|
What does Bayes' theorem
tell us about a likelihood ratio? You may not know the definition,
but look at Bayes' theorem and choose the answer that you think
is most appropriate.
| The
likelihood ratio of a test is the only characteristic of the
test that influences the meaning of the test result. |
| The
likelihood ratio of a test is one of three characteristics
that influences the meaning of the test result. |
| The
likelihood ratio of a test is a standardized number that locks
in the meaning of the test result. |
| All
of the above. |
|