In: Statistics and Probability
In medicine, epidemiologists are often concerned about the accuracy of their tests for the
presence of certain diseases. That is, if a person has a disease, will our test accurately tell
us the truth? Below I will use the \+" to indicate getting a positive test back and \?" to
indicate a negative test. A positive test means the test indicates the patient has the disease.
Specically, we have the following terms:
Sensitivity: The probability that a test gives a positive result given the subject actually has
the disease; P(+j D)
Specicity: The probability that a test gives a negative result given the subject does not
have the disease; P(? j : D)
Positive Predictive Value: The probability that a subject has the disease given the test
comes back positive; P(D j +)
Negative Predictive Value: The probability that a subject does not have the disease given
the test comes back negative; P(: Dj ?)
Assuming that a disease has a known prevalence in the population of 0.10%, meaning 1 person
in 1000 has the disease, and a given test for the disease has a sensitivity value of 0.997 and a
specicity of 0.985: (18 points total).
(a) Calculate the Positive Predictive Value: Suppose a subject from our population gets a
positive test result. What is the probability this subject has the disease (9 points)?
(b) Calculate the Negative Predictive Value: Suppose a subject from our population gets a
negative test result. What is the probability this subject does not have the disease