Final: Lectures 20-21

Question 0

How well a test can detect absence of disease when in fact the disease is absent is known as _______.

Answer 0

• Specificity
• Proportion of the time a TEST is negative in a patient that does not have the disease, low false positive rate
• Specificity = TN/(All Not Diseased) x 100%

Question 1

How well a test can detect presence of disease when in fact disease is present is known as _______.

Answer 1

• Sensitivity
• Proportion of time that a TEST is positive in a patient that does have the disease, test has a low false negative rate
• Sensitivity = TP/(All diseased) x 100%

Question 2

Positive Predictive Value (PPV)

Answer 2

• How accurately a positive test predicts the presence of disease
• Percentage of TP’s in patients with a positive test (correct prediction)
• PPV = TP/(All Positive Tests) x 100%

Question 3

Negative Predictive Value (NPV)

Answer 3

• How accurately a negative test predicts the absence of disease
• Percentage of TN’s in patients with a negative test (correct prediction)
• NPV = TN/(All Negative Tests) x 100%

Question 4

Diagnostic Accuracy (DA) or Diagnostic Precision (DP)

Answer 4

• Proportion of the time that a patient is correctly identified as either having a disease or not having a disease with a positive or negative test
• DA/DP = (TP/TN)/(All Patients) x 100%

Question 5

Likelihood Ratio’s (LR)

Answer 5

• Probability of a given test result for a person with the disease ÷ probability of the same test results for a person without the disease
• LR+: probability of + test in presence of disease ÷ probability of + test in absence of disease (Sensitivity ÷ (1-Specificity) Should be >10*
• LR-: Probability of - test in the presence of disease ÷ probability of - test in absence of disease ((1-Sensitivity) ÷ Specificity) Should be <0.1

Question 6

Multiple Cutoff Values

Answer 6

• Many diagnoses typically have 2 dichotomous outcomes (+ or -)
• For screening tests with numerical values, we use ROCs (Receiver Operator Curves)
• More efficient way to show a relationship between sensitivity and specificity for test with numerical outcomes
• Want to hug Y axis until you get close to 1

Question 7

Validity

Answer 7

• Ability to accurately discern between those that do have the disease and those that do not have the disease, “Telling the Truth”
• Internal validity: extent to which results accurately reflect the true situation of the study population
• External validity: Extent to which results are applicable to other populations not included in original study (a.k.a. Generalizability)

Question 8

Reliability

Answer 8

• Ability of a test to give the same result on repeated uses

* Reproducibility/Consistency

Question 9

T/F A valid test is always reliable, so a reliable test always has to be valid.

Answer 9

•False, A reliable test is not always valid!